Commit
·
5141843
1
Parent(s):
4d52695
Upload folder using huggingface_hub (#1)
Browse files- Upload folder using huggingface_hub (eda17abad36d35144e8d134f533bb7d04495a1ce)
- all_results.json +8 -0
- checkpoint-800/config.json +180 -0
- checkpoint-800/generation_config.json +13 -0
- checkpoint-800/optimizer.pt +3 -0
- checkpoint-800/preprocessor_config.json +22 -0
- checkpoint-800/pytorch_model.bin +3 -0
- checkpoint-800/rng_state.pth +3 -0
- checkpoint-800/scheduler.pt +3 -0
- checkpoint-800/trainer_state.json +2488 -0
- checkpoint-800/training_args.bin +3 -0
- config.json +180 -0
- generation_config.json +13 -0
- preprocessor_config.json +22 -0
- pytorch_model.bin +3 -0
- runs/Aug10_05-52-09_abe03c06359e/events.out.tfevents.1691646743.abe03c06359e.238.0 +3 -0
- runs/Aug10_05-53-22_abe03c06359e/events.out.tfevents.1691646809.abe03c06359e.238.1 +3 -0
- runs/Aug10_05-53-49_abe03c06359e/events.out.tfevents.1691646834.abe03c06359e.238.2 +3 -0
- runs/Aug10_05-54-16_abe03c06359e/events.out.tfevents.1691646860.abe03c06359e.238.3 +3 -0
- runs/Aug10_05-58-15_abe03c06359e/events.out.tfevents.1691647104.abe03c06359e.3240.0 +3 -0
- train_results.json +8 -0
- trainer_state.json +2725 -0
- training_args.bin +3 -0
all_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 8.368943131860664e+17,
|
| 4 |
+
"train_loss": 6.569673863720132,
|
| 5 |
+
"train_runtime": 1471.7945,
|
| 6 |
+
"train_samples_per_second": 4.753,
|
| 7 |
+
"train_steps_per_second": 0.595
|
| 8 |
+
}
|
checkpoint-800/config.json
ADDED
|
@@ -0,0 +1,180 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_commit_hash": "55eb2010aeaaa246defc329d42939e0253d55c99",
|
| 3 |
+
"_name_or_path": "microsoft/trocr-small-handwritten",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"VisionEncoderDecoderModel"
|
| 6 |
+
],
|
| 7 |
+
"decoder": {
|
| 8 |
+
"_name_or_path": "",
|
| 9 |
+
"activation_dropout": 0.0,
|
| 10 |
+
"activation_function": "relu",
|
| 11 |
+
"add_cross_attention": true,
|
| 12 |
+
"architectures": null,
|
| 13 |
+
"attention_dropout": 0.0,
|
| 14 |
+
"bad_words_ids": null,
|
| 15 |
+
"begin_suppress_tokens": null,
|
| 16 |
+
"bos_token_id": 0,
|
| 17 |
+
"chunk_size_feed_forward": 0,
|
| 18 |
+
"classifier_dropout": 0.0,
|
| 19 |
+
"cross_attention_hidden_size": 384,
|
| 20 |
+
"d_model": 256,
|
| 21 |
+
"decoder_attention_heads": 8,
|
| 22 |
+
"decoder_ffn_dim": 1024,
|
| 23 |
+
"decoder_layerdrop": 0.0,
|
| 24 |
+
"decoder_layers": 6,
|
| 25 |
+
"decoder_start_token_id": 2,
|
| 26 |
+
"diversity_penalty": 0.0,
|
| 27 |
+
"do_sample": false,
|
| 28 |
+
"dropout": 0.1,
|
| 29 |
+
"early_stopping": false,
|
| 30 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 31 |
+
"eos_token_id": 2,
|
| 32 |
+
"exponential_decay_length_penalty": null,
|
| 33 |
+
"finetuning_task": null,
|
| 34 |
+
"forced_bos_token_id": null,
|
| 35 |
+
"forced_eos_token_id": null,
|
| 36 |
+
"id2label": {
|
| 37 |
+
"0": "LABEL_0",
|
| 38 |
+
"1": "LABEL_1"
|
| 39 |
+
},
|
| 40 |
+
"init_std": 0.02,
|
| 41 |
+
"is_decoder": true,
|
| 42 |
+
"is_encoder_decoder": false,
|
| 43 |
+
"label2id": {
|
| 44 |
+
"LABEL_0": 0,
|
| 45 |
+
"LABEL_1": 1
|
| 46 |
+
},
|
| 47 |
+
"layernorm_embedding": true,
|
| 48 |
+
"length_penalty": 1.0,
|
| 49 |
+
"max_length": 20,
|
| 50 |
+
"max_position_embeddings": 512,
|
| 51 |
+
"min_length": 0,
|
| 52 |
+
"model_type": "trocr",
|
| 53 |
+
"no_repeat_ngram_size": 0,
|
| 54 |
+
"num_beam_groups": 1,
|
| 55 |
+
"num_beams": 1,
|
| 56 |
+
"num_return_sequences": 1,
|
| 57 |
+
"output_attentions": false,
|
| 58 |
+
"output_hidden_states": false,
|
| 59 |
+
"output_scores": false,
|
| 60 |
+
"pad_token_id": 1,
|
| 61 |
+
"prefix": null,
|
| 62 |
+
"problem_type": null,
|
| 63 |
+
"pruned_heads": {},
|
| 64 |
+
"remove_invalid_values": false,
|
| 65 |
+
"repetition_penalty": 1.0,
|
| 66 |
+
"return_dict": true,
|
| 67 |
+
"return_dict_in_generate": false,
|
| 68 |
+
"scale_embedding": true,
|
| 69 |
+
"sep_token_id": null,
|
| 70 |
+
"suppress_tokens": null,
|
| 71 |
+
"task_specific_params": null,
|
| 72 |
+
"temperature": 1.0,
|
| 73 |
+
"tf_legacy_loss": false,
|
| 74 |
+
"tie_encoder_decoder": false,
|
| 75 |
+
"tie_word_embeddings": false,
|
| 76 |
+
"tokenizer_class": null,
|
| 77 |
+
"top_k": 50,
|
| 78 |
+
"top_p": 1.0,
|
| 79 |
+
"torch_dtype": null,
|
| 80 |
+
"torchscript": false,
|
| 81 |
+
"transformers_version": "4.31.0",
|
| 82 |
+
"typical_p": 1.0,
|
| 83 |
+
"use_bfloat16": false,
|
| 84 |
+
"use_cache": false,
|
| 85 |
+
"use_learned_position_embeddings": true,
|
| 86 |
+
"vocab_size": 64044
|
| 87 |
+
},
|
| 88 |
+
"decoder_start_token_id": 0,
|
| 89 |
+
"early_stopping": true,
|
| 90 |
+
"encoder": {
|
| 91 |
+
"_name_or_path": "",
|
| 92 |
+
"add_cross_attention": false,
|
| 93 |
+
"architectures": null,
|
| 94 |
+
"attention_probs_dropout_prob": 0.0,
|
| 95 |
+
"bad_words_ids": null,
|
| 96 |
+
"begin_suppress_tokens": null,
|
| 97 |
+
"bos_token_id": null,
|
| 98 |
+
"chunk_size_feed_forward": 0,
|
| 99 |
+
"cross_attention_hidden_size": null,
|
| 100 |
+
"decoder_start_token_id": null,
|
| 101 |
+
"diversity_penalty": 0.0,
|
| 102 |
+
"do_sample": false,
|
| 103 |
+
"early_stopping": false,
|
| 104 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 105 |
+
"encoder_stride": 16,
|
| 106 |
+
"eos_token_id": null,
|
| 107 |
+
"exponential_decay_length_penalty": null,
|
| 108 |
+
"finetuning_task": null,
|
| 109 |
+
"forced_bos_token_id": null,
|
| 110 |
+
"forced_eos_token_id": null,
|
| 111 |
+
"hidden_act": "gelu",
|
| 112 |
+
"hidden_dropout_prob": 0.0,
|
| 113 |
+
"hidden_size": 384,
|
| 114 |
+
"id2label": {
|
| 115 |
+
"0": "LABEL_0",
|
| 116 |
+
"1": "LABEL_1"
|
| 117 |
+
},
|
| 118 |
+
"image_size": 384,
|
| 119 |
+
"initializer_range": 0.02,
|
| 120 |
+
"intermediate_size": 1536,
|
| 121 |
+
"is_decoder": false,
|
| 122 |
+
"is_encoder_decoder": false,
|
| 123 |
+
"label2id": {
|
| 124 |
+
"LABEL_0": 0,
|
| 125 |
+
"LABEL_1": 1
|
| 126 |
+
},
|
| 127 |
+
"layer_norm_eps": 1e-12,
|
| 128 |
+
"length_penalty": 1.0,
|
| 129 |
+
"max_length": 20,
|
| 130 |
+
"min_length": 0,
|
| 131 |
+
"model_type": "deit",
|
| 132 |
+
"no_repeat_ngram_size": 0,
|
| 133 |
+
"num_attention_heads": 6,
|
| 134 |
+
"num_beam_groups": 1,
|
| 135 |
+
"num_beams": 1,
|
| 136 |
+
"num_channels": 3,
|
| 137 |
+
"num_hidden_layers": 12,
|
| 138 |
+
"num_return_sequences": 1,
|
| 139 |
+
"output_attentions": false,
|
| 140 |
+
"output_hidden_states": false,
|
| 141 |
+
"output_scores": false,
|
| 142 |
+
"pad_token_id": null,
|
| 143 |
+
"patch_size": 16,
|
| 144 |
+
"prefix": null,
|
| 145 |
+
"problem_type": null,
|
| 146 |
+
"pruned_heads": {},
|
| 147 |
+
"qkv_bias": true,
|
| 148 |
+
"remove_invalid_values": false,
|
| 149 |
+
"repetition_penalty": 1.0,
|
| 150 |
+
"return_dict": true,
|
| 151 |
+
"return_dict_in_generate": false,
|
| 152 |
+
"sep_token_id": null,
|
| 153 |
+
"suppress_tokens": null,
|
| 154 |
+
"task_specific_params": null,
|
| 155 |
+
"temperature": 1.0,
|
| 156 |
+
"tf_legacy_loss": false,
|
| 157 |
+
"tie_encoder_decoder": false,
|
| 158 |
+
"tie_word_embeddings": true,
|
| 159 |
+
"tokenizer_class": null,
|
| 160 |
+
"top_k": 50,
|
| 161 |
+
"top_p": 1.0,
|
| 162 |
+
"torch_dtype": null,
|
| 163 |
+
"torchscript": false,
|
| 164 |
+
"transformers_version": "4.31.0",
|
| 165 |
+
"typical_p": 1.0,
|
| 166 |
+
"use_bfloat16": false
|
| 167 |
+
},
|
| 168 |
+
"eos_token_id": 2,
|
| 169 |
+
"is_encoder_decoder": true,
|
| 170 |
+
"length_penalty": 2.0,
|
| 171 |
+
"max_length": 64,
|
| 172 |
+
"model_type": "vision-encoder-decoder",
|
| 173 |
+
"no_repeat_ngram_size": 4,
|
| 174 |
+
"num_beams": 4,
|
| 175 |
+
"pad_token_id": 1,
|
| 176 |
+
"tie_word_embeddings": false,
|
| 177 |
+
"torch_dtype": "float32",
|
| 178 |
+
"transformers_version": null,
|
| 179 |
+
"vocab_size": 64044
|
| 180 |
+
}
|
checkpoint-800/generation_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 0,
|
| 3 |
+
"decoder_start_token_id": 0,
|
| 4 |
+
"early_stopping": true,
|
| 5 |
+
"eos_token_id": 2,
|
| 6 |
+
"length_penalty": 2.0,
|
| 7 |
+
"max_length": 64,
|
| 8 |
+
"no_repeat_ngram_size": 4,
|
| 9 |
+
"num_beams": 4,
|
| 10 |
+
"pad_token_id": 1,
|
| 11 |
+
"transformers_version": "4.31.0",
|
| 12 |
+
"use_cache": false
|
| 13 |
+
}
|
checkpoint-800/optimizer.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:aa0274fd498081cfb164698fbce9f67890c53433058cedf2b4291f2a8f0f8cab
|
| 3 |
+
size 491802245
|
checkpoint-800/preprocessor_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"do_rescale": true,
|
| 4 |
+
"do_resize": true,
|
| 5 |
+
"image_mean": [
|
| 6 |
+
0.5,
|
| 7 |
+
0.5,
|
| 8 |
+
0.5
|
| 9 |
+
],
|
| 10 |
+
"image_processor_type": "ViTImageProcessor",
|
| 11 |
+
"image_std": [
|
| 12 |
+
0.5,
|
| 13 |
+
0.5,
|
| 14 |
+
0.5
|
| 15 |
+
],
|
| 16 |
+
"resample": 2,
|
| 17 |
+
"rescale_factor": 0.00392156862745098,
|
| 18 |
+
"size": {
|
| 19 |
+
"height": 384,
|
| 20 |
+
"width": 384
|
| 21 |
+
}
|
| 22 |
+
}
|
checkpoint-800/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8556be61324a5710eff0caf21f2d5038e5d2096ae76708ef27ed10b00487ec3c
|
| 3 |
+
size 246512117
|
checkpoint-800/rng_state.pth
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:419fd5db5aaa26f35d45327d2d94e08af5e9efb2b80b017d1c3c15e505ec3cc7
|
| 3 |
+
size 14575
|
checkpoint-800/scheduler.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:06108ef71ef9e2245c7be69207ed5de8b4498768de6b1cf8ed1e59a8833a4c6e
|
| 3 |
+
size 627
|
checkpoint-800/trainer_state.json
ADDED
|
@@ -0,0 +1,2488 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 6.38665771484375,
|
| 3 |
+
"best_model_checkpoint": "/content/drive/MyDrive/AI\\ Camp/words-detector/checkpoint-800",
|
| 4 |
+
"epoch": 2.73972602739726,
|
| 5 |
+
"global_step": 800,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 4.9942922374429226e-05,
|
| 13 |
+
"loss": 17.8686,
|
| 14 |
+
"step": 2
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.01,
|
| 18 |
+
"learning_rate": 4.9942922374429226e-05,
|
| 19 |
+
"loss": 18.2644,
|
| 20 |
+
"step": 4
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.02,
|
| 24 |
+
"learning_rate": 4.9942922374429226e-05,
|
| 25 |
+
"loss": 17.8102,
|
| 26 |
+
"step": 6
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.03,
|
| 30 |
+
"learning_rate": 4.9942922374429226e-05,
|
| 31 |
+
"loss": 18.0033,
|
| 32 |
+
"step": 8
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.03,
|
| 36 |
+
"learning_rate": 4.9828767123287674e-05,
|
| 37 |
+
"loss": 15.0071,
|
| 38 |
+
"step": 10
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.04,
|
| 42 |
+
"learning_rate": 4.971461187214612e-05,
|
| 43 |
+
"loss": 10.0967,
|
| 44 |
+
"step": 12
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.05,
|
| 48 |
+
"learning_rate": 4.960045662100457e-05,
|
| 49 |
+
"loss": 9.4344,
|
| 50 |
+
"step": 14
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.05,
|
| 54 |
+
"learning_rate": 4.948630136986301e-05,
|
| 55 |
+
"loss": 9.3036,
|
| 56 |
+
"step": 16
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.06,
|
| 60 |
+
"learning_rate": 4.937214611872146e-05,
|
| 61 |
+
"loss": 8.8597,
|
| 62 |
+
"step": 18
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.07,
|
| 66 |
+
"learning_rate": 4.9257990867579914e-05,
|
| 67 |
+
"loss": 8.7451,
|
| 68 |
+
"step": 20
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.08,
|
| 72 |
+
"learning_rate": 4.914383561643836e-05,
|
| 73 |
+
"loss": 9.0271,
|
| 74 |
+
"step": 22
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.08,
|
| 78 |
+
"learning_rate": 4.9029680365296804e-05,
|
| 79 |
+
"loss": 8.3687,
|
| 80 |
+
"step": 24
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.09,
|
| 84 |
+
"learning_rate": 4.891552511415525e-05,
|
| 85 |
+
"loss": 8.7046,
|
| 86 |
+
"step": 26
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.1,
|
| 90 |
+
"learning_rate": 4.88013698630137e-05,
|
| 91 |
+
"loss": 8.2987,
|
| 92 |
+
"step": 28
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.1,
|
| 96 |
+
"learning_rate": 4.868721461187215e-05,
|
| 97 |
+
"loss": 8.3789,
|
| 98 |
+
"step": 30
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.11,
|
| 102 |
+
"learning_rate": 4.8573059360730596e-05,
|
| 103 |
+
"loss": 7.7792,
|
| 104 |
+
"step": 32
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.12,
|
| 108 |
+
"learning_rate": 4.8458904109589044e-05,
|
| 109 |
+
"loss": 8.1005,
|
| 110 |
+
"step": 34
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.12,
|
| 114 |
+
"learning_rate": 4.834474885844749e-05,
|
| 115 |
+
"loss": 8.0337,
|
| 116 |
+
"step": 36
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.13,
|
| 120 |
+
"learning_rate": 4.823059360730594e-05,
|
| 121 |
+
"loss": 7.8095,
|
| 122 |
+
"step": 38
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.14,
|
| 126 |
+
"learning_rate": 4.811643835616438e-05,
|
| 127 |
+
"loss": 7.5034,
|
| 128 |
+
"step": 40
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.14,
|
| 132 |
+
"learning_rate": 4.8002283105022836e-05,
|
| 133 |
+
"loss": 7.6029,
|
| 134 |
+
"step": 42
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.15,
|
| 138 |
+
"learning_rate": 4.7888127853881284e-05,
|
| 139 |
+
"loss": 7.7257,
|
| 140 |
+
"step": 44
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.16,
|
| 144 |
+
"learning_rate": 4.777397260273973e-05,
|
| 145 |
+
"loss": 7.4601,
|
| 146 |
+
"step": 46
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.16,
|
| 150 |
+
"learning_rate": 4.765981735159817e-05,
|
| 151 |
+
"loss": 7.7843,
|
| 152 |
+
"step": 48
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.17,
|
| 156 |
+
"learning_rate": 4.754566210045662e-05,
|
| 157 |
+
"loss": 7.7357,
|
| 158 |
+
"step": 50
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.18,
|
| 162 |
+
"learning_rate": 4.743150684931507e-05,
|
| 163 |
+
"loss": 7.1895,
|
| 164 |
+
"step": 52
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.18,
|
| 168 |
+
"learning_rate": 4.7317351598173523e-05,
|
| 169 |
+
"loss": 7.2758,
|
| 170 |
+
"step": 54
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.19,
|
| 174 |
+
"learning_rate": 4.7203196347031965e-05,
|
| 175 |
+
"loss": 7.818,
|
| 176 |
+
"step": 56
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.2,
|
| 180 |
+
"learning_rate": 4.708904109589041e-05,
|
| 181 |
+
"loss": 7.4382,
|
| 182 |
+
"step": 58
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.21,
|
| 186 |
+
"learning_rate": 4.697488584474886e-05,
|
| 187 |
+
"loss": 7.4169,
|
| 188 |
+
"step": 60
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.21,
|
| 192 |
+
"learning_rate": 4.686073059360731e-05,
|
| 193 |
+
"loss": 7.5099,
|
| 194 |
+
"step": 62
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.22,
|
| 198 |
+
"learning_rate": 4.674657534246576e-05,
|
| 199 |
+
"loss": 7.309,
|
| 200 |
+
"step": 64
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.23,
|
| 204 |
+
"learning_rate": 4.6632420091324205e-05,
|
| 205 |
+
"loss": 7.4911,
|
| 206 |
+
"step": 66
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.23,
|
| 210 |
+
"learning_rate": 4.651826484018265e-05,
|
| 211 |
+
"loss": 7.595,
|
| 212 |
+
"step": 68
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.24,
|
| 216 |
+
"learning_rate": 4.64041095890411e-05,
|
| 217 |
+
"loss": 7.3767,
|
| 218 |
+
"step": 70
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.25,
|
| 222 |
+
"learning_rate": 4.628995433789954e-05,
|
| 223 |
+
"loss": 7.3467,
|
| 224 |
+
"step": 72
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.25,
|
| 228 |
+
"learning_rate": 4.617579908675799e-05,
|
| 229 |
+
"loss": 7.4988,
|
| 230 |
+
"step": 74
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.26,
|
| 234 |
+
"learning_rate": 4.6061643835616445e-05,
|
| 235 |
+
"loss": 7.2302,
|
| 236 |
+
"step": 76
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.27,
|
| 240 |
+
"learning_rate": 4.594748858447489e-05,
|
| 241 |
+
"loss": 7.0727,
|
| 242 |
+
"step": 78
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.27,
|
| 246 |
+
"learning_rate": 4.5833333333333334e-05,
|
| 247 |
+
"loss": 7.5031,
|
| 248 |
+
"step": 80
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.28,
|
| 252 |
+
"learning_rate": 4.571917808219178e-05,
|
| 253 |
+
"loss": 7.1357,
|
| 254 |
+
"step": 82
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.29,
|
| 258 |
+
"learning_rate": 4.560502283105023e-05,
|
| 259 |
+
"loss": 7.1202,
|
| 260 |
+
"step": 84
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.29,
|
| 264 |
+
"learning_rate": 4.549086757990868e-05,
|
| 265 |
+
"loss": 7.2538,
|
| 266 |
+
"step": 86
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.3,
|
| 270 |
+
"learning_rate": 4.5376712328767126e-05,
|
| 271 |
+
"loss": 7.071,
|
| 272 |
+
"step": 88
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.31,
|
| 276 |
+
"learning_rate": 4.5262557077625574e-05,
|
| 277 |
+
"loss": 7.5201,
|
| 278 |
+
"step": 90
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.32,
|
| 282 |
+
"learning_rate": 4.514840182648402e-05,
|
| 283 |
+
"loss": 7.5082,
|
| 284 |
+
"step": 92
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.32,
|
| 288 |
+
"learning_rate": 4.503424657534247e-05,
|
| 289 |
+
"loss": 7.1858,
|
| 290 |
+
"step": 94
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.33,
|
| 294 |
+
"learning_rate": 4.492009132420091e-05,
|
| 295 |
+
"loss": 7.2599,
|
| 296 |
+
"step": 96
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.34,
|
| 300 |
+
"learning_rate": 4.4805936073059366e-05,
|
| 301 |
+
"loss": 7.3367,
|
| 302 |
+
"step": 98
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.34,
|
| 306 |
+
"learning_rate": 4.4691780821917814e-05,
|
| 307 |
+
"loss": 6.9713,
|
| 308 |
+
"step": 100
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.34,
|
| 312 |
+
"eval_cer": 0.7946802397175929,
|
| 313 |
+
"eval_loss": 7.137301445007324,
|
| 314 |
+
"eval_runtime": 208.567,
|
| 315 |
+
"eval_samples_per_second": 2.795,
|
| 316 |
+
"eval_steps_per_second": 0.35,
|
| 317 |
+
"step": 100
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.35,
|
| 321 |
+
"learning_rate": 4.457762557077626e-05,
|
| 322 |
+
"loss": 6.8404,
|
| 323 |
+
"step": 102
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 0.36,
|
| 327 |
+
"learning_rate": 4.44634703196347e-05,
|
| 328 |
+
"loss": 7.0133,
|
| 329 |
+
"step": 104
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 0.36,
|
| 333 |
+
"learning_rate": 4.434931506849315e-05,
|
| 334 |
+
"loss": 7.2514,
|
| 335 |
+
"step": 106
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.37,
|
| 339 |
+
"learning_rate": 4.42351598173516e-05,
|
| 340 |
+
"loss": 7.2344,
|
| 341 |
+
"step": 108
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 0.38,
|
| 345 |
+
"learning_rate": 4.412100456621005e-05,
|
| 346 |
+
"loss": 7.1263,
|
| 347 |
+
"step": 110
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 0.38,
|
| 351 |
+
"learning_rate": 4.4006849315068495e-05,
|
| 352 |
+
"loss": 7.1429,
|
| 353 |
+
"step": 112
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.39,
|
| 357 |
+
"learning_rate": 4.389269406392694e-05,
|
| 358 |
+
"loss": 7.422,
|
| 359 |
+
"step": 114
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.4,
|
| 363 |
+
"learning_rate": 4.377853881278539e-05,
|
| 364 |
+
"loss": 7.2147,
|
| 365 |
+
"step": 116
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 0.4,
|
| 369 |
+
"learning_rate": 4.366438356164384e-05,
|
| 370 |
+
"loss": 6.9941,
|
| 371 |
+
"step": 118
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.41,
|
| 375 |
+
"learning_rate": 4.355022831050228e-05,
|
| 376 |
+
"loss": 7.0812,
|
| 377 |
+
"step": 120
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 0.42,
|
| 381 |
+
"learning_rate": 4.3436073059360735e-05,
|
| 382 |
+
"loss": 7.4399,
|
| 383 |
+
"step": 122
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 0.42,
|
| 387 |
+
"learning_rate": 4.332191780821918e-05,
|
| 388 |
+
"loss": 6.9418,
|
| 389 |
+
"step": 124
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 0.43,
|
| 393 |
+
"learning_rate": 4.320776255707763e-05,
|
| 394 |
+
"loss": 7.0305,
|
| 395 |
+
"step": 126
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.44,
|
| 399 |
+
"learning_rate": 4.309360730593607e-05,
|
| 400 |
+
"loss": 7.262,
|
| 401 |
+
"step": 128
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.45,
|
| 405 |
+
"learning_rate": 4.297945205479452e-05,
|
| 406 |
+
"loss": 7.092,
|
| 407 |
+
"step": 130
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 0.45,
|
| 411 |
+
"learning_rate": 4.286529680365297e-05,
|
| 412 |
+
"loss": 6.9877,
|
| 413 |
+
"step": 132
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 0.46,
|
| 417 |
+
"learning_rate": 4.275114155251142e-05,
|
| 418 |
+
"loss": 6.825,
|
| 419 |
+
"step": 134
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 0.47,
|
| 423 |
+
"learning_rate": 4.2636986301369864e-05,
|
| 424 |
+
"loss": 6.6798,
|
| 425 |
+
"step": 136
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 0.47,
|
| 429 |
+
"learning_rate": 4.252283105022831e-05,
|
| 430 |
+
"loss": 6.9598,
|
| 431 |
+
"step": 138
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.48,
|
| 435 |
+
"learning_rate": 4.240867579908676e-05,
|
| 436 |
+
"loss": 6.6182,
|
| 437 |
+
"step": 140
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.49,
|
| 441 |
+
"learning_rate": 4.229452054794521e-05,
|
| 442 |
+
"loss": 7.0775,
|
| 443 |
+
"step": 142
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.49,
|
| 447 |
+
"learning_rate": 4.2180365296803656e-05,
|
| 448 |
+
"loss": 7.0996,
|
| 449 |
+
"step": 144
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 0.5,
|
| 453 |
+
"learning_rate": 4.2066210045662104e-05,
|
| 454 |
+
"loss": 6.9547,
|
| 455 |
+
"step": 146
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 0.51,
|
| 459 |
+
"learning_rate": 4.195205479452055e-05,
|
| 460 |
+
"loss": 6.8326,
|
| 461 |
+
"step": 148
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.51,
|
| 465 |
+
"learning_rate": 4.1837899543379e-05,
|
| 466 |
+
"loss": 7.1501,
|
| 467 |
+
"step": 150
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 0.52,
|
| 471 |
+
"learning_rate": 4.172374429223744e-05,
|
| 472 |
+
"loss": 7.0974,
|
| 473 |
+
"step": 152
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 0.53,
|
| 477 |
+
"learning_rate": 4.160958904109589e-05,
|
| 478 |
+
"loss": 6.8626,
|
| 479 |
+
"step": 154
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.53,
|
| 483 |
+
"learning_rate": 4.1495433789954344e-05,
|
| 484 |
+
"loss": 6.4989,
|
| 485 |
+
"step": 156
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.54,
|
| 489 |
+
"learning_rate": 4.138127853881279e-05,
|
| 490 |
+
"loss": 6.7234,
|
| 491 |
+
"step": 158
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 0.55,
|
| 495 |
+
"learning_rate": 4.126712328767123e-05,
|
| 496 |
+
"loss": 6.7773,
|
| 497 |
+
"step": 160
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 0.55,
|
| 501 |
+
"learning_rate": 4.115296803652968e-05,
|
| 502 |
+
"loss": 6.9259,
|
| 503 |
+
"step": 162
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 0.56,
|
| 507 |
+
"learning_rate": 4.103881278538813e-05,
|
| 508 |
+
"loss": 7.0744,
|
| 509 |
+
"step": 164
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 0.57,
|
| 513 |
+
"learning_rate": 4.092465753424658e-05,
|
| 514 |
+
"loss": 6.8039,
|
| 515 |
+
"step": 166
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 0.58,
|
| 519 |
+
"learning_rate": 4.0810502283105025e-05,
|
| 520 |
+
"loss": 6.8271,
|
| 521 |
+
"step": 168
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.58,
|
| 525 |
+
"learning_rate": 4.069634703196347e-05,
|
| 526 |
+
"loss": 6.7427,
|
| 527 |
+
"step": 170
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.59,
|
| 531 |
+
"learning_rate": 4.058219178082192e-05,
|
| 532 |
+
"loss": 6.5537,
|
| 533 |
+
"step": 172
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 0.6,
|
| 537 |
+
"learning_rate": 4.046803652968037e-05,
|
| 538 |
+
"loss": 6.6001,
|
| 539 |
+
"step": 174
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 0.6,
|
| 543 |
+
"learning_rate": 4.035388127853881e-05,
|
| 544 |
+
"loss": 6.6144,
|
| 545 |
+
"step": 176
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 0.61,
|
| 549 |
+
"learning_rate": 4.0239726027397265e-05,
|
| 550 |
+
"loss": 6.8667,
|
| 551 |
+
"step": 178
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 0.62,
|
| 555 |
+
"learning_rate": 4.012557077625571e-05,
|
| 556 |
+
"loss": 7.0148,
|
| 557 |
+
"step": 180
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.62,
|
| 561 |
+
"learning_rate": 4.001141552511416e-05,
|
| 562 |
+
"loss": 6.7102,
|
| 563 |
+
"step": 182
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.63,
|
| 567 |
+
"learning_rate": 3.98972602739726e-05,
|
| 568 |
+
"loss": 6.8212,
|
| 569 |
+
"step": 184
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.64,
|
| 573 |
+
"learning_rate": 3.978310502283105e-05,
|
| 574 |
+
"loss": 6.6284,
|
| 575 |
+
"step": 186
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 0.64,
|
| 579 |
+
"learning_rate": 3.96689497716895e-05,
|
| 580 |
+
"loss": 6.5087,
|
| 581 |
+
"step": 188
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 0.65,
|
| 585 |
+
"learning_rate": 3.9554794520547946e-05,
|
| 586 |
+
"loss": 7.1345,
|
| 587 |
+
"step": 190
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 0.66,
|
| 591 |
+
"learning_rate": 3.9440639269406394e-05,
|
| 592 |
+
"loss": 7.0863,
|
| 593 |
+
"step": 192
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 0.66,
|
| 597 |
+
"learning_rate": 3.932648401826484e-05,
|
| 598 |
+
"loss": 6.8418,
|
| 599 |
+
"step": 194
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 0.67,
|
| 603 |
+
"learning_rate": 3.921232876712329e-05,
|
| 604 |
+
"loss": 7.0545,
|
| 605 |
+
"step": 196
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.68,
|
| 609 |
+
"learning_rate": 3.909817351598174e-05,
|
| 610 |
+
"loss": 7.1057,
|
| 611 |
+
"step": 198
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.68,
|
| 615 |
+
"learning_rate": 3.8984018264840186e-05,
|
| 616 |
+
"loss": 6.6551,
|
| 617 |
+
"step": 200
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 0.68,
|
| 621 |
+
"eval_cer": 0.8836712913553896,
|
| 622 |
+
"eval_loss": 6.700944900512695,
|
| 623 |
+
"eval_runtime": 40.0981,
|
| 624 |
+
"eval_samples_per_second": 14.539,
|
| 625 |
+
"eval_steps_per_second": 1.821,
|
| 626 |
+
"step": 200
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.69,
|
| 630 |
+
"learning_rate": 3.8869863013698634e-05,
|
| 631 |
+
"loss": 6.7477,
|
| 632 |
+
"step": 202
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.7,
|
| 636 |
+
"learning_rate": 3.875570776255708e-05,
|
| 637 |
+
"loss": 6.9312,
|
| 638 |
+
"step": 204
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.71,
|
| 642 |
+
"learning_rate": 3.864155251141553e-05,
|
| 643 |
+
"loss": 6.6784,
|
| 644 |
+
"step": 206
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.71,
|
| 648 |
+
"learning_rate": 3.852739726027397e-05,
|
| 649 |
+
"loss": 6.9644,
|
| 650 |
+
"step": 208
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.72,
|
| 654 |
+
"learning_rate": 3.841324200913242e-05,
|
| 655 |
+
"loss": 6.5752,
|
| 656 |
+
"step": 210
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.73,
|
| 660 |
+
"learning_rate": 3.829908675799087e-05,
|
| 661 |
+
"loss": 7.0168,
|
| 662 |
+
"step": 212
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.73,
|
| 666 |
+
"learning_rate": 3.818493150684932e-05,
|
| 667 |
+
"loss": 6.7026,
|
| 668 |
+
"step": 214
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.74,
|
| 672 |
+
"learning_rate": 3.8070776255707764e-05,
|
| 673 |
+
"loss": 6.8185,
|
| 674 |
+
"step": 216
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.75,
|
| 678 |
+
"learning_rate": 3.795662100456621e-05,
|
| 679 |
+
"loss": 6.7238,
|
| 680 |
+
"step": 218
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.75,
|
| 684 |
+
"learning_rate": 3.784246575342466e-05,
|
| 685 |
+
"loss": 6.7133,
|
| 686 |
+
"step": 220
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.76,
|
| 690 |
+
"learning_rate": 3.772831050228311e-05,
|
| 691 |
+
"loss": 6.771,
|
| 692 |
+
"step": 222
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.77,
|
| 696 |
+
"learning_rate": 3.7614155251141555e-05,
|
| 697 |
+
"loss": 6.8149,
|
| 698 |
+
"step": 224
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.77,
|
| 702 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 703 |
+
"loss": 6.6512,
|
| 704 |
+
"step": 226
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.78,
|
| 708 |
+
"learning_rate": 3.738584474885845e-05,
|
| 709 |
+
"loss": 6.213,
|
| 710 |
+
"step": 228
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.79,
|
| 714 |
+
"learning_rate": 3.727168949771689e-05,
|
| 715 |
+
"loss": 6.7743,
|
| 716 |
+
"step": 230
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.79,
|
| 720 |
+
"learning_rate": 3.715753424657534e-05,
|
| 721 |
+
"loss": 6.7843,
|
| 722 |
+
"step": 232
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.8,
|
| 726 |
+
"learning_rate": 3.704337899543379e-05,
|
| 727 |
+
"loss": 6.6052,
|
| 728 |
+
"step": 234
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.81,
|
| 732 |
+
"learning_rate": 3.6929223744292243e-05,
|
| 733 |
+
"loss": 6.5319,
|
| 734 |
+
"step": 236
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.82,
|
| 738 |
+
"learning_rate": 3.6815068493150685e-05,
|
| 739 |
+
"loss": 6.4869,
|
| 740 |
+
"step": 238
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.82,
|
| 744 |
+
"learning_rate": 3.670091324200913e-05,
|
| 745 |
+
"loss": 6.7245,
|
| 746 |
+
"step": 240
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.83,
|
| 750 |
+
"learning_rate": 3.658675799086758e-05,
|
| 751 |
+
"loss": 6.8325,
|
| 752 |
+
"step": 242
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.84,
|
| 756 |
+
"learning_rate": 3.647260273972603e-05,
|
| 757 |
+
"loss": 6.6019,
|
| 758 |
+
"step": 244
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.84,
|
| 762 |
+
"learning_rate": 3.635844748858448e-05,
|
| 763 |
+
"loss": 6.2614,
|
| 764 |
+
"step": 246
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 0.85,
|
| 768 |
+
"learning_rate": 3.6244292237442925e-05,
|
| 769 |
+
"loss": 7.0083,
|
| 770 |
+
"step": 248
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.86,
|
| 774 |
+
"learning_rate": 3.613013698630137e-05,
|
| 775 |
+
"loss": 6.4853,
|
| 776 |
+
"step": 250
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.86,
|
| 780 |
+
"learning_rate": 3.601598173515982e-05,
|
| 781 |
+
"loss": 6.7758,
|
| 782 |
+
"step": 252
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.87,
|
| 786 |
+
"learning_rate": 3.590182648401826e-05,
|
| 787 |
+
"loss": 6.7792,
|
| 788 |
+
"step": 254
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.88,
|
| 792 |
+
"learning_rate": 3.578767123287671e-05,
|
| 793 |
+
"loss": 6.6991,
|
| 794 |
+
"step": 256
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.88,
|
| 798 |
+
"learning_rate": 3.5673515981735165e-05,
|
| 799 |
+
"loss": 6.5657,
|
| 800 |
+
"step": 258
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.89,
|
| 804 |
+
"learning_rate": 3.555936073059361e-05,
|
| 805 |
+
"loss": 6.7322,
|
| 806 |
+
"step": 260
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 0.9,
|
| 810 |
+
"learning_rate": 3.5445205479452054e-05,
|
| 811 |
+
"loss": 6.4798,
|
| 812 |
+
"step": 262
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.9,
|
| 816 |
+
"learning_rate": 3.53310502283105e-05,
|
| 817 |
+
"loss": 6.7884,
|
| 818 |
+
"step": 264
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.91,
|
| 822 |
+
"learning_rate": 3.521689497716895e-05,
|
| 823 |
+
"loss": 6.424,
|
| 824 |
+
"step": 266
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.92,
|
| 828 |
+
"learning_rate": 3.51027397260274e-05,
|
| 829 |
+
"loss": 6.6925,
|
| 830 |
+
"step": 268
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.92,
|
| 834 |
+
"learning_rate": 3.4988584474885846e-05,
|
| 835 |
+
"loss": 7.0121,
|
| 836 |
+
"step": 270
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.93,
|
| 840 |
+
"learning_rate": 3.4874429223744294e-05,
|
| 841 |
+
"loss": 6.8289,
|
| 842 |
+
"step": 272
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.94,
|
| 846 |
+
"learning_rate": 3.476027397260274e-05,
|
| 847 |
+
"loss": 6.7504,
|
| 848 |
+
"step": 274
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.95,
|
| 852 |
+
"learning_rate": 3.464611872146119e-05,
|
| 853 |
+
"loss": 6.6898,
|
| 854 |
+
"step": 276
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.95,
|
| 858 |
+
"learning_rate": 3.453196347031963e-05,
|
| 859 |
+
"loss": 6.7379,
|
| 860 |
+
"step": 278
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.96,
|
| 864 |
+
"learning_rate": 3.4417808219178086e-05,
|
| 865 |
+
"loss": 6.4448,
|
| 866 |
+
"step": 280
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.97,
|
| 870 |
+
"learning_rate": 3.4303652968036534e-05,
|
| 871 |
+
"loss": 6.678,
|
| 872 |
+
"step": 282
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.97,
|
| 876 |
+
"learning_rate": 3.418949771689498e-05,
|
| 877 |
+
"loss": 6.4519,
|
| 878 |
+
"step": 284
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.98,
|
| 882 |
+
"learning_rate": 3.407534246575342e-05,
|
| 883 |
+
"loss": 6.5809,
|
| 884 |
+
"step": 286
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.99,
|
| 888 |
+
"learning_rate": 3.396118721461187e-05,
|
| 889 |
+
"loss": 6.5462,
|
| 890 |
+
"step": 288
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.99,
|
| 894 |
+
"learning_rate": 3.384703196347032e-05,
|
| 895 |
+
"loss": 6.4562,
|
| 896 |
+
"step": 290
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 1.0,
|
| 900 |
+
"learning_rate": 3.373287671232877e-05,
|
| 901 |
+
"loss": 6.5566,
|
| 902 |
+
"step": 292
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 1.01,
|
| 906 |
+
"learning_rate": 3.3618721461187215e-05,
|
| 907 |
+
"loss": 6.4929,
|
| 908 |
+
"step": 294
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 1.01,
|
| 912 |
+
"learning_rate": 3.350456621004566e-05,
|
| 913 |
+
"loss": 6.5228,
|
| 914 |
+
"step": 296
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 1.02,
|
| 918 |
+
"learning_rate": 3.339041095890411e-05,
|
| 919 |
+
"loss": 6.6275,
|
| 920 |
+
"step": 298
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 1.03,
|
| 924 |
+
"learning_rate": 3.327625570776256e-05,
|
| 925 |
+
"loss": 6.2089,
|
| 926 |
+
"step": 300
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 1.03,
|
| 930 |
+
"eval_cer": 0.7978408997619243,
|
| 931 |
+
"eval_loss": 6.618570327758789,
|
| 932 |
+
"eval_runtime": 34.5444,
|
| 933 |
+
"eval_samples_per_second": 16.877,
|
| 934 |
+
"eval_steps_per_second": 2.113,
|
| 935 |
+
"step": 300
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 1.03,
|
| 939 |
+
"learning_rate": 3.316210045662101e-05,
|
| 940 |
+
"loss": 6.3685,
|
| 941 |
+
"step": 302
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 1.04,
|
| 945 |
+
"learning_rate": 3.3047945205479455e-05,
|
| 946 |
+
"loss": 6.1731,
|
| 947 |
+
"step": 304
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 1.05,
|
| 951 |
+
"learning_rate": 3.29337899543379e-05,
|
| 952 |
+
"loss": 6.359,
|
| 953 |
+
"step": 306
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 1.05,
|
| 957 |
+
"learning_rate": 3.281963470319635e-05,
|
| 958 |
+
"loss": 6.2789,
|
| 959 |
+
"step": 308
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 1.06,
|
| 963 |
+
"learning_rate": 3.270547945205479e-05,
|
| 964 |
+
"loss": 6.2938,
|
| 965 |
+
"step": 310
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 1.07,
|
| 969 |
+
"learning_rate": 3.259132420091324e-05,
|
| 970 |
+
"loss": 6.3706,
|
| 971 |
+
"step": 312
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 1.08,
|
| 975 |
+
"learning_rate": 3.247716894977169e-05,
|
| 976 |
+
"loss": 6.4628,
|
| 977 |
+
"step": 314
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 1.08,
|
| 981 |
+
"learning_rate": 3.236301369863014e-05,
|
| 982 |
+
"loss": 5.9578,
|
| 983 |
+
"step": 316
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 1.09,
|
| 987 |
+
"learning_rate": 3.2248858447488584e-05,
|
| 988 |
+
"loss": 6.3769,
|
| 989 |
+
"step": 318
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 1.1,
|
| 993 |
+
"learning_rate": 3.213470319634703e-05,
|
| 994 |
+
"loss": 6.7954,
|
| 995 |
+
"step": 320
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 1.1,
|
| 999 |
+
"learning_rate": 3.202054794520548e-05,
|
| 1000 |
+
"loss": 6.7632,
|
| 1001 |
+
"step": 322
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 1.11,
|
| 1005 |
+
"learning_rate": 3.190639269406393e-05,
|
| 1006 |
+
"loss": 6.3459,
|
| 1007 |
+
"step": 324
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 1.12,
|
| 1011 |
+
"learning_rate": 3.1792237442922376e-05,
|
| 1012 |
+
"loss": 6.4654,
|
| 1013 |
+
"step": 326
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 1.12,
|
| 1017 |
+
"learning_rate": 3.1678082191780824e-05,
|
| 1018 |
+
"loss": 6.6141,
|
| 1019 |
+
"step": 328
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 1.13,
|
| 1023 |
+
"learning_rate": 3.156392694063927e-05,
|
| 1024 |
+
"loss": 6.2931,
|
| 1025 |
+
"step": 330
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 1.14,
|
| 1029 |
+
"learning_rate": 3.144977168949772e-05,
|
| 1030 |
+
"loss": 5.9657,
|
| 1031 |
+
"step": 332
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 1.14,
|
| 1035 |
+
"learning_rate": 3.133561643835616e-05,
|
| 1036 |
+
"loss": 6.6218,
|
| 1037 |
+
"step": 334
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 1.15,
|
| 1041 |
+
"learning_rate": 3.122146118721461e-05,
|
| 1042 |
+
"loss": 6.2199,
|
| 1043 |
+
"step": 336
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 1.16,
|
| 1047 |
+
"learning_rate": 3.1107305936073064e-05,
|
| 1048 |
+
"loss": 6.5139,
|
| 1049 |
+
"step": 338
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 1.16,
|
| 1053 |
+
"learning_rate": 3.099315068493151e-05,
|
| 1054 |
+
"loss": 6.5002,
|
| 1055 |
+
"step": 340
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 1.17,
|
| 1059 |
+
"learning_rate": 3.087899543378995e-05,
|
| 1060 |
+
"loss": 6.3457,
|
| 1061 |
+
"step": 342
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 1.18,
|
| 1065 |
+
"learning_rate": 3.07648401826484e-05,
|
| 1066 |
+
"loss": 6.1393,
|
| 1067 |
+
"step": 344
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 1.18,
|
| 1071 |
+
"learning_rate": 3.065068493150685e-05,
|
| 1072 |
+
"loss": 6.0003,
|
| 1073 |
+
"step": 346
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 1.19,
|
| 1077 |
+
"learning_rate": 3.05365296803653e-05,
|
| 1078 |
+
"loss": 6.237,
|
| 1079 |
+
"step": 348
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 1.2,
|
| 1083 |
+
"learning_rate": 3.0422374429223742e-05,
|
| 1084 |
+
"loss": 6.4162,
|
| 1085 |
+
"step": 350
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 1.21,
|
| 1089 |
+
"learning_rate": 3.0308219178082193e-05,
|
| 1090 |
+
"loss": 6.4468,
|
| 1091 |
+
"step": 352
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 1.21,
|
| 1095 |
+
"learning_rate": 3.019406392694064e-05,
|
| 1096 |
+
"loss": 6.6031,
|
| 1097 |
+
"step": 354
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 1.22,
|
| 1101 |
+
"learning_rate": 3.007990867579909e-05,
|
| 1102 |
+
"loss": 6.3145,
|
| 1103 |
+
"step": 356
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 1.23,
|
| 1107 |
+
"learning_rate": 2.9965753424657534e-05,
|
| 1108 |
+
"loss": 6.3375,
|
| 1109 |
+
"step": 358
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.23,
|
| 1113 |
+
"learning_rate": 2.9851598173515982e-05,
|
| 1114 |
+
"loss": 6.4769,
|
| 1115 |
+
"step": 360
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 1.24,
|
| 1119 |
+
"learning_rate": 2.973744292237443e-05,
|
| 1120 |
+
"loss": 6.0631,
|
| 1121 |
+
"step": 362
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 1.25,
|
| 1125 |
+
"learning_rate": 2.962328767123288e-05,
|
| 1126 |
+
"loss": 6.2347,
|
| 1127 |
+
"step": 364
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 1.25,
|
| 1131 |
+
"learning_rate": 2.9509132420091322e-05,
|
| 1132 |
+
"loss": 6.0945,
|
| 1133 |
+
"step": 366
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 1.26,
|
| 1137 |
+
"learning_rate": 2.9394977168949774e-05,
|
| 1138 |
+
"loss": 6.2858,
|
| 1139 |
+
"step": 368
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 1.27,
|
| 1143 |
+
"learning_rate": 2.9280821917808222e-05,
|
| 1144 |
+
"loss": 6.3601,
|
| 1145 |
+
"step": 370
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 1.27,
|
| 1149 |
+
"learning_rate": 2.916666666666667e-05,
|
| 1150 |
+
"loss": 6.3449,
|
| 1151 |
+
"step": 372
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 1.28,
|
| 1155 |
+
"learning_rate": 2.9052511415525114e-05,
|
| 1156 |
+
"loss": 6.3585,
|
| 1157 |
+
"step": 374
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 1.29,
|
| 1161 |
+
"learning_rate": 2.8938356164383562e-05,
|
| 1162 |
+
"loss": 6.0235,
|
| 1163 |
+
"step": 376
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 1.29,
|
| 1167 |
+
"learning_rate": 2.882420091324201e-05,
|
| 1168 |
+
"loss": 6.1927,
|
| 1169 |
+
"step": 378
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 1.3,
|
| 1173 |
+
"learning_rate": 2.871004566210046e-05,
|
| 1174 |
+
"loss": 6.4136,
|
| 1175 |
+
"step": 380
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 1.31,
|
| 1179 |
+
"learning_rate": 2.8595890410958903e-05,
|
| 1180 |
+
"loss": 6.4073,
|
| 1181 |
+
"step": 382
|
| 1182 |
+
},
|
| 1183 |
+
{
|
| 1184 |
+
"epoch": 1.32,
|
| 1185 |
+
"learning_rate": 2.848173515981735e-05,
|
| 1186 |
+
"loss": 6.4599,
|
| 1187 |
+
"step": 384
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 1.32,
|
| 1191 |
+
"learning_rate": 2.8367579908675802e-05,
|
| 1192 |
+
"loss": 6.1341,
|
| 1193 |
+
"step": 386
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 1.33,
|
| 1197 |
+
"learning_rate": 2.825342465753425e-05,
|
| 1198 |
+
"loss": 6.0782,
|
| 1199 |
+
"step": 388
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 1.34,
|
| 1203 |
+
"learning_rate": 2.813926940639269e-05,
|
| 1204 |
+
"loss": 6.1504,
|
| 1205 |
+
"step": 390
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 1.34,
|
| 1209 |
+
"learning_rate": 2.8025114155251143e-05,
|
| 1210 |
+
"loss": 6.3133,
|
| 1211 |
+
"step": 392
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 1.35,
|
| 1215 |
+
"learning_rate": 2.791095890410959e-05,
|
| 1216 |
+
"loss": 6.4356,
|
| 1217 |
+
"step": 394
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 1.36,
|
| 1221 |
+
"learning_rate": 2.779680365296804e-05,
|
| 1222 |
+
"loss": 6.4857,
|
| 1223 |
+
"step": 396
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 1.36,
|
| 1227 |
+
"learning_rate": 2.7682648401826484e-05,
|
| 1228 |
+
"loss": 6.4011,
|
| 1229 |
+
"step": 398
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 1.37,
|
| 1233 |
+
"learning_rate": 2.756849315068493e-05,
|
| 1234 |
+
"loss": 6.1637,
|
| 1235 |
+
"step": 400
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 1.37,
|
| 1239 |
+
"eval_cer": 0.9765208110992529,
|
| 1240 |
+
"eval_loss": 6.528327941894531,
|
| 1241 |
+
"eval_runtime": 37.1998,
|
| 1242 |
+
"eval_samples_per_second": 15.672,
|
| 1243 |
+
"eval_steps_per_second": 1.962,
|
| 1244 |
+
"step": 400
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 1.38,
|
| 1248 |
+
"learning_rate": 2.745433789954338e-05,
|
| 1249 |
+
"loss": 6.3984,
|
| 1250 |
+
"step": 402
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 1.38,
|
| 1254 |
+
"learning_rate": 2.734018264840183e-05,
|
| 1255 |
+
"loss": 6.0221,
|
| 1256 |
+
"step": 404
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 1.39,
|
| 1260 |
+
"learning_rate": 2.7226027397260272e-05,
|
| 1261 |
+
"loss": 6.2472,
|
| 1262 |
+
"step": 406
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 1.4,
|
| 1266 |
+
"learning_rate": 2.7111872146118723e-05,
|
| 1267 |
+
"loss": 6.333,
|
| 1268 |
+
"step": 408
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 1.4,
|
| 1272 |
+
"learning_rate": 2.699771689497717e-05,
|
| 1273 |
+
"loss": 6.2516,
|
| 1274 |
+
"step": 410
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 1.41,
|
| 1278 |
+
"learning_rate": 2.688356164383562e-05,
|
| 1279 |
+
"loss": 6.2314,
|
| 1280 |
+
"step": 412
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 1.42,
|
| 1284 |
+
"learning_rate": 2.6769406392694064e-05,
|
| 1285 |
+
"loss": 6.1203,
|
| 1286 |
+
"step": 414
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 1.42,
|
| 1290 |
+
"learning_rate": 2.6655251141552512e-05,
|
| 1291 |
+
"loss": 6.311,
|
| 1292 |
+
"step": 416
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 1.43,
|
| 1296 |
+
"learning_rate": 2.654109589041096e-05,
|
| 1297 |
+
"loss": 6.1741,
|
| 1298 |
+
"step": 418
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.44,
|
| 1302 |
+
"learning_rate": 2.6426940639269408e-05,
|
| 1303 |
+
"loss": 6.0226,
|
| 1304 |
+
"step": 420
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 1.45,
|
| 1308 |
+
"learning_rate": 2.6312785388127853e-05,
|
| 1309 |
+
"loss": 6.1573,
|
| 1310 |
+
"step": 422
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 1.45,
|
| 1314 |
+
"learning_rate": 2.61986301369863e-05,
|
| 1315 |
+
"loss": 6.3298,
|
| 1316 |
+
"step": 424
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 1.46,
|
| 1320 |
+
"learning_rate": 2.6084474885844752e-05,
|
| 1321 |
+
"loss": 6.3197,
|
| 1322 |
+
"step": 426
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 1.47,
|
| 1326 |
+
"learning_rate": 2.59703196347032e-05,
|
| 1327 |
+
"loss": 6.3265,
|
| 1328 |
+
"step": 428
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 1.47,
|
| 1332 |
+
"learning_rate": 2.5856164383561645e-05,
|
| 1333 |
+
"loss": 6.5904,
|
| 1334 |
+
"step": 430
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 1.48,
|
| 1338 |
+
"learning_rate": 2.5742009132420093e-05,
|
| 1339 |
+
"loss": 6.0295,
|
| 1340 |
+
"step": 432
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 1.49,
|
| 1344 |
+
"learning_rate": 2.562785388127854e-05,
|
| 1345 |
+
"loss": 6.2887,
|
| 1346 |
+
"step": 434
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 1.49,
|
| 1350 |
+
"learning_rate": 2.551369863013699e-05,
|
| 1351 |
+
"loss": 6.2315,
|
| 1352 |
+
"step": 436
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 1.5,
|
| 1356 |
+
"learning_rate": 2.5399543378995433e-05,
|
| 1357 |
+
"loss": 6.1131,
|
| 1358 |
+
"step": 438
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 1.51,
|
| 1362 |
+
"learning_rate": 2.528538812785388e-05,
|
| 1363 |
+
"loss": 6.0897,
|
| 1364 |
+
"step": 440
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 1.51,
|
| 1368 |
+
"learning_rate": 2.517123287671233e-05,
|
| 1369 |
+
"loss": 6.2543,
|
| 1370 |
+
"step": 442
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 1.52,
|
| 1374 |
+
"learning_rate": 2.505707762557078e-05,
|
| 1375 |
+
"loss": 6.2021,
|
| 1376 |
+
"step": 444
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 1.53,
|
| 1380 |
+
"learning_rate": 2.4942922374429225e-05,
|
| 1381 |
+
"loss": 6.2414,
|
| 1382 |
+
"step": 446
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 1.53,
|
| 1386 |
+
"learning_rate": 2.4828767123287673e-05,
|
| 1387 |
+
"loss": 6.3754,
|
| 1388 |
+
"step": 448
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 1.54,
|
| 1392 |
+
"learning_rate": 2.471461187214612e-05,
|
| 1393 |
+
"loss": 6.4199,
|
| 1394 |
+
"step": 450
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 1.55,
|
| 1398 |
+
"learning_rate": 2.4600456621004566e-05,
|
| 1399 |
+
"loss": 6.1184,
|
| 1400 |
+
"step": 452
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 1.55,
|
| 1404 |
+
"learning_rate": 2.4486301369863017e-05,
|
| 1405 |
+
"loss": 6.308,
|
| 1406 |
+
"step": 454
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 1.56,
|
| 1410 |
+
"learning_rate": 2.4372146118721462e-05,
|
| 1411 |
+
"loss": 6.4365,
|
| 1412 |
+
"step": 456
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 1.57,
|
| 1416 |
+
"learning_rate": 2.425799086757991e-05,
|
| 1417 |
+
"loss": 6.0473,
|
| 1418 |
+
"step": 458
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 1.58,
|
| 1422 |
+
"learning_rate": 2.4143835616438358e-05,
|
| 1423 |
+
"loss": 6.055,
|
| 1424 |
+
"step": 460
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 1.58,
|
| 1428 |
+
"learning_rate": 2.4029680365296806e-05,
|
| 1429 |
+
"loss": 6.0705,
|
| 1430 |
+
"step": 462
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 1.59,
|
| 1434 |
+
"learning_rate": 2.391552511415525e-05,
|
| 1435 |
+
"loss": 6.0425,
|
| 1436 |
+
"step": 464
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 1.6,
|
| 1440 |
+
"learning_rate": 2.3801369863013702e-05,
|
| 1441 |
+
"loss": 5.9382,
|
| 1442 |
+
"step": 466
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 1.6,
|
| 1446 |
+
"learning_rate": 2.3687214611872146e-05,
|
| 1447 |
+
"loss": 6.0501,
|
| 1448 |
+
"step": 468
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 1.61,
|
| 1452 |
+
"learning_rate": 2.3573059360730594e-05,
|
| 1453 |
+
"loss": 6.1394,
|
| 1454 |
+
"step": 470
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 1.62,
|
| 1458 |
+
"learning_rate": 2.3458904109589042e-05,
|
| 1459 |
+
"loss": 6.4409,
|
| 1460 |
+
"step": 472
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 1.62,
|
| 1464 |
+
"learning_rate": 2.334474885844749e-05,
|
| 1465 |
+
"loss": 6.1642,
|
| 1466 |
+
"step": 474
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 1.63,
|
| 1470 |
+
"learning_rate": 2.3230593607305935e-05,
|
| 1471 |
+
"loss": 6.1099,
|
| 1472 |
+
"step": 476
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 1.64,
|
| 1476 |
+
"learning_rate": 2.3116438356164386e-05,
|
| 1477 |
+
"loss": 6.1933,
|
| 1478 |
+
"step": 478
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 1.64,
|
| 1482 |
+
"learning_rate": 2.300228310502283e-05,
|
| 1483 |
+
"loss": 6.2047,
|
| 1484 |
+
"step": 480
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 1.65,
|
| 1488 |
+
"learning_rate": 2.288812785388128e-05,
|
| 1489 |
+
"loss": 6.0044,
|
| 1490 |
+
"step": 482
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 1.66,
|
| 1494 |
+
"learning_rate": 2.2773972602739727e-05,
|
| 1495 |
+
"loss": 6.0892,
|
| 1496 |
+
"step": 484
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 1.66,
|
| 1500 |
+
"learning_rate": 2.2659817351598175e-05,
|
| 1501 |
+
"loss": 6.245,
|
| 1502 |
+
"step": 486
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 1.67,
|
| 1506 |
+
"learning_rate": 2.2545662100456623e-05,
|
| 1507 |
+
"loss": 6.1838,
|
| 1508 |
+
"step": 488
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 1.68,
|
| 1512 |
+
"learning_rate": 2.243150684931507e-05,
|
| 1513 |
+
"loss": 5.7336,
|
| 1514 |
+
"step": 490
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 1.68,
|
| 1518 |
+
"learning_rate": 2.2317351598173515e-05,
|
| 1519 |
+
"loss": 6.1649,
|
| 1520 |
+
"step": 492
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 1.69,
|
| 1524 |
+
"learning_rate": 2.2203196347031967e-05,
|
| 1525 |
+
"loss": 6.001,
|
| 1526 |
+
"step": 494
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 1.7,
|
| 1530 |
+
"learning_rate": 2.208904109589041e-05,
|
| 1531 |
+
"loss": 6.6607,
|
| 1532 |
+
"step": 496
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 1.71,
|
| 1536 |
+
"learning_rate": 2.197488584474886e-05,
|
| 1537 |
+
"loss": 5.9693,
|
| 1538 |
+
"step": 498
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 1.71,
|
| 1542 |
+
"learning_rate": 2.1860730593607307e-05,
|
| 1543 |
+
"loss": 6.2159,
|
| 1544 |
+
"step": 500
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 1.71,
|
| 1548 |
+
"eval_cer": 0.8505459321894754,
|
| 1549 |
+
"eval_loss": 6.468786716461182,
|
| 1550 |
+
"eval_runtime": 38.2571,
|
| 1551 |
+
"eval_samples_per_second": 15.239,
|
| 1552 |
+
"eval_steps_per_second": 1.908,
|
| 1553 |
+
"step": 500
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 1.72,
|
| 1557 |
+
"learning_rate": 2.1746575342465755e-05,
|
| 1558 |
+
"loss": 6.5233,
|
| 1559 |
+
"step": 502
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 1.73,
|
| 1563 |
+
"learning_rate": 2.16324200913242e-05,
|
| 1564 |
+
"loss": 6.0243,
|
| 1565 |
+
"step": 504
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 1.73,
|
| 1569 |
+
"learning_rate": 2.151826484018265e-05,
|
| 1570 |
+
"loss": 6.0025,
|
| 1571 |
+
"step": 506
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.74,
|
| 1575 |
+
"learning_rate": 2.1404109589041096e-05,
|
| 1576 |
+
"loss": 6.19,
|
| 1577 |
+
"step": 508
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 1.75,
|
| 1581 |
+
"learning_rate": 2.1289954337899544e-05,
|
| 1582 |
+
"loss": 6.1283,
|
| 1583 |
+
"step": 510
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 1.75,
|
| 1587 |
+
"learning_rate": 2.1175799086757992e-05,
|
| 1588 |
+
"loss": 5.9461,
|
| 1589 |
+
"step": 512
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 1.76,
|
| 1593 |
+
"learning_rate": 2.106164383561644e-05,
|
| 1594 |
+
"loss": 6.163,
|
| 1595 |
+
"step": 514
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 1.77,
|
| 1599 |
+
"learning_rate": 2.0947488584474888e-05,
|
| 1600 |
+
"loss": 5.9075,
|
| 1601 |
+
"step": 516
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 1.77,
|
| 1605 |
+
"learning_rate": 2.0833333333333336e-05,
|
| 1606 |
+
"loss": 6.061,
|
| 1607 |
+
"step": 518
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 1.78,
|
| 1611 |
+
"learning_rate": 2.071917808219178e-05,
|
| 1612 |
+
"loss": 5.8001,
|
| 1613 |
+
"step": 520
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 1.79,
|
| 1617 |
+
"learning_rate": 2.060502283105023e-05,
|
| 1618 |
+
"loss": 6.3543,
|
| 1619 |
+
"step": 522
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 1.79,
|
| 1623 |
+
"learning_rate": 2.0490867579908677e-05,
|
| 1624 |
+
"loss": 6.1044,
|
| 1625 |
+
"step": 524
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 1.8,
|
| 1629 |
+
"learning_rate": 2.0376712328767125e-05,
|
| 1630 |
+
"loss": 6.0392,
|
| 1631 |
+
"step": 526
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 1.81,
|
| 1635 |
+
"learning_rate": 2.0262557077625573e-05,
|
| 1636 |
+
"loss": 6.3613,
|
| 1637 |
+
"step": 528
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 1.82,
|
| 1641 |
+
"learning_rate": 2.014840182648402e-05,
|
| 1642 |
+
"loss": 6.1969,
|
| 1643 |
+
"step": 530
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 1.82,
|
| 1647 |
+
"learning_rate": 2.0034246575342465e-05,
|
| 1648 |
+
"loss": 6.2186,
|
| 1649 |
+
"step": 532
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 1.83,
|
| 1653 |
+
"learning_rate": 1.9920091324200917e-05,
|
| 1654 |
+
"loss": 6.5516,
|
| 1655 |
+
"step": 534
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 1.84,
|
| 1659 |
+
"learning_rate": 1.980593607305936e-05,
|
| 1660 |
+
"loss": 6.2319,
|
| 1661 |
+
"step": 536
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 1.84,
|
| 1665 |
+
"learning_rate": 1.969178082191781e-05,
|
| 1666 |
+
"loss": 6.1139,
|
| 1667 |
+
"step": 538
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 1.85,
|
| 1671 |
+
"learning_rate": 1.9577625570776257e-05,
|
| 1672 |
+
"loss": 6.6036,
|
| 1673 |
+
"step": 540
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 1.86,
|
| 1677 |
+
"learning_rate": 1.9463470319634705e-05,
|
| 1678 |
+
"loss": 6.3099,
|
| 1679 |
+
"step": 542
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"epoch": 1.86,
|
| 1683 |
+
"learning_rate": 1.934931506849315e-05,
|
| 1684 |
+
"loss": 6.2986,
|
| 1685 |
+
"step": 544
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 1.87,
|
| 1689 |
+
"learning_rate": 1.92351598173516e-05,
|
| 1690 |
+
"loss": 6.0614,
|
| 1691 |
+
"step": 546
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 1.88,
|
| 1695 |
+
"learning_rate": 1.9121004566210046e-05,
|
| 1696 |
+
"loss": 5.8707,
|
| 1697 |
+
"step": 548
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 1.88,
|
| 1701 |
+
"learning_rate": 1.9006849315068494e-05,
|
| 1702 |
+
"loss": 6.4741,
|
| 1703 |
+
"step": 550
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 1.89,
|
| 1707 |
+
"learning_rate": 1.8892694063926942e-05,
|
| 1708 |
+
"loss": 5.9351,
|
| 1709 |
+
"step": 552
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 1.9,
|
| 1713 |
+
"learning_rate": 1.877853881278539e-05,
|
| 1714 |
+
"loss": 6.3108,
|
| 1715 |
+
"step": 554
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 1.9,
|
| 1719 |
+
"learning_rate": 1.8664383561643838e-05,
|
| 1720 |
+
"loss": 5.8719,
|
| 1721 |
+
"step": 556
|
| 1722 |
+
},
|
| 1723 |
+
{
|
| 1724 |
+
"epoch": 1.91,
|
| 1725 |
+
"learning_rate": 1.8550228310502286e-05,
|
| 1726 |
+
"loss": 5.9978,
|
| 1727 |
+
"step": 558
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 1.92,
|
| 1731 |
+
"learning_rate": 1.843607305936073e-05,
|
| 1732 |
+
"loss": 5.9883,
|
| 1733 |
+
"step": 560
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 1.92,
|
| 1737 |
+
"learning_rate": 1.832191780821918e-05,
|
| 1738 |
+
"loss": 5.7932,
|
| 1739 |
+
"step": 562
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 1.93,
|
| 1743 |
+
"learning_rate": 1.8207762557077626e-05,
|
| 1744 |
+
"loss": 6.1634,
|
| 1745 |
+
"step": 564
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 1.94,
|
| 1749 |
+
"learning_rate": 1.8093607305936074e-05,
|
| 1750 |
+
"loss": 6.2713,
|
| 1751 |
+
"step": 566
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 1.95,
|
| 1755 |
+
"learning_rate": 1.7979452054794522e-05,
|
| 1756 |
+
"loss": 6.2679,
|
| 1757 |
+
"step": 568
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 1.95,
|
| 1761 |
+
"learning_rate": 1.786529680365297e-05,
|
| 1762 |
+
"loss": 6.2969,
|
| 1763 |
+
"step": 570
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 1.96,
|
| 1767 |
+
"learning_rate": 1.7751141552511415e-05,
|
| 1768 |
+
"loss": 6.0192,
|
| 1769 |
+
"step": 572
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 1.97,
|
| 1773 |
+
"learning_rate": 1.7636986301369866e-05,
|
| 1774 |
+
"loss": 5.9912,
|
| 1775 |
+
"step": 574
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 1.97,
|
| 1779 |
+
"learning_rate": 1.752283105022831e-05,
|
| 1780 |
+
"loss": 6.2804,
|
| 1781 |
+
"step": 576
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 1.98,
|
| 1785 |
+
"learning_rate": 1.740867579908676e-05,
|
| 1786 |
+
"loss": 6.1972,
|
| 1787 |
+
"step": 578
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 1.99,
|
| 1791 |
+
"learning_rate": 1.7294520547945207e-05,
|
| 1792 |
+
"loss": 6.0236,
|
| 1793 |
+
"step": 580
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 1.99,
|
| 1797 |
+
"learning_rate": 1.7180365296803655e-05,
|
| 1798 |
+
"loss": 5.9499,
|
| 1799 |
+
"step": 582
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 2.0,
|
| 1803 |
+
"learning_rate": 1.70662100456621e-05,
|
| 1804 |
+
"loss": 6.1688,
|
| 1805 |
+
"step": 584
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 2.01,
|
| 1809 |
+
"learning_rate": 1.695205479452055e-05,
|
| 1810 |
+
"loss": 5.9289,
|
| 1811 |
+
"step": 586
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 2.01,
|
| 1815 |
+
"learning_rate": 1.6837899543378995e-05,
|
| 1816 |
+
"loss": 5.5866,
|
| 1817 |
+
"step": 588
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 2.02,
|
| 1821 |
+
"learning_rate": 1.6723744292237443e-05,
|
| 1822 |
+
"loss": 6.2259,
|
| 1823 |
+
"step": 590
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 2.03,
|
| 1827 |
+
"learning_rate": 1.660958904109589e-05,
|
| 1828 |
+
"loss": 5.9349,
|
| 1829 |
+
"step": 592
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 2.03,
|
| 1833 |
+
"learning_rate": 1.649543378995434e-05,
|
| 1834 |
+
"loss": 6.0916,
|
| 1835 |
+
"step": 594
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 2.04,
|
| 1839 |
+
"learning_rate": 1.6381278538812787e-05,
|
| 1840 |
+
"loss": 5.8322,
|
| 1841 |
+
"step": 596
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 2.05,
|
| 1845 |
+
"learning_rate": 1.6267123287671235e-05,
|
| 1846 |
+
"loss": 6.1475,
|
| 1847 |
+
"step": 598
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 2.05,
|
| 1851 |
+
"learning_rate": 1.615296803652968e-05,
|
| 1852 |
+
"loss": 6.1342,
|
| 1853 |
+
"step": 600
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 2.05,
|
| 1857 |
+
"eval_cer": 0.8386831951399721,
|
| 1858 |
+
"eval_loss": 6.42148494720459,
|
| 1859 |
+
"eval_runtime": 43.5835,
|
| 1860 |
+
"eval_samples_per_second": 13.377,
|
| 1861 |
+
"eval_steps_per_second": 1.675,
|
| 1862 |
+
"step": 600
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 2.06,
|
| 1866 |
+
"learning_rate": 1.603881278538813e-05,
|
| 1867 |
+
"loss": 5.7958,
|
| 1868 |
+
"step": 602
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 2.07,
|
| 1872 |
+
"learning_rate": 1.5924657534246576e-05,
|
| 1873 |
+
"loss": 6.3319,
|
| 1874 |
+
"step": 604
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 2.08,
|
| 1878 |
+
"learning_rate": 1.5810502283105024e-05,
|
| 1879 |
+
"loss": 6.181,
|
| 1880 |
+
"step": 606
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 2.08,
|
| 1884 |
+
"learning_rate": 1.5696347031963472e-05,
|
| 1885 |
+
"loss": 6.1645,
|
| 1886 |
+
"step": 608
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 2.09,
|
| 1890 |
+
"learning_rate": 1.558219178082192e-05,
|
| 1891 |
+
"loss": 5.8628,
|
| 1892 |
+
"step": 610
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 2.1,
|
| 1896 |
+
"learning_rate": 1.5468036529680365e-05,
|
| 1897 |
+
"loss": 6.1039,
|
| 1898 |
+
"step": 612
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 2.1,
|
| 1902 |
+
"learning_rate": 1.5353881278538816e-05,
|
| 1903 |
+
"loss": 6.196,
|
| 1904 |
+
"step": 614
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 2.11,
|
| 1908 |
+
"learning_rate": 1.523972602739726e-05,
|
| 1909 |
+
"loss": 6.0791,
|
| 1910 |
+
"step": 616
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 2.12,
|
| 1914 |
+
"learning_rate": 1.512557077625571e-05,
|
| 1915 |
+
"loss": 6.2871,
|
| 1916 |
+
"step": 618
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 2.12,
|
| 1920 |
+
"learning_rate": 1.5011415525114157e-05,
|
| 1921 |
+
"loss": 5.8731,
|
| 1922 |
+
"step": 620
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 2.13,
|
| 1926 |
+
"learning_rate": 1.4897260273972605e-05,
|
| 1927 |
+
"loss": 6.1108,
|
| 1928 |
+
"step": 622
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 2.14,
|
| 1932 |
+
"learning_rate": 1.4783105022831051e-05,
|
| 1933 |
+
"loss": 5.9225,
|
| 1934 |
+
"step": 624
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 2.14,
|
| 1938 |
+
"learning_rate": 1.4668949771689499e-05,
|
| 1939 |
+
"loss": 6.1894,
|
| 1940 |
+
"step": 626
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 2.15,
|
| 1944 |
+
"learning_rate": 1.4554794520547945e-05,
|
| 1945 |
+
"loss": 5.7433,
|
| 1946 |
+
"step": 628
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 2.16,
|
| 1950 |
+
"learning_rate": 1.4440639269406395e-05,
|
| 1951 |
+
"loss": 6.0826,
|
| 1952 |
+
"step": 630
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 2.16,
|
| 1956 |
+
"learning_rate": 1.4326484018264841e-05,
|
| 1957 |
+
"loss": 6.0097,
|
| 1958 |
+
"step": 632
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 2.17,
|
| 1962 |
+
"learning_rate": 1.421232876712329e-05,
|
| 1963 |
+
"loss": 5.8454,
|
| 1964 |
+
"step": 634
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 2.18,
|
| 1968 |
+
"learning_rate": 1.4098173515981735e-05,
|
| 1969 |
+
"loss": 6.1941,
|
| 1970 |
+
"step": 636
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 2.18,
|
| 1974 |
+
"learning_rate": 1.3984018264840185e-05,
|
| 1975 |
+
"loss": 5.8484,
|
| 1976 |
+
"step": 638
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 2.19,
|
| 1980 |
+
"learning_rate": 1.3869863013698631e-05,
|
| 1981 |
+
"loss": 6.328,
|
| 1982 |
+
"step": 640
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 2.2,
|
| 1986 |
+
"learning_rate": 1.375570776255708e-05,
|
| 1987 |
+
"loss": 6.3417,
|
| 1988 |
+
"step": 642
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 2.21,
|
| 1992 |
+
"learning_rate": 1.3641552511415526e-05,
|
| 1993 |
+
"loss": 6.1993,
|
| 1994 |
+
"step": 644
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 2.21,
|
| 1998 |
+
"learning_rate": 1.3527397260273974e-05,
|
| 1999 |
+
"loss": 6.0958,
|
| 2000 |
+
"step": 646
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 2.22,
|
| 2004 |
+
"learning_rate": 1.341324200913242e-05,
|
| 2005 |
+
"loss": 6.0639,
|
| 2006 |
+
"step": 648
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 2.23,
|
| 2010 |
+
"learning_rate": 1.329908675799087e-05,
|
| 2011 |
+
"loss": 6.0299,
|
| 2012 |
+
"step": 650
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 2.23,
|
| 2016 |
+
"learning_rate": 1.3184931506849316e-05,
|
| 2017 |
+
"loss": 6.1234,
|
| 2018 |
+
"step": 652
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 2.24,
|
| 2022 |
+
"learning_rate": 1.3070776255707764e-05,
|
| 2023 |
+
"loss": 6.0303,
|
| 2024 |
+
"step": 654
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 2.25,
|
| 2028 |
+
"learning_rate": 1.295662100456621e-05,
|
| 2029 |
+
"loss": 6.0098,
|
| 2030 |
+
"step": 656
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 2.25,
|
| 2034 |
+
"learning_rate": 1.284246575342466e-05,
|
| 2035 |
+
"loss": 5.9058,
|
| 2036 |
+
"step": 658
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 2.26,
|
| 2040 |
+
"learning_rate": 1.2728310502283106e-05,
|
| 2041 |
+
"loss": 6.2086,
|
| 2042 |
+
"step": 660
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 2.27,
|
| 2046 |
+
"learning_rate": 1.2614155251141554e-05,
|
| 2047 |
+
"loss": 6.1667,
|
| 2048 |
+
"step": 662
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 2.27,
|
| 2052 |
+
"learning_rate": 1.25e-05,
|
| 2053 |
+
"loss": 6.26,
|
| 2054 |
+
"step": 664
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 2.28,
|
| 2058 |
+
"learning_rate": 1.2385844748858449e-05,
|
| 2059 |
+
"loss": 5.698,
|
| 2060 |
+
"step": 666
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 2.29,
|
| 2064 |
+
"learning_rate": 1.2271689497716895e-05,
|
| 2065 |
+
"loss": 6.0933,
|
| 2066 |
+
"step": 668
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 2.29,
|
| 2070 |
+
"learning_rate": 1.2157534246575343e-05,
|
| 2071 |
+
"loss": 6.0448,
|
| 2072 |
+
"step": 670
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 2.3,
|
| 2076 |
+
"learning_rate": 1.2043378995433791e-05,
|
| 2077 |
+
"loss": 6.1017,
|
| 2078 |
+
"step": 672
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 2.31,
|
| 2082 |
+
"learning_rate": 1.1929223744292239e-05,
|
| 2083 |
+
"loss": 6.2529,
|
| 2084 |
+
"step": 674
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 2.32,
|
| 2088 |
+
"learning_rate": 1.1815068493150685e-05,
|
| 2089 |
+
"loss": 5.9125,
|
| 2090 |
+
"step": 676
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 2.32,
|
| 2094 |
+
"learning_rate": 1.1700913242009133e-05,
|
| 2095 |
+
"loss": 6.0414,
|
| 2096 |
+
"step": 678
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 2.33,
|
| 2100 |
+
"learning_rate": 1.1586757990867581e-05,
|
| 2101 |
+
"loss": 6.1012,
|
| 2102 |
+
"step": 680
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 2.34,
|
| 2106 |
+
"learning_rate": 1.1472602739726027e-05,
|
| 2107 |
+
"loss": 6.1578,
|
| 2108 |
+
"step": 682
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 2.34,
|
| 2112 |
+
"learning_rate": 1.1358447488584475e-05,
|
| 2113 |
+
"loss": 6.1924,
|
| 2114 |
+
"step": 684
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 2.35,
|
| 2118 |
+
"learning_rate": 1.1244292237442923e-05,
|
| 2119 |
+
"loss": 5.9418,
|
| 2120 |
+
"step": 686
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 2.36,
|
| 2124 |
+
"learning_rate": 1.113013698630137e-05,
|
| 2125 |
+
"loss": 6.0778,
|
| 2126 |
+
"step": 688
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 2.36,
|
| 2130 |
+
"learning_rate": 1.1015981735159818e-05,
|
| 2131 |
+
"loss": 6.1784,
|
| 2132 |
+
"step": 690
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 2.37,
|
| 2136 |
+
"learning_rate": 1.0901826484018266e-05,
|
| 2137 |
+
"loss": 6.2346,
|
| 2138 |
+
"step": 692
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 2.38,
|
| 2142 |
+
"learning_rate": 1.0787671232876714e-05,
|
| 2143 |
+
"loss": 6.0849,
|
| 2144 |
+
"step": 694
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 2.38,
|
| 2148 |
+
"learning_rate": 1.067351598173516e-05,
|
| 2149 |
+
"loss": 6.0025,
|
| 2150 |
+
"step": 696
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 2.39,
|
| 2154 |
+
"learning_rate": 1.0559360730593608e-05,
|
| 2155 |
+
"loss": 6.2221,
|
| 2156 |
+
"step": 698
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 2.4,
|
| 2160 |
+
"learning_rate": 1.0445205479452056e-05,
|
| 2161 |
+
"loss": 5.8295,
|
| 2162 |
+
"step": 700
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 2.4,
|
| 2166 |
+
"eval_cer": 0.7918069124045645,
|
| 2167 |
+
"eval_loss": 6.3999199867248535,
|
| 2168 |
+
"eval_runtime": 40.7886,
|
| 2169 |
+
"eval_samples_per_second": 14.293,
|
| 2170 |
+
"eval_steps_per_second": 1.79,
|
| 2171 |
+
"step": 700
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 2.4,
|
| 2175 |
+
"learning_rate": 1.0331050228310502e-05,
|
| 2176 |
+
"loss": 5.8848,
|
| 2177 |
+
"step": 702
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 2.41,
|
| 2181 |
+
"learning_rate": 1.021689497716895e-05,
|
| 2182 |
+
"loss": 6.0501,
|
| 2183 |
+
"step": 704
|
| 2184 |
+
},
|
| 2185 |
+
{
|
| 2186 |
+
"epoch": 2.42,
|
| 2187 |
+
"learning_rate": 1.0102739726027398e-05,
|
| 2188 |
+
"loss": 6.1242,
|
| 2189 |
+
"step": 706
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 2.42,
|
| 2193 |
+
"learning_rate": 9.988584474885845e-06,
|
| 2194 |
+
"loss": 5.9263,
|
| 2195 |
+
"step": 708
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 2.43,
|
| 2199 |
+
"learning_rate": 9.874429223744293e-06,
|
| 2200 |
+
"loss": 5.897,
|
| 2201 |
+
"step": 710
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 2.44,
|
| 2205 |
+
"learning_rate": 9.76027397260274e-06,
|
| 2206 |
+
"loss": 6.3139,
|
| 2207 |
+
"step": 712
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 2.45,
|
| 2211 |
+
"learning_rate": 9.646118721461189e-06,
|
| 2212 |
+
"loss": 5.9507,
|
| 2213 |
+
"step": 714
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 2.45,
|
| 2217 |
+
"learning_rate": 9.531963470319635e-06,
|
| 2218 |
+
"loss": 5.9616,
|
| 2219 |
+
"step": 716
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"epoch": 2.46,
|
| 2223 |
+
"learning_rate": 9.417808219178083e-06,
|
| 2224 |
+
"loss": 6.0359,
|
| 2225 |
+
"step": 718
|
| 2226 |
+
},
|
| 2227 |
+
{
|
| 2228 |
+
"epoch": 2.47,
|
| 2229 |
+
"learning_rate": 9.303652968036531e-06,
|
| 2230 |
+
"loss": 5.8949,
|
| 2231 |
+
"step": 720
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 2.47,
|
| 2235 |
+
"learning_rate": 9.189497716894977e-06,
|
| 2236 |
+
"loss": 6.1793,
|
| 2237 |
+
"step": 722
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 2.48,
|
| 2241 |
+
"learning_rate": 9.075342465753425e-06,
|
| 2242 |
+
"loss": 5.9515,
|
| 2243 |
+
"step": 724
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 2.49,
|
| 2247 |
+
"learning_rate": 8.961187214611873e-06,
|
| 2248 |
+
"loss": 5.8504,
|
| 2249 |
+
"step": 726
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 2.49,
|
| 2253 |
+
"learning_rate": 8.84703196347032e-06,
|
| 2254 |
+
"loss": 6.0409,
|
| 2255 |
+
"step": 728
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 2.5,
|
| 2259 |
+
"learning_rate": 8.732876712328767e-06,
|
| 2260 |
+
"loss": 6.2968,
|
| 2261 |
+
"step": 730
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 2.51,
|
| 2265 |
+
"learning_rate": 8.618721461187215e-06,
|
| 2266 |
+
"loss": 5.8559,
|
| 2267 |
+
"step": 732
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 2.51,
|
| 2271 |
+
"learning_rate": 8.504566210045663e-06,
|
| 2272 |
+
"loss": 5.7911,
|
| 2273 |
+
"step": 734
|
| 2274 |
+
},
|
| 2275 |
+
{
|
| 2276 |
+
"epoch": 2.52,
|
| 2277 |
+
"learning_rate": 8.39041095890411e-06,
|
| 2278 |
+
"loss": 6.1524,
|
| 2279 |
+
"step": 736
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 2.53,
|
| 2283 |
+
"learning_rate": 8.276255707762558e-06,
|
| 2284 |
+
"loss": 5.7204,
|
| 2285 |
+
"step": 738
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 2.53,
|
| 2289 |
+
"learning_rate": 8.162100456621006e-06,
|
| 2290 |
+
"loss": 6.1016,
|
| 2291 |
+
"step": 740
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 2.54,
|
| 2295 |
+
"learning_rate": 8.047945205479452e-06,
|
| 2296 |
+
"loss": 5.9837,
|
| 2297 |
+
"step": 742
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 2.55,
|
| 2301 |
+
"learning_rate": 7.9337899543379e-06,
|
| 2302 |
+
"loss": 5.8908,
|
| 2303 |
+
"step": 744
|
| 2304 |
+
},
|
| 2305 |
+
{
|
| 2306 |
+
"epoch": 2.55,
|
| 2307 |
+
"learning_rate": 7.819634703196348e-06,
|
| 2308 |
+
"loss": 5.843,
|
| 2309 |
+
"step": 746
|
| 2310 |
+
},
|
| 2311 |
+
{
|
| 2312 |
+
"epoch": 2.56,
|
| 2313 |
+
"learning_rate": 7.705479452054794e-06,
|
| 2314 |
+
"loss": 5.9089,
|
| 2315 |
+
"step": 748
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 2.57,
|
| 2319 |
+
"learning_rate": 7.591324200913242e-06,
|
| 2320 |
+
"loss": 6.0063,
|
| 2321 |
+
"step": 750
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 2.58,
|
| 2325 |
+
"learning_rate": 7.47716894977169e-06,
|
| 2326 |
+
"loss": 5.9461,
|
| 2327 |
+
"step": 752
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 2.58,
|
| 2331 |
+
"learning_rate": 7.3630136986301374e-06,
|
| 2332 |
+
"loss": 5.6393,
|
| 2333 |
+
"step": 754
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 2.59,
|
| 2337 |
+
"learning_rate": 7.2488584474885854e-06,
|
| 2338 |
+
"loss": 6.0984,
|
| 2339 |
+
"step": 756
|
| 2340 |
+
},
|
| 2341 |
+
{
|
| 2342 |
+
"epoch": 2.6,
|
| 2343 |
+
"learning_rate": 7.134703196347033e-06,
|
| 2344 |
+
"loss": 5.7783,
|
| 2345 |
+
"step": 758
|
| 2346 |
+
},
|
| 2347 |
+
{
|
| 2348 |
+
"epoch": 2.6,
|
| 2349 |
+
"learning_rate": 7.02054794520548e-06,
|
| 2350 |
+
"loss": 5.8534,
|
| 2351 |
+
"step": 760
|
| 2352 |
+
},
|
| 2353 |
+
{
|
| 2354 |
+
"epoch": 2.61,
|
| 2355 |
+
"learning_rate": 6.906392694063928e-06,
|
| 2356 |
+
"loss": 5.9853,
|
| 2357 |
+
"step": 762
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 2.62,
|
| 2361 |
+
"learning_rate": 6.792237442922375e-06,
|
| 2362 |
+
"loss": 6.2198,
|
| 2363 |
+
"step": 764
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 2.62,
|
| 2367 |
+
"learning_rate": 6.678082191780823e-06,
|
| 2368 |
+
"loss": 6.1982,
|
| 2369 |
+
"step": 766
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 2.63,
|
| 2373 |
+
"learning_rate": 6.56392694063927e-06,
|
| 2374 |
+
"loss": 5.9753,
|
| 2375 |
+
"step": 768
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 2.64,
|
| 2379 |
+
"learning_rate": 6.449771689497717e-06,
|
| 2380 |
+
"loss": 5.806,
|
| 2381 |
+
"step": 770
|
| 2382 |
+
},
|
| 2383 |
+
{
|
| 2384 |
+
"epoch": 2.64,
|
| 2385 |
+
"learning_rate": 6.335616438356165e-06,
|
| 2386 |
+
"loss": 5.8563,
|
| 2387 |
+
"step": 772
|
| 2388 |
+
},
|
| 2389 |
+
{
|
| 2390 |
+
"epoch": 2.65,
|
| 2391 |
+
"learning_rate": 6.2214611872146115e-06,
|
| 2392 |
+
"loss": 6.0158,
|
| 2393 |
+
"step": 774
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 2.66,
|
| 2397 |
+
"learning_rate": 6.1073059360730594e-06,
|
| 2398 |
+
"loss": 6.2569,
|
| 2399 |
+
"step": 776
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 2.66,
|
| 2403 |
+
"learning_rate": 5.993150684931507e-06,
|
| 2404 |
+
"loss": 5.9379,
|
| 2405 |
+
"step": 778
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 2.67,
|
| 2409 |
+
"learning_rate": 5.878995433789955e-06,
|
| 2410 |
+
"loss": 5.914,
|
| 2411 |
+
"step": 780
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 2.68,
|
| 2415 |
+
"learning_rate": 5.764840182648402e-06,
|
| 2416 |
+
"loss": 5.847,
|
| 2417 |
+
"step": 782
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 2.68,
|
| 2421 |
+
"learning_rate": 5.650684931506849e-06,
|
| 2422 |
+
"loss": 5.9925,
|
| 2423 |
+
"step": 784
|
| 2424 |
+
},
|
| 2425 |
+
{
|
| 2426 |
+
"epoch": 2.69,
|
| 2427 |
+
"learning_rate": 5.536529680365297e-06,
|
| 2428 |
+
"loss": 5.7515,
|
| 2429 |
+
"step": 786
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 2.7,
|
| 2433 |
+
"learning_rate": 5.422374429223744e-06,
|
| 2434 |
+
"loss": 6.0283,
|
| 2435 |
+
"step": 788
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 2.71,
|
| 2439 |
+
"learning_rate": 5.308219178082192e-06,
|
| 2440 |
+
"loss": 5.8207,
|
| 2441 |
+
"step": 790
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 2.71,
|
| 2445 |
+
"learning_rate": 5.194063926940639e-06,
|
| 2446 |
+
"loss": 5.9529,
|
| 2447 |
+
"step": 792
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 2.72,
|
| 2451 |
+
"learning_rate": 5.079908675799086e-06,
|
| 2452 |
+
"loss": 6.3816,
|
| 2453 |
+
"step": 794
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 2.73,
|
| 2457 |
+
"learning_rate": 4.965753424657534e-06,
|
| 2458 |
+
"loss": 5.921,
|
| 2459 |
+
"step": 796
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 2.73,
|
| 2463 |
+
"learning_rate": 4.8515981735159814e-06,
|
| 2464 |
+
"loss": 5.9369,
|
| 2465 |
+
"step": 798
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 2.74,
|
| 2469 |
+
"learning_rate": 4.7374429223744294e-06,
|
| 2470 |
+
"loss": 6.0542,
|
| 2471 |
+
"step": 800
|
| 2472 |
+
},
|
| 2473 |
+
{
|
| 2474 |
+
"epoch": 2.74,
|
| 2475 |
+
"eval_cer": 0.7485838601100074,
|
| 2476 |
+
"eval_loss": 6.38665771484375,
|
| 2477 |
+
"eval_runtime": 43.0264,
|
| 2478 |
+
"eval_samples_per_second": 13.55,
|
| 2479 |
+
"eval_steps_per_second": 1.697,
|
| 2480 |
+
"step": 800
|
| 2481 |
+
}
|
| 2482 |
+
],
|
| 2483 |
+
"max_steps": 876,
|
| 2484 |
+
"num_train_epochs": 3,
|
| 2485 |
+
"total_flos": 7.646410019847537e+17,
|
| 2486 |
+
"trial_name": null,
|
| 2487 |
+
"trial_params": null
|
| 2488 |
+
}
|
checkpoint-800/training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33bde102253d7e820fef744fab4f7922c36508e4f38c4994d8a8022cc5871b9a
|
| 3 |
+
size 4155
|
config.json
ADDED
|
@@ -0,0 +1,180 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_commit_hash": "55eb2010aeaaa246defc329d42939e0253d55c99",
|
| 3 |
+
"_name_or_path": "microsoft/trocr-small-handwritten",
|
| 4 |
+
"architectures": [
|
| 5 |
+
"VisionEncoderDecoderModel"
|
| 6 |
+
],
|
| 7 |
+
"decoder": {
|
| 8 |
+
"_name_or_path": "",
|
| 9 |
+
"activation_dropout": 0.0,
|
| 10 |
+
"activation_function": "relu",
|
| 11 |
+
"add_cross_attention": true,
|
| 12 |
+
"architectures": null,
|
| 13 |
+
"attention_dropout": 0.0,
|
| 14 |
+
"bad_words_ids": null,
|
| 15 |
+
"begin_suppress_tokens": null,
|
| 16 |
+
"bos_token_id": 0,
|
| 17 |
+
"chunk_size_feed_forward": 0,
|
| 18 |
+
"classifier_dropout": 0.0,
|
| 19 |
+
"cross_attention_hidden_size": 384,
|
| 20 |
+
"d_model": 256,
|
| 21 |
+
"decoder_attention_heads": 8,
|
| 22 |
+
"decoder_ffn_dim": 1024,
|
| 23 |
+
"decoder_layerdrop": 0.0,
|
| 24 |
+
"decoder_layers": 6,
|
| 25 |
+
"decoder_start_token_id": 2,
|
| 26 |
+
"diversity_penalty": 0.0,
|
| 27 |
+
"do_sample": false,
|
| 28 |
+
"dropout": 0.1,
|
| 29 |
+
"early_stopping": false,
|
| 30 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 31 |
+
"eos_token_id": 2,
|
| 32 |
+
"exponential_decay_length_penalty": null,
|
| 33 |
+
"finetuning_task": null,
|
| 34 |
+
"forced_bos_token_id": null,
|
| 35 |
+
"forced_eos_token_id": null,
|
| 36 |
+
"id2label": {
|
| 37 |
+
"0": "LABEL_0",
|
| 38 |
+
"1": "LABEL_1"
|
| 39 |
+
},
|
| 40 |
+
"init_std": 0.02,
|
| 41 |
+
"is_decoder": true,
|
| 42 |
+
"is_encoder_decoder": false,
|
| 43 |
+
"label2id": {
|
| 44 |
+
"LABEL_0": 0,
|
| 45 |
+
"LABEL_1": 1
|
| 46 |
+
},
|
| 47 |
+
"layernorm_embedding": true,
|
| 48 |
+
"length_penalty": 1.0,
|
| 49 |
+
"max_length": 20,
|
| 50 |
+
"max_position_embeddings": 512,
|
| 51 |
+
"min_length": 0,
|
| 52 |
+
"model_type": "trocr",
|
| 53 |
+
"no_repeat_ngram_size": 0,
|
| 54 |
+
"num_beam_groups": 1,
|
| 55 |
+
"num_beams": 1,
|
| 56 |
+
"num_return_sequences": 1,
|
| 57 |
+
"output_attentions": false,
|
| 58 |
+
"output_hidden_states": false,
|
| 59 |
+
"output_scores": false,
|
| 60 |
+
"pad_token_id": 1,
|
| 61 |
+
"prefix": null,
|
| 62 |
+
"problem_type": null,
|
| 63 |
+
"pruned_heads": {},
|
| 64 |
+
"remove_invalid_values": false,
|
| 65 |
+
"repetition_penalty": 1.0,
|
| 66 |
+
"return_dict": true,
|
| 67 |
+
"return_dict_in_generate": false,
|
| 68 |
+
"scale_embedding": true,
|
| 69 |
+
"sep_token_id": null,
|
| 70 |
+
"suppress_tokens": null,
|
| 71 |
+
"task_specific_params": null,
|
| 72 |
+
"temperature": 1.0,
|
| 73 |
+
"tf_legacy_loss": false,
|
| 74 |
+
"tie_encoder_decoder": false,
|
| 75 |
+
"tie_word_embeddings": false,
|
| 76 |
+
"tokenizer_class": null,
|
| 77 |
+
"top_k": 50,
|
| 78 |
+
"top_p": 1.0,
|
| 79 |
+
"torch_dtype": null,
|
| 80 |
+
"torchscript": false,
|
| 81 |
+
"transformers_version": "4.31.0",
|
| 82 |
+
"typical_p": 1.0,
|
| 83 |
+
"use_bfloat16": false,
|
| 84 |
+
"use_cache": false,
|
| 85 |
+
"use_learned_position_embeddings": true,
|
| 86 |
+
"vocab_size": 64044
|
| 87 |
+
},
|
| 88 |
+
"decoder_start_token_id": 0,
|
| 89 |
+
"early_stopping": true,
|
| 90 |
+
"encoder": {
|
| 91 |
+
"_name_or_path": "",
|
| 92 |
+
"add_cross_attention": false,
|
| 93 |
+
"architectures": null,
|
| 94 |
+
"attention_probs_dropout_prob": 0.0,
|
| 95 |
+
"bad_words_ids": null,
|
| 96 |
+
"begin_suppress_tokens": null,
|
| 97 |
+
"bos_token_id": null,
|
| 98 |
+
"chunk_size_feed_forward": 0,
|
| 99 |
+
"cross_attention_hidden_size": null,
|
| 100 |
+
"decoder_start_token_id": null,
|
| 101 |
+
"diversity_penalty": 0.0,
|
| 102 |
+
"do_sample": false,
|
| 103 |
+
"early_stopping": false,
|
| 104 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 105 |
+
"encoder_stride": 16,
|
| 106 |
+
"eos_token_id": null,
|
| 107 |
+
"exponential_decay_length_penalty": null,
|
| 108 |
+
"finetuning_task": null,
|
| 109 |
+
"forced_bos_token_id": null,
|
| 110 |
+
"forced_eos_token_id": null,
|
| 111 |
+
"hidden_act": "gelu",
|
| 112 |
+
"hidden_dropout_prob": 0.0,
|
| 113 |
+
"hidden_size": 384,
|
| 114 |
+
"id2label": {
|
| 115 |
+
"0": "LABEL_0",
|
| 116 |
+
"1": "LABEL_1"
|
| 117 |
+
},
|
| 118 |
+
"image_size": 384,
|
| 119 |
+
"initializer_range": 0.02,
|
| 120 |
+
"intermediate_size": 1536,
|
| 121 |
+
"is_decoder": false,
|
| 122 |
+
"is_encoder_decoder": false,
|
| 123 |
+
"label2id": {
|
| 124 |
+
"LABEL_0": 0,
|
| 125 |
+
"LABEL_1": 1
|
| 126 |
+
},
|
| 127 |
+
"layer_norm_eps": 1e-12,
|
| 128 |
+
"length_penalty": 1.0,
|
| 129 |
+
"max_length": 20,
|
| 130 |
+
"min_length": 0,
|
| 131 |
+
"model_type": "deit",
|
| 132 |
+
"no_repeat_ngram_size": 0,
|
| 133 |
+
"num_attention_heads": 6,
|
| 134 |
+
"num_beam_groups": 1,
|
| 135 |
+
"num_beams": 1,
|
| 136 |
+
"num_channels": 3,
|
| 137 |
+
"num_hidden_layers": 12,
|
| 138 |
+
"num_return_sequences": 1,
|
| 139 |
+
"output_attentions": false,
|
| 140 |
+
"output_hidden_states": false,
|
| 141 |
+
"output_scores": false,
|
| 142 |
+
"pad_token_id": null,
|
| 143 |
+
"patch_size": 16,
|
| 144 |
+
"prefix": null,
|
| 145 |
+
"problem_type": null,
|
| 146 |
+
"pruned_heads": {},
|
| 147 |
+
"qkv_bias": true,
|
| 148 |
+
"remove_invalid_values": false,
|
| 149 |
+
"repetition_penalty": 1.0,
|
| 150 |
+
"return_dict": true,
|
| 151 |
+
"return_dict_in_generate": false,
|
| 152 |
+
"sep_token_id": null,
|
| 153 |
+
"suppress_tokens": null,
|
| 154 |
+
"task_specific_params": null,
|
| 155 |
+
"temperature": 1.0,
|
| 156 |
+
"tf_legacy_loss": false,
|
| 157 |
+
"tie_encoder_decoder": false,
|
| 158 |
+
"tie_word_embeddings": true,
|
| 159 |
+
"tokenizer_class": null,
|
| 160 |
+
"top_k": 50,
|
| 161 |
+
"top_p": 1.0,
|
| 162 |
+
"torch_dtype": null,
|
| 163 |
+
"torchscript": false,
|
| 164 |
+
"transformers_version": "4.31.0",
|
| 165 |
+
"typical_p": 1.0,
|
| 166 |
+
"use_bfloat16": false
|
| 167 |
+
},
|
| 168 |
+
"eos_token_id": 2,
|
| 169 |
+
"is_encoder_decoder": true,
|
| 170 |
+
"length_penalty": 2.0,
|
| 171 |
+
"max_length": 64,
|
| 172 |
+
"model_type": "vision-encoder-decoder",
|
| 173 |
+
"no_repeat_ngram_size": 4,
|
| 174 |
+
"num_beams": 4,
|
| 175 |
+
"pad_token_id": 1,
|
| 176 |
+
"tie_word_embeddings": false,
|
| 177 |
+
"torch_dtype": "float32",
|
| 178 |
+
"transformers_version": null,
|
| 179 |
+
"vocab_size": 64044
|
| 180 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,13 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bos_token_id": 0,
|
| 3 |
+
"decoder_start_token_id": 0,
|
| 4 |
+
"early_stopping": true,
|
| 5 |
+
"eos_token_id": 2,
|
| 6 |
+
"length_penalty": 2.0,
|
| 7 |
+
"max_length": 64,
|
| 8 |
+
"no_repeat_ngram_size": 4,
|
| 9 |
+
"num_beams": 4,
|
| 10 |
+
"pad_token_id": 1,
|
| 11 |
+
"transformers_version": "4.31.0",
|
| 12 |
+
"use_cache": false
|
| 13 |
+
}
|
preprocessor_config.json
ADDED
|
@@ -0,0 +1,22 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_normalize": true,
|
| 3 |
+
"do_rescale": true,
|
| 4 |
+
"do_resize": true,
|
| 5 |
+
"image_mean": [
|
| 6 |
+
0.5,
|
| 7 |
+
0.5,
|
| 8 |
+
0.5
|
| 9 |
+
],
|
| 10 |
+
"image_processor_type": "ViTImageProcessor",
|
| 11 |
+
"image_std": [
|
| 12 |
+
0.5,
|
| 13 |
+
0.5,
|
| 14 |
+
0.5
|
| 15 |
+
],
|
| 16 |
+
"resample": 2,
|
| 17 |
+
"rescale_factor": 0.00392156862745098,
|
| 18 |
+
"size": {
|
| 19 |
+
"height": 384,
|
| 20 |
+
"width": 384
|
| 21 |
+
}
|
| 22 |
+
}
|
pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8556be61324a5710eff0caf21f2d5038e5d2096ae76708ef27ed10b00487ec3c
|
| 3 |
+
size 246512117
|
runs/Aug10_05-52-09_abe03c06359e/events.out.tfevents.1691646743.abe03c06359e.238.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8f32735ae35e0693cc9b6d6201da454f272a295233d352ce23c8a7b1fbe697ce
|
| 3 |
+
size 8524
|
runs/Aug10_05-53-22_abe03c06359e/events.out.tfevents.1691646809.abe03c06359e.238.1
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:fc026064c79cdcf393d378be4a86fa4a07ac3bfba1f13036258b1e19e57868ae
|
| 3 |
+
size 8521
|
runs/Aug10_05-53-49_abe03c06359e/events.out.tfevents.1691646834.abe03c06359e.238.2
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:bda626e7da64a2487e79a765bac003e829084ad7e9be92abd5da620724722682
|
| 3 |
+
size 8829
|
runs/Aug10_05-54-16_abe03c06359e/events.out.tfevents.1691646860.abe03c06359e.238.3
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:8e6fd4ef1253734f409d55e8dd7f96de5023fdc433760817699e0d351973a93b
|
| 3 |
+
size 8521
|
runs/Aug10_05-58-15_abe03c06359e/events.out.tfevents.1691647104.abe03c06359e.3240.0
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b0a0f48e2d6a0cf47a321c7547f04a0e4ec2a3e1d5044866c91ea46e1d822724
|
| 3 |
+
size 79984
|
train_results.json
ADDED
|
@@ -0,0 +1,8 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"epoch": 3.0,
|
| 3 |
+
"total_flos": 8.368943131860664e+17,
|
| 4 |
+
"train_loss": 6.569673863720132,
|
| 5 |
+
"train_runtime": 1471.7945,
|
| 6 |
+
"train_samples_per_second": 4.753,
|
| 7 |
+
"train_steps_per_second": 0.595
|
| 8 |
+
}
|
trainer_state.json
ADDED
|
@@ -0,0 +1,2725 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"best_metric": 6.38665771484375,
|
| 3 |
+
"best_model_checkpoint": "/content/drive/MyDrive/AI\\ Camp/words-detector/checkpoint-800",
|
| 4 |
+
"epoch": 3.0,
|
| 5 |
+
"global_step": 876,
|
| 6 |
+
"is_hyper_param_search": false,
|
| 7 |
+
"is_local_process_zero": true,
|
| 8 |
+
"is_world_process_zero": true,
|
| 9 |
+
"log_history": [
|
| 10 |
+
{
|
| 11 |
+
"epoch": 0.01,
|
| 12 |
+
"learning_rate": 4.9942922374429226e-05,
|
| 13 |
+
"loss": 17.8686,
|
| 14 |
+
"step": 2
|
| 15 |
+
},
|
| 16 |
+
{
|
| 17 |
+
"epoch": 0.01,
|
| 18 |
+
"learning_rate": 4.9942922374429226e-05,
|
| 19 |
+
"loss": 18.2644,
|
| 20 |
+
"step": 4
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"epoch": 0.02,
|
| 24 |
+
"learning_rate": 4.9942922374429226e-05,
|
| 25 |
+
"loss": 17.8102,
|
| 26 |
+
"step": 6
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"epoch": 0.03,
|
| 30 |
+
"learning_rate": 4.9942922374429226e-05,
|
| 31 |
+
"loss": 18.0033,
|
| 32 |
+
"step": 8
|
| 33 |
+
},
|
| 34 |
+
{
|
| 35 |
+
"epoch": 0.03,
|
| 36 |
+
"learning_rate": 4.9828767123287674e-05,
|
| 37 |
+
"loss": 15.0071,
|
| 38 |
+
"step": 10
|
| 39 |
+
},
|
| 40 |
+
{
|
| 41 |
+
"epoch": 0.04,
|
| 42 |
+
"learning_rate": 4.971461187214612e-05,
|
| 43 |
+
"loss": 10.0967,
|
| 44 |
+
"step": 12
|
| 45 |
+
},
|
| 46 |
+
{
|
| 47 |
+
"epoch": 0.05,
|
| 48 |
+
"learning_rate": 4.960045662100457e-05,
|
| 49 |
+
"loss": 9.4344,
|
| 50 |
+
"step": 14
|
| 51 |
+
},
|
| 52 |
+
{
|
| 53 |
+
"epoch": 0.05,
|
| 54 |
+
"learning_rate": 4.948630136986301e-05,
|
| 55 |
+
"loss": 9.3036,
|
| 56 |
+
"step": 16
|
| 57 |
+
},
|
| 58 |
+
{
|
| 59 |
+
"epoch": 0.06,
|
| 60 |
+
"learning_rate": 4.937214611872146e-05,
|
| 61 |
+
"loss": 8.8597,
|
| 62 |
+
"step": 18
|
| 63 |
+
},
|
| 64 |
+
{
|
| 65 |
+
"epoch": 0.07,
|
| 66 |
+
"learning_rate": 4.9257990867579914e-05,
|
| 67 |
+
"loss": 8.7451,
|
| 68 |
+
"step": 20
|
| 69 |
+
},
|
| 70 |
+
{
|
| 71 |
+
"epoch": 0.08,
|
| 72 |
+
"learning_rate": 4.914383561643836e-05,
|
| 73 |
+
"loss": 9.0271,
|
| 74 |
+
"step": 22
|
| 75 |
+
},
|
| 76 |
+
{
|
| 77 |
+
"epoch": 0.08,
|
| 78 |
+
"learning_rate": 4.9029680365296804e-05,
|
| 79 |
+
"loss": 8.3687,
|
| 80 |
+
"step": 24
|
| 81 |
+
},
|
| 82 |
+
{
|
| 83 |
+
"epoch": 0.09,
|
| 84 |
+
"learning_rate": 4.891552511415525e-05,
|
| 85 |
+
"loss": 8.7046,
|
| 86 |
+
"step": 26
|
| 87 |
+
},
|
| 88 |
+
{
|
| 89 |
+
"epoch": 0.1,
|
| 90 |
+
"learning_rate": 4.88013698630137e-05,
|
| 91 |
+
"loss": 8.2987,
|
| 92 |
+
"step": 28
|
| 93 |
+
},
|
| 94 |
+
{
|
| 95 |
+
"epoch": 0.1,
|
| 96 |
+
"learning_rate": 4.868721461187215e-05,
|
| 97 |
+
"loss": 8.3789,
|
| 98 |
+
"step": 30
|
| 99 |
+
},
|
| 100 |
+
{
|
| 101 |
+
"epoch": 0.11,
|
| 102 |
+
"learning_rate": 4.8573059360730596e-05,
|
| 103 |
+
"loss": 7.7792,
|
| 104 |
+
"step": 32
|
| 105 |
+
},
|
| 106 |
+
{
|
| 107 |
+
"epoch": 0.12,
|
| 108 |
+
"learning_rate": 4.8458904109589044e-05,
|
| 109 |
+
"loss": 8.1005,
|
| 110 |
+
"step": 34
|
| 111 |
+
},
|
| 112 |
+
{
|
| 113 |
+
"epoch": 0.12,
|
| 114 |
+
"learning_rate": 4.834474885844749e-05,
|
| 115 |
+
"loss": 8.0337,
|
| 116 |
+
"step": 36
|
| 117 |
+
},
|
| 118 |
+
{
|
| 119 |
+
"epoch": 0.13,
|
| 120 |
+
"learning_rate": 4.823059360730594e-05,
|
| 121 |
+
"loss": 7.8095,
|
| 122 |
+
"step": 38
|
| 123 |
+
},
|
| 124 |
+
{
|
| 125 |
+
"epoch": 0.14,
|
| 126 |
+
"learning_rate": 4.811643835616438e-05,
|
| 127 |
+
"loss": 7.5034,
|
| 128 |
+
"step": 40
|
| 129 |
+
},
|
| 130 |
+
{
|
| 131 |
+
"epoch": 0.14,
|
| 132 |
+
"learning_rate": 4.8002283105022836e-05,
|
| 133 |
+
"loss": 7.6029,
|
| 134 |
+
"step": 42
|
| 135 |
+
},
|
| 136 |
+
{
|
| 137 |
+
"epoch": 0.15,
|
| 138 |
+
"learning_rate": 4.7888127853881284e-05,
|
| 139 |
+
"loss": 7.7257,
|
| 140 |
+
"step": 44
|
| 141 |
+
},
|
| 142 |
+
{
|
| 143 |
+
"epoch": 0.16,
|
| 144 |
+
"learning_rate": 4.777397260273973e-05,
|
| 145 |
+
"loss": 7.4601,
|
| 146 |
+
"step": 46
|
| 147 |
+
},
|
| 148 |
+
{
|
| 149 |
+
"epoch": 0.16,
|
| 150 |
+
"learning_rate": 4.765981735159817e-05,
|
| 151 |
+
"loss": 7.7843,
|
| 152 |
+
"step": 48
|
| 153 |
+
},
|
| 154 |
+
{
|
| 155 |
+
"epoch": 0.17,
|
| 156 |
+
"learning_rate": 4.754566210045662e-05,
|
| 157 |
+
"loss": 7.7357,
|
| 158 |
+
"step": 50
|
| 159 |
+
},
|
| 160 |
+
{
|
| 161 |
+
"epoch": 0.18,
|
| 162 |
+
"learning_rate": 4.743150684931507e-05,
|
| 163 |
+
"loss": 7.1895,
|
| 164 |
+
"step": 52
|
| 165 |
+
},
|
| 166 |
+
{
|
| 167 |
+
"epoch": 0.18,
|
| 168 |
+
"learning_rate": 4.7317351598173523e-05,
|
| 169 |
+
"loss": 7.2758,
|
| 170 |
+
"step": 54
|
| 171 |
+
},
|
| 172 |
+
{
|
| 173 |
+
"epoch": 0.19,
|
| 174 |
+
"learning_rate": 4.7203196347031965e-05,
|
| 175 |
+
"loss": 7.818,
|
| 176 |
+
"step": 56
|
| 177 |
+
},
|
| 178 |
+
{
|
| 179 |
+
"epoch": 0.2,
|
| 180 |
+
"learning_rate": 4.708904109589041e-05,
|
| 181 |
+
"loss": 7.4382,
|
| 182 |
+
"step": 58
|
| 183 |
+
},
|
| 184 |
+
{
|
| 185 |
+
"epoch": 0.21,
|
| 186 |
+
"learning_rate": 4.697488584474886e-05,
|
| 187 |
+
"loss": 7.4169,
|
| 188 |
+
"step": 60
|
| 189 |
+
},
|
| 190 |
+
{
|
| 191 |
+
"epoch": 0.21,
|
| 192 |
+
"learning_rate": 4.686073059360731e-05,
|
| 193 |
+
"loss": 7.5099,
|
| 194 |
+
"step": 62
|
| 195 |
+
},
|
| 196 |
+
{
|
| 197 |
+
"epoch": 0.22,
|
| 198 |
+
"learning_rate": 4.674657534246576e-05,
|
| 199 |
+
"loss": 7.309,
|
| 200 |
+
"step": 64
|
| 201 |
+
},
|
| 202 |
+
{
|
| 203 |
+
"epoch": 0.23,
|
| 204 |
+
"learning_rate": 4.6632420091324205e-05,
|
| 205 |
+
"loss": 7.4911,
|
| 206 |
+
"step": 66
|
| 207 |
+
},
|
| 208 |
+
{
|
| 209 |
+
"epoch": 0.23,
|
| 210 |
+
"learning_rate": 4.651826484018265e-05,
|
| 211 |
+
"loss": 7.595,
|
| 212 |
+
"step": 68
|
| 213 |
+
},
|
| 214 |
+
{
|
| 215 |
+
"epoch": 0.24,
|
| 216 |
+
"learning_rate": 4.64041095890411e-05,
|
| 217 |
+
"loss": 7.3767,
|
| 218 |
+
"step": 70
|
| 219 |
+
},
|
| 220 |
+
{
|
| 221 |
+
"epoch": 0.25,
|
| 222 |
+
"learning_rate": 4.628995433789954e-05,
|
| 223 |
+
"loss": 7.3467,
|
| 224 |
+
"step": 72
|
| 225 |
+
},
|
| 226 |
+
{
|
| 227 |
+
"epoch": 0.25,
|
| 228 |
+
"learning_rate": 4.617579908675799e-05,
|
| 229 |
+
"loss": 7.4988,
|
| 230 |
+
"step": 74
|
| 231 |
+
},
|
| 232 |
+
{
|
| 233 |
+
"epoch": 0.26,
|
| 234 |
+
"learning_rate": 4.6061643835616445e-05,
|
| 235 |
+
"loss": 7.2302,
|
| 236 |
+
"step": 76
|
| 237 |
+
},
|
| 238 |
+
{
|
| 239 |
+
"epoch": 0.27,
|
| 240 |
+
"learning_rate": 4.594748858447489e-05,
|
| 241 |
+
"loss": 7.0727,
|
| 242 |
+
"step": 78
|
| 243 |
+
},
|
| 244 |
+
{
|
| 245 |
+
"epoch": 0.27,
|
| 246 |
+
"learning_rate": 4.5833333333333334e-05,
|
| 247 |
+
"loss": 7.5031,
|
| 248 |
+
"step": 80
|
| 249 |
+
},
|
| 250 |
+
{
|
| 251 |
+
"epoch": 0.28,
|
| 252 |
+
"learning_rate": 4.571917808219178e-05,
|
| 253 |
+
"loss": 7.1357,
|
| 254 |
+
"step": 82
|
| 255 |
+
},
|
| 256 |
+
{
|
| 257 |
+
"epoch": 0.29,
|
| 258 |
+
"learning_rate": 4.560502283105023e-05,
|
| 259 |
+
"loss": 7.1202,
|
| 260 |
+
"step": 84
|
| 261 |
+
},
|
| 262 |
+
{
|
| 263 |
+
"epoch": 0.29,
|
| 264 |
+
"learning_rate": 4.549086757990868e-05,
|
| 265 |
+
"loss": 7.2538,
|
| 266 |
+
"step": 86
|
| 267 |
+
},
|
| 268 |
+
{
|
| 269 |
+
"epoch": 0.3,
|
| 270 |
+
"learning_rate": 4.5376712328767126e-05,
|
| 271 |
+
"loss": 7.071,
|
| 272 |
+
"step": 88
|
| 273 |
+
},
|
| 274 |
+
{
|
| 275 |
+
"epoch": 0.31,
|
| 276 |
+
"learning_rate": 4.5262557077625574e-05,
|
| 277 |
+
"loss": 7.5201,
|
| 278 |
+
"step": 90
|
| 279 |
+
},
|
| 280 |
+
{
|
| 281 |
+
"epoch": 0.32,
|
| 282 |
+
"learning_rate": 4.514840182648402e-05,
|
| 283 |
+
"loss": 7.5082,
|
| 284 |
+
"step": 92
|
| 285 |
+
},
|
| 286 |
+
{
|
| 287 |
+
"epoch": 0.32,
|
| 288 |
+
"learning_rate": 4.503424657534247e-05,
|
| 289 |
+
"loss": 7.1858,
|
| 290 |
+
"step": 94
|
| 291 |
+
},
|
| 292 |
+
{
|
| 293 |
+
"epoch": 0.33,
|
| 294 |
+
"learning_rate": 4.492009132420091e-05,
|
| 295 |
+
"loss": 7.2599,
|
| 296 |
+
"step": 96
|
| 297 |
+
},
|
| 298 |
+
{
|
| 299 |
+
"epoch": 0.34,
|
| 300 |
+
"learning_rate": 4.4805936073059366e-05,
|
| 301 |
+
"loss": 7.3367,
|
| 302 |
+
"step": 98
|
| 303 |
+
},
|
| 304 |
+
{
|
| 305 |
+
"epoch": 0.34,
|
| 306 |
+
"learning_rate": 4.4691780821917814e-05,
|
| 307 |
+
"loss": 6.9713,
|
| 308 |
+
"step": 100
|
| 309 |
+
},
|
| 310 |
+
{
|
| 311 |
+
"epoch": 0.34,
|
| 312 |
+
"eval_cer": 0.7946802397175929,
|
| 313 |
+
"eval_loss": 7.137301445007324,
|
| 314 |
+
"eval_runtime": 208.567,
|
| 315 |
+
"eval_samples_per_second": 2.795,
|
| 316 |
+
"eval_steps_per_second": 0.35,
|
| 317 |
+
"step": 100
|
| 318 |
+
},
|
| 319 |
+
{
|
| 320 |
+
"epoch": 0.35,
|
| 321 |
+
"learning_rate": 4.457762557077626e-05,
|
| 322 |
+
"loss": 6.8404,
|
| 323 |
+
"step": 102
|
| 324 |
+
},
|
| 325 |
+
{
|
| 326 |
+
"epoch": 0.36,
|
| 327 |
+
"learning_rate": 4.44634703196347e-05,
|
| 328 |
+
"loss": 7.0133,
|
| 329 |
+
"step": 104
|
| 330 |
+
},
|
| 331 |
+
{
|
| 332 |
+
"epoch": 0.36,
|
| 333 |
+
"learning_rate": 4.434931506849315e-05,
|
| 334 |
+
"loss": 7.2514,
|
| 335 |
+
"step": 106
|
| 336 |
+
},
|
| 337 |
+
{
|
| 338 |
+
"epoch": 0.37,
|
| 339 |
+
"learning_rate": 4.42351598173516e-05,
|
| 340 |
+
"loss": 7.2344,
|
| 341 |
+
"step": 108
|
| 342 |
+
},
|
| 343 |
+
{
|
| 344 |
+
"epoch": 0.38,
|
| 345 |
+
"learning_rate": 4.412100456621005e-05,
|
| 346 |
+
"loss": 7.1263,
|
| 347 |
+
"step": 110
|
| 348 |
+
},
|
| 349 |
+
{
|
| 350 |
+
"epoch": 0.38,
|
| 351 |
+
"learning_rate": 4.4006849315068495e-05,
|
| 352 |
+
"loss": 7.1429,
|
| 353 |
+
"step": 112
|
| 354 |
+
},
|
| 355 |
+
{
|
| 356 |
+
"epoch": 0.39,
|
| 357 |
+
"learning_rate": 4.389269406392694e-05,
|
| 358 |
+
"loss": 7.422,
|
| 359 |
+
"step": 114
|
| 360 |
+
},
|
| 361 |
+
{
|
| 362 |
+
"epoch": 0.4,
|
| 363 |
+
"learning_rate": 4.377853881278539e-05,
|
| 364 |
+
"loss": 7.2147,
|
| 365 |
+
"step": 116
|
| 366 |
+
},
|
| 367 |
+
{
|
| 368 |
+
"epoch": 0.4,
|
| 369 |
+
"learning_rate": 4.366438356164384e-05,
|
| 370 |
+
"loss": 6.9941,
|
| 371 |
+
"step": 118
|
| 372 |
+
},
|
| 373 |
+
{
|
| 374 |
+
"epoch": 0.41,
|
| 375 |
+
"learning_rate": 4.355022831050228e-05,
|
| 376 |
+
"loss": 7.0812,
|
| 377 |
+
"step": 120
|
| 378 |
+
},
|
| 379 |
+
{
|
| 380 |
+
"epoch": 0.42,
|
| 381 |
+
"learning_rate": 4.3436073059360735e-05,
|
| 382 |
+
"loss": 7.4399,
|
| 383 |
+
"step": 122
|
| 384 |
+
},
|
| 385 |
+
{
|
| 386 |
+
"epoch": 0.42,
|
| 387 |
+
"learning_rate": 4.332191780821918e-05,
|
| 388 |
+
"loss": 6.9418,
|
| 389 |
+
"step": 124
|
| 390 |
+
},
|
| 391 |
+
{
|
| 392 |
+
"epoch": 0.43,
|
| 393 |
+
"learning_rate": 4.320776255707763e-05,
|
| 394 |
+
"loss": 7.0305,
|
| 395 |
+
"step": 126
|
| 396 |
+
},
|
| 397 |
+
{
|
| 398 |
+
"epoch": 0.44,
|
| 399 |
+
"learning_rate": 4.309360730593607e-05,
|
| 400 |
+
"loss": 7.262,
|
| 401 |
+
"step": 128
|
| 402 |
+
},
|
| 403 |
+
{
|
| 404 |
+
"epoch": 0.45,
|
| 405 |
+
"learning_rate": 4.297945205479452e-05,
|
| 406 |
+
"loss": 7.092,
|
| 407 |
+
"step": 130
|
| 408 |
+
},
|
| 409 |
+
{
|
| 410 |
+
"epoch": 0.45,
|
| 411 |
+
"learning_rate": 4.286529680365297e-05,
|
| 412 |
+
"loss": 6.9877,
|
| 413 |
+
"step": 132
|
| 414 |
+
},
|
| 415 |
+
{
|
| 416 |
+
"epoch": 0.46,
|
| 417 |
+
"learning_rate": 4.275114155251142e-05,
|
| 418 |
+
"loss": 6.825,
|
| 419 |
+
"step": 134
|
| 420 |
+
},
|
| 421 |
+
{
|
| 422 |
+
"epoch": 0.47,
|
| 423 |
+
"learning_rate": 4.2636986301369864e-05,
|
| 424 |
+
"loss": 6.6798,
|
| 425 |
+
"step": 136
|
| 426 |
+
},
|
| 427 |
+
{
|
| 428 |
+
"epoch": 0.47,
|
| 429 |
+
"learning_rate": 4.252283105022831e-05,
|
| 430 |
+
"loss": 6.9598,
|
| 431 |
+
"step": 138
|
| 432 |
+
},
|
| 433 |
+
{
|
| 434 |
+
"epoch": 0.48,
|
| 435 |
+
"learning_rate": 4.240867579908676e-05,
|
| 436 |
+
"loss": 6.6182,
|
| 437 |
+
"step": 140
|
| 438 |
+
},
|
| 439 |
+
{
|
| 440 |
+
"epoch": 0.49,
|
| 441 |
+
"learning_rate": 4.229452054794521e-05,
|
| 442 |
+
"loss": 7.0775,
|
| 443 |
+
"step": 142
|
| 444 |
+
},
|
| 445 |
+
{
|
| 446 |
+
"epoch": 0.49,
|
| 447 |
+
"learning_rate": 4.2180365296803656e-05,
|
| 448 |
+
"loss": 7.0996,
|
| 449 |
+
"step": 144
|
| 450 |
+
},
|
| 451 |
+
{
|
| 452 |
+
"epoch": 0.5,
|
| 453 |
+
"learning_rate": 4.2066210045662104e-05,
|
| 454 |
+
"loss": 6.9547,
|
| 455 |
+
"step": 146
|
| 456 |
+
},
|
| 457 |
+
{
|
| 458 |
+
"epoch": 0.51,
|
| 459 |
+
"learning_rate": 4.195205479452055e-05,
|
| 460 |
+
"loss": 6.8326,
|
| 461 |
+
"step": 148
|
| 462 |
+
},
|
| 463 |
+
{
|
| 464 |
+
"epoch": 0.51,
|
| 465 |
+
"learning_rate": 4.1837899543379e-05,
|
| 466 |
+
"loss": 7.1501,
|
| 467 |
+
"step": 150
|
| 468 |
+
},
|
| 469 |
+
{
|
| 470 |
+
"epoch": 0.52,
|
| 471 |
+
"learning_rate": 4.172374429223744e-05,
|
| 472 |
+
"loss": 7.0974,
|
| 473 |
+
"step": 152
|
| 474 |
+
},
|
| 475 |
+
{
|
| 476 |
+
"epoch": 0.53,
|
| 477 |
+
"learning_rate": 4.160958904109589e-05,
|
| 478 |
+
"loss": 6.8626,
|
| 479 |
+
"step": 154
|
| 480 |
+
},
|
| 481 |
+
{
|
| 482 |
+
"epoch": 0.53,
|
| 483 |
+
"learning_rate": 4.1495433789954344e-05,
|
| 484 |
+
"loss": 6.4989,
|
| 485 |
+
"step": 156
|
| 486 |
+
},
|
| 487 |
+
{
|
| 488 |
+
"epoch": 0.54,
|
| 489 |
+
"learning_rate": 4.138127853881279e-05,
|
| 490 |
+
"loss": 6.7234,
|
| 491 |
+
"step": 158
|
| 492 |
+
},
|
| 493 |
+
{
|
| 494 |
+
"epoch": 0.55,
|
| 495 |
+
"learning_rate": 4.126712328767123e-05,
|
| 496 |
+
"loss": 6.7773,
|
| 497 |
+
"step": 160
|
| 498 |
+
},
|
| 499 |
+
{
|
| 500 |
+
"epoch": 0.55,
|
| 501 |
+
"learning_rate": 4.115296803652968e-05,
|
| 502 |
+
"loss": 6.9259,
|
| 503 |
+
"step": 162
|
| 504 |
+
},
|
| 505 |
+
{
|
| 506 |
+
"epoch": 0.56,
|
| 507 |
+
"learning_rate": 4.103881278538813e-05,
|
| 508 |
+
"loss": 7.0744,
|
| 509 |
+
"step": 164
|
| 510 |
+
},
|
| 511 |
+
{
|
| 512 |
+
"epoch": 0.57,
|
| 513 |
+
"learning_rate": 4.092465753424658e-05,
|
| 514 |
+
"loss": 6.8039,
|
| 515 |
+
"step": 166
|
| 516 |
+
},
|
| 517 |
+
{
|
| 518 |
+
"epoch": 0.58,
|
| 519 |
+
"learning_rate": 4.0810502283105025e-05,
|
| 520 |
+
"loss": 6.8271,
|
| 521 |
+
"step": 168
|
| 522 |
+
},
|
| 523 |
+
{
|
| 524 |
+
"epoch": 0.58,
|
| 525 |
+
"learning_rate": 4.069634703196347e-05,
|
| 526 |
+
"loss": 6.7427,
|
| 527 |
+
"step": 170
|
| 528 |
+
},
|
| 529 |
+
{
|
| 530 |
+
"epoch": 0.59,
|
| 531 |
+
"learning_rate": 4.058219178082192e-05,
|
| 532 |
+
"loss": 6.5537,
|
| 533 |
+
"step": 172
|
| 534 |
+
},
|
| 535 |
+
{
|
| 536 |
+
"epoch": 0.6,
|
| 537 |
+
"learning_rate": 4.046803652968037e-05,
|
| 538 |
+
"loss": 6.6001,
|
| 539 |
+
"step": 174
|
| 540 |
+
},
|
| 541 |
+
{
|
| 542 |
+
"epoch": 0.6,
|
| 543 |
+
"learning_rate": 4.035388127853881e-05,
|
| 544 |
+
"loss": 6.6144,
|
| 545 |
+
"step": 176
|
| 546 |
+
},
|
| 547 |
+
{
|
| 548 |
+
"epoch": 0.61,
|
| 549 |
+
"learning_rate": 4.0239726027397265e-05,
|
| 550 |
+
"loss": 6.8667,
|
| 551 |
+
"step": 178
|
| 552 |
+
},
|
| 553 |
+
{
|
| 554 |
+
"epoch": 0.62,
|
| 555 |
+
"learning_rate": 4.012557077625571e-05,
|
| 556 |
+
"loss": 7.0148,
|
| 557 |
+
"step": 180
|
| 558 |
+
},
|
| 559 |
+
{
|
| 560 |
+
"epoch": 0.62,
|
| 561 |
+
"learning_rate": 4.001141552511416e-05,
|
| 562 |
+
"loss": 6.7102,
|
| 563 |
+
"step": 182
|
| 564 |
+
},
|
| 565 |
+
{
|
| 566 |
+
"epoch": 0.63,
|
| 567 |
+
"learning_rate": 3.98972602739726e-05,
|
| 568 |
+
"loss": 6.8212,
|
| 569 |
+
"step": 184
|
| 570 |
+
},
|
| 571 |
+
{
|
| 572 |
+
"epoch": 0.64,
|
| 573 |
+
"learning_rate": 3.978310502283105e-05,
|
| 574 |
+
"loss": 6.6284,
|
| 575 |
+
"step": 186
|
| 576 |
+
},
|
| 577 |
+
{
|
| 578 |
+
"epoch": 0.64,
|
| 579 |
+
"learning_rate": 3.96689497716895e-05,
|
| 580 |
+
"loss": 6.5087,
|
| 581 |
+
"step": 188
|
| 582 |
+
},
|
| 583 |
+
{
|
| 584 |
+
"epoch": 0.65,
|
| 585 |
+
"learning_rate": 3.9554794520547946e-05,
|
| 586 |
+
"loss": 7.1345,
|
| 587 |
+
"step": 190
|
| 588 |
+
},
|
| 589 |
+
{
|
| 590 |
+
"epoch": 0.66,
|
| 591 |
+
"learning_rate": 3.9440639269406394e-05,
|
| 592 |
+
"loss": 7.0863,
|
| 593 |
+
"step": 192
|
| 594 |
+
},
|
| 595 |
+
{
|
| 596 |
+
"epoch": 0.66,
|
| 597 |
+
"learning_rate": 3.932648401826484e-05,
|
| 598 |
+
"loss": 6.8418,
|
| 599 |
+
"step": 194
|
| 600 |
+
},
|
| 601 |
+
{
|
| 602 |
+
"epoch": 0.67,
|
| 603 |
+
"learning_rate": 3.921232876712329e-05,
|
| 604 |
+
"loss": 7.0545,
|
| 605 |
+
"step": 196
|
| 606 |
+
},
|
| 607 |
+
{
|
| 608 |
+
"epoch": 0.68,
|
| 609 |
+
"learning_rate": 3.909817351598174e-05,
|
| 610 |
+
"loss": 7.1057,
|
| 611 |
+
"step": 198
|
| 612 |
+
},
|
| 613 |
+
{
|
| 614 |
+
"epoch": 0.68,
|
| 615 |
+
"learning_rate": 3.8984018264840186e-05,
|
| 616 |
+
"loss": 6.6551,
|
| 617 |
+
"step": 200
|
| 618 |
+
},
|
| 619 |
+
{
|
| 620 |
+
"epoch": 0.68,
|
| 621 |
+
"eval_cer": 0.8836712913553896,
|
| 622 |
+
"eval_loss": 6.700944900512695,
|
| 623 |
+
"eval_runtime": 40.0981,
|
| 624 |
+
"eval_samples_per_second": 14.539,
|
| 625 |
+
"eval_steps_per_second": 1.821,
|
| 626 |
+
"step": 200
|
| 627 |
+
},
|
| 628 |
+
{
|
| 629 |
+
"epoch": 0.69,
|
| 630 |
+
"learning_rate": 3.8869863013698634e-05,
|
| 631 |
+
"loss": 6.7477,
|
| 632 |
+
"step": 202
|
| 633 |
+
},
|
| 634 |
+
{
|
| 635 |
+
"epoch": 0.7,
|
| 636 |
+
"learning_rate": 3.875570776255708e-05,
|
| 637 |
+
"loss": 6.9312,
|
| 638 |
+
"step": 204
|
| 639 |
+
},
|
| 640 |
+
{
|
| 641 |
+
"epoch": 0.71,
|
| 642 |
+
"learning_rate": 3.864155251141553e-05,
|
| 643 |
+
"loss": 6.6784,
|
| 644 |
+
"step": 206
|
| 645 |
+
},
|
| 646 |
+
{
|
| 647 |
+
"epoch": 0.71,
|
| 648 |
+
"learning_rate": 3.852739726027397e-05,
|
| 649 |
+
"loss": 6.9644,
|
| 650 |
+
"step": 208
|
| 651 |
+
},
|
| 652 |
+
{
|
| 653 |
+
"epoch": 0.72,
|
| 654 |
+
"learning_rate": 3.841324200913242e-05,
|
| 655 |
+
"loss": 6.5752,
|
| 656 |
+
"step": 210
|
| 657 |
+
},
|
| 658 |
+
{
|
| 659 |
+
"epoch": 0.73,
|
| 660 |
+
"learning_rate": 3.829908675799087e-05,
|
| 661 |
+
"loss": 7.0168,
|
| 662 |
+
"step": 212
|
| 663 |
+
},
|
| 664 |
+
{
|
| 665 |
+
"epoch": 0.73,
|
| 666 |
+
"learning_rate": 3.818493150684932e-05,
|
| 667 |
+
"loss": 6.7026,
|
| 668 |
+
"step": 214
|
| 669 |
+
},
|
| 670 |
+
{
|
| 671 |
+
"epoch": 0.74,
|
| 672 |
+
"learning_rate": 3.8070776255707764e-05,
|
| 673 |
+
"loss": 6.8185,
|
| 674 |
+
"step": 216
|
| 675 |
+
},
|
| 676 |
+
{
|
| 677 |
+
"epoch": 0.75,
|
| 678 |
+
"learning_rate": 3.795662100456621e-05,
|
| 679 |
+
"loss": 6.7238,
|
| 680 |
+
"step": 218
|
| 681 |
+
},
|
| 682 |
+
{
|
| 683 |
+
"epoch": 0.75,
|
| 684 |
+
"learning_rate": 3.784246575342466e-05,
|
| 685 |
+
"loss": 6.7133,
|
| 686 |
+
"step": 220
|
| 687 |
+
},
|
| 688 |
+
{
|
| 689 |
+
"epoch": 0.76,
|
| 690 |
+
"learning_rate": 3.772831050228311e-05,
|
| 691 |
+
"loss": 6.771,
|
| 692 |
+
"step": 222
|
| 693 |
+
},
|
| 694 |
+
{
|
| 695 |
+
"epoch": 0.77,
|
| 696 |
+
"learning_rate": 3.7614155251141555e-05,
|
| 697 |
+
"loss": 6.8149,
|
| 698 |
+
"step": 224
|
| 699 |
+
},
|
| 700 |
+
{
|
| 701 |
+
"epoch": 0.77,
|
| 702 |
+
"learning_rate": 3.7500000000000003e-05,
|
| 703 |
+
"loss": 6.6512,
|
| 704 |
+
"step": 226
|
| 705 |
+
},
|
| 706 |
+
{
|
| 707 |
+
"epoch": 0.78,
|
| 708 |
+
"learning_rate": 3.738584474885845e-05,
|
| 709 |
+
"loss": 6.213,
|
| 710 |
+
"step": 228
|
| 711 |
+
},
|
| 712 |
+
{
|
| 713 |
+
"epoch": 0.79,
|
| 714 |
+
"learning_rate": 3.727168949771689e-05,
|
| 715 |
+
"loss": 6.7743,
|
| 716 |
+
"step": 230
|
| 717 |
+
},
|
| 718 |
+
{
|
| 719 |
+
"epoch": 0.79,
|
| 720 |
+
"learning_rate": 3.715753424657534e-05,
|
| 721 |
+
"loss": 6.7843,
|
| 722 |
+
"step": 232
|
| 723 |
+
},
|
| 724 |
+
{
|
| 725 |
+
"epoch": 0.8,
|
| 726 |
+
"learning_rate": 3.704337899543379e-05,
|
| 727 |
+
"loss": 6.6052,
|
| 728 |
+
"step": 234
|
| 729 |
+
},
|
| 730 |
+
{
|
| 731 |
+
"epoch": 0.81,
|
| 732 |
+
"learning_rate": 3.6929223744292243e-05,
|
| 733 |
+
"loss": 6.5319,
|
| 734 |
+
"step": 236
|
| 735 |
+
},
|
| 736 |
+
{
|
| 737 |
+
"epoch": 0.82,
|
| 738 |
+
"learning_rate": 3.6815068493150685e-05,
|
| 739 |
+
"loss": 6.4869,
|
| 740 |
+
"step": 238
|
| 741 |
+
},
|
| 742 |
+
{
|
| 743 |
+
"epoch": 0.82,
|
| 744 |
+
"learning_rate": 3.670091324200913e-05,
|
| 745 |
+
"loss": 6.7245,
|
| 746 |
+
"step": 240
|
| 747 |
+
},
|
| 748 |
+
{
|
| 749 |
+
"epoch": 0.83,
|
| 750 |
+
"learning_rate": 3.658675799086758e-05,
|
| 751 |
+
"loss": 6.8325,
|
| 752 |
+
"step": 242
|
| 753 |
+
},
|
| 754 |
+
{
|
| 755 |
+
"epoch": 0.84,
|
| 756 |
+
"learning_rate": 3.647260273972603e-05,
|
| 757 |
+
"loss": 6.6019,
|
| 758 |
+
"step": 244
|
| 759 |
+
},
|
| 760 |
+
{
|
| 761 |
+
"epoch": 0.84,
|
| 762 |
+
"learning_rate": 3.635844748858448e-05,
|
| 763 |
+
"loss": 6.2614,
|
| 764 |
+
"step": 246
|
| 765 |
+
},
|
| 766 |
+
{
|
| 767 |
+
"epoch": 0.85,
|
| 768 |
+
"learning_rate": 3.6244292237442925e-05,
|
| 769 |
+
"loss": 7.0083,
|
| 770 |
+
"step": 248
|
| 771 |
+
},
|
| 772 |
+
{
|
| 773 |
+
"epoch": 0.86,
|
| 774 |
+
"learning_rate": 3.613013698630137e-05,
|
| 775 |
+
"loss": 6.4853,
|
| 776 |
+
"step": 250
|
| 777 |
+
},
|
| 778 |
+
{
|
| 779 |
+
"epoch": 0.86,
|
| 780 |
+
"learning_rate": 3.601598173515982e-05,
|
| 781 |
+
"loss": 6.7758,
|
| 782 |
+
"step": 252
|
| 783 |
+
},
|
| 784 |
+
{
|
| 785 |
+
"epoch": 0.87,
|
| 786 |
+
"learning_rate": 3.590182648401826e-05,
|
| 787 |
+
"loss": 6.7792,
|
| 788 |
+
"step": 254
|
| 789 |
+
},
|
| 790 |
+
{
|
| 791 |
+
"epoch": 0.88,
|
| 792 |
+
"learning_rate": 3.578767123287671e-05,
|
| 793 |
+
"loss": 6.6991,
|
| 794 |
+
"step": 256
|
| 795 |
+
},
|
| 796 |
+
{
|
| 797 |
+
"epoch": 0.88,
|
| 798 |
+
"learning_rate": 3.5673515981735165e-05,
|
| 799 |
+
"loss": 6.5657,
|
| 800 |
+
"step": 258
|
| 801 |
+
},
|
| 802 |
+
{
|
| 803 |
+
"epoch": 0.89,
|
| 804 |
+
"learning_rate": 3.555936073059361e-05,
|
| 805 |
+
"loss": 6.7322,
|
| 806 |
+
"step": 260
|
| 807 |
+
},
|
| 808 |
+
{
|
| 809 |
+
"epoch": 0.9,
|
| 810 |
+
"learning_rate": 3.5445205479452054e-05,
|
| 811 |
+
"loss": 6.4798,
|
| 812 |
+
"step": 262
|
| 813 |
+
},
|
| 814 |
+
{
|
| 815 |
+
"epoch": 0.9,
|
| 816 |
+
"learning_rate": 3.53310502283105e-05,
|
| 817 |
+
"loss": 6.7884,
|
| 818 |
+
"step": 264
|
| 819 |
+
},
|
| 820 |
+
{
|
| 821 |
+
"epoch": 0.91,
|
| 822 |
+
"learning_rate": 3.521689497716895e-05,
|
| 823 |
+
"loss": 6.424,
|
| 824 |
+
"step": 266
|
| 825 |
+
},
|
| 826 |
+
{
|
| 827 |
+
"epoch": 0.92,
|
| 828 |
+
"learning_rate": 3.51027397260274e-05,
|
| 829 |
+
"loss": 6.6925,
|
| 830 |
+
"step": 268
|
| 831 |
+
},
|
| 832 |
+
{
|
| 833 |
+
"epoch": 0.92,
|
| 834 |
+
"learning_rate": 3.4988584474885846e-05,
|
| 835 |
+
"loss": 7.0121,
|
| 836 |
+
"step": 270
|
| 837 |
+
},
|
| 838 |
+
{
|
| 839 |
+
"epoch": 0.93,
|
| 840 |
+
"learning_rate": 3.4874429223744294e-05,
|
| 841 |
+
"loss": 6.8289,
|
| 842 |
+
"step": 272
|
| 843 |
+
},
|
| 844 |
+
{
|
| 845 |
+
"epoch": 0.94,
|
| 846 |
+
"learning_rate": 3.476027397260274e-05,
|
| 847 |
+
"loss": 6.7504,
|
| 848 |
+
"step": 274
|
| 849 |
+
},
|
| 850 |
+
{
|
| 851 |
+
"epoch": 0.95,
|
| 852 |
+
"learning_rate": 3.464611872146119e-05,
|
| 853 |
+
"loss": 6.6898,
|
| 854 |
+
"step": 276
|
| 855 |
+
},
|
| 856 |
+
{
|
| 857 |
+
"epoch": 0.95,
|
| 858 |
+
"learning_rate": 3.453196347031963e-05,
|
| 859 |
+
"loss": 6.7379,
|
| 860 |
+
"step": 278
|
| 861 |
+
},
|
| 862 |
+
{
|
| 863 |
+
"epoch": 0.96,
|
| 864 |
+
"learning_rate": 3.4417808219178086e-05,
|
| 865 |
+
"loss": 6.4448,
|
| 866 |
+
"step": 280
|
| 867 |
+
},
|
| 868 |
+
{
|
| 869 |
+
"epoch": 0.97,
|
| 870 |
+
"learning_rate": 3.4303652968036534e-05,
|
| 871 |
+
"loss": 6.678,
|
| 872 |
+
"step": 282
|
| 873 |
+
},
|
| 874 |
+
{
|
| 875 |
+
"epoch": 0.97,
|
| 876 |
+
"learning_rate": 3.418949771689498e-05,
|
| 877 |
+
"loss": 6.4519,
|
| 878 |
+
"step": 284
|
| 879 |
+
},
|
| 880 |
+
{
|
| 881 |
+
"epoch": 0.98,
|
| 882 |
+
"learning_rate": 3.407534246575342e-05,
|
| 883 |
+
"loss": 6.5809,
|
| 884 |
+
"step": 286
|
| 885 |
+
},
|
| 886 |
+
{
|
| 887 |
+
"epoch": 0.99,
|
| 888 |
+
"learning_rate": 3.396118721461187e-05,
|
| 889 |
+
"loss": 6.5462,
|
| 890 |
+
"step": 288
|
| 891 |
+
},
|
| 892 |
+
{
|
| 893 |
+
"epoch": 0.99,
|
| 894 |
+
"learning_rate": 3.384703196347032e-05,
|
| 895 |
+
"loss": 6.4562,
|
| 896 |
+
"step": 290
|
| 897 |
+
},
|
| 898 |
+
{
|
| 899 |
+
"epoch": 1.0,
|
| 900 |
+
"learning_rate": 3.373287671232877e-05,
|
| 901 |
+
"loss": 6.5566,
|
| 902 |
+
"step": 292
|
| 903 |
+
},
|
| 904 |
+
{
|
| 905 |
+
"epoch": 1.01,
|
| 906 |
+
"learning_rate": 3.3618721461187215e-05,
|
| 907 |
+
"loss": 6.4929,
|
| 908 |
+
"step": 294
|
| 909 |
+
},
|
| 910 |
+
{
|
| 911 |
+
"epoch": 1.01,
|
| 912 |
+
"learning_rate": 3.350456621004566e-05,
|
| 913 |
+
"loss": 6.5228,
|
| 914 |
+
"step": 296
|
| 915 |
+
},
|
| 916 |
+
{
|
| 917 |
+
"epoch": 1.02,
|
| 918 |
+
"learning_rate": 3.339041095890411e-05,
|
| 919 |
+
"loss": 6.6275,
|
| 920 |
+
"step": 298
|
| 921 |
+
},
|
| 922 |
+
{
|
| 923 |
+
"epoch": 1.03,
|
| 924 |
+
"learning_rate": 3.327625570776256e-05,
|
| 925 |
+
"loss": 6.2089,
|
| 926 |
+
"step": 300
|
| 927 |
+
},
|
| 928 |
+
{
|
| 929 |
+
"epoch": 1.03,
|
| 930 |
+
"eval_cer": 0.7978408997619243,
|
| 931 |
+
"eval_loss": 6.618570327758789,
|
| 932 |
+
"eval_runtime": 34.5444,
|
| 933 |
+
"eval_samples_per_second": 16.877,
|
| 934 |
+
"eval_steps_per_second": 2.113,
|
| 935 |
+
"step": 300
|
| 936 |
+
},
|
| 937 |
+
{
|
| 938 |
+
"epoch": 1.03,
|
| 939 |
+
"learning_rate": 3.316210045662101e-05,
|
| 940 |
+
"loss": 6.3685,
|
| 941 |
+
"step": 302
|
| 942 |
+
},
|
| 943 |
+
{
|
| 944 |
+
"epoch": 1.04,
|
| 945 |
+
"learning_rate": 3.3047945205479455e-05,
|
| 946 |
+
"loss": 6.1731,
|
| 947 |
+
"step": 304
|
| 948 |
+
},
|
| 949 |
+
{
|
| 950 |
+
"epoch": 1.05,
|
| 951 |
+
"learning_rate": 3.29337899543379e-05,
|
| 952 |
+
"loss": 6.359,
|
| 953 |
+
"step": 306
|
| 954 |
+
},
|
| 955 |
+
{
|
| 956 |
+
"epoch": 1.05,
|
| 957 |
+
"learning_rate": 3.281963470319635e-05,
|
| 958 |
+
"loss": 6.2789,
|
| 959 |
+
"step": 308
|
| 960 |
+
},
|
| 961 |
+
{
|
| 962 |
+
"epoch": 1.06,
|
| 963 |
+
"learning_rate": 3.270547945205479e-05,
|
| 964 |
+
"loss": 6.2938,
|
| 965 |
+
"step": 310
|
| 966 |
+
},
|
| 967 |
+
{
|
| 968 |
+
"epoch": 1.07,
|
| 969 |
+
"learning_rate": 3.259132420091324e-05,
|
| 970 |
+
"loss": 6.3706,
|
| 971 |
+
"step": 312
|
| 972 |
+
},
|
| 973 |
+
{
|
| 974 |
+
"epoch": 1.08,
|
| 975 |
+
"learning_rate": 3.247716894977169e-05,
|
| 976 |
+
"loss": 6.4628,
|
| 977 |
+
"step": 314
|
| 978 |
+
},
|
| 979 |
+
{
|
| 980 |
+
"epoch": 1.08,
|
| 981 |
+
"learning_rate": 3.236301369863014e-05,
|
| 982 |
+
"loss": 5.9578,
|
| 983 |
+
"step": 316
|
| 984 |
+
},
|
| 985 |
+
{
|
| 986 |
+
"epoch": 1.09,
|
| 987 |
+
"learning_rate": 3.2248858447488584e-05,
|
| 988 |
+
"loss": 6.3769,
|
| 989 |
+
"step": 318
|
| 990 |
+
},
|
| 991 |
+
{
|
| 992 |
+
"epoch": 1.1,
|
| 993 |
+
"learning_rate": 3.213470319634703e-05,
|
| 994 |
+
"loss": 6.7954,
|
| 995 |
+
"step": 320
|
| 996 |
+
},
|
| 997 |
+
{
|
| 998 |
+
"epoch": 1.1,
|
| 999 |
+
"learning_rate": 3.202054794520548e-05,
|
| 1000 |
+
"loss": 6.7632,
|
| 1001 |
+
"step": 322
|
| 1002 |
+
},
|
| 1003 |
+
{
|
| 1004 |
+
"epoch": 1.11,
|
| 1005 |
+
"learning_rate": 3.190639269406393e-05,
|
| 1006 |
+
"loss": 6.3459,
|
| 1007 |
+
"step": 324
|
| 1008 |
+
},
|
| 1009 |
+
{
|
| 1010 |
+
"epoch": 1.12,
|
| 1011 |
+
"learning_rate": 3.1792237442922376e-05,
|
| 1012 |
+
"loss": 6.4654,
|
| 1013 |
+
"step": 326
|
| 1014 |
+
},
|
| 1015 |
+
{
|
| 1016 |
+
"epoch": 1.12,
|
| 1017 |
+
"learning_rate": 3.1678082191780824e-05,
|
| 1018 |
+
"loss": 6.6141,
|
| 1019 |
+
"step": 328
|
| 1020 |
+
},
|
| 1021 |
+
{
|
| 1022 |
+
"epoch": 1.13,
|
| 1023 |
+
"learning_rate": 3.156392694063927e-05,
|
| 1024 |
+
"loss": 6.2931,
|
| 1025 |
+
"step": 330
|
| 1026 |
+
},
|
| 1027 |
+
{
|
| 1028 |
+
"epoch": 1.14,
|
| 1029 |
+
"learning_rate": 3.144977168949772e-05,
|
| 1030 |
+
"loss": 5.9657,
|
| 1031 |
+
"step": 332
|
| 1032 |
+
},
|
| 1033 |
+
{
|
| 1034 |
+
"epoch": 1.14,
|
| 1035 |
+
"learning_rate": 3.133561643835616e-05,
|
| 1036 |
+
"loss": 6.6218,
|
| 1037 |
+
"step": 334
|
| 1038 |
+
},
|
| 1039 |
+
{
|
| 1040 |
+
"epoch": 1.15,
|
| 1041 |
+
"learning_rate": 3.122146118721461e-05,
|
| 1042 |
+
"loss": 6.2199,
|
| 1043 |
+
"step": 336
|
| 1044 |
+
},
|
| 1045 |
+
{
|
| 1046 |
+
"epoch": 1.16,
|
| 1047 |
+
"learning_rate": 3.1107305936073064e-05,
|
| 1048 |
+
"loss": 6.5139,
|
| 1049 |
+
"step": 338
|
| 1050 |
+
},
|
| 1051 |
+
{
|
| 1052 |
+
"epoch": 1.16,
|
| 1053 |
+
"learning_rate": 3.099315068493151e-05,
|
| 1054 |
+
"loss": 6.5002,
|
| 1055 |
+
"step": 340
|
| 1056 |
+
},
|
| 1057 |
+
{
|
| 1058 |
+
"epoch": 1.17,
|
| 1059 |
+
"learning_rate": 3.087899543378995e-05,
|
| 1060 |
+
"loss": 6.3457,
|
| 1061 |
+
"step": 342
|
| 1062 |
+
},
|
| 1063 |
+
{
|
| 1064 |
+
"epoch": 1.18,
|
| 1065 |
+
"learning_rate": 3.07648401826484e-05,
|
| 1066 |
+
"loss": 6.1393,
|
| 1067 |
+
"step": 344
|
| 1068 |
+
},
|
| 1069 |
+
{
|
| 1070 |
+
"epoch": 1.18,
|
| 1071 |
+
"learning_rate": 3.065068493150685e-05,
|
| 1072 |
+
"loss": 6.0003,
|
| 1073 |
+
"step": 346
|
| 1074 |
+
},
|
| 1075 |
+
{
|
| 1076 |
+
"epoch": 1.19,
|
| 1077 |
+
"learning_rate": 3.05365296803653e-05,
|
| 1078 |
+
"loss": 6.237,
|
| 1079 |
+
"step": 348
|
| 1080 |
+
},
|
| 1081 |
+
{
|
| 1082 |
+
"epoch": 1.2,
|
| 1083 |
+
"learning_rate": 3.0422374429223742e-05,
|
| 1084 |
+
"loss": 6.4162,
|
| 1085 |
+
"step": 350
|
| 1086 |
+
},
|
| 1087 |
+
{
|
| 1088 |
+
"epoch": 1.21,
|
| 1089 |
+
"learning_rate": 3.0308219178082193e-05,
|
| 1090 |
+
"loss": 6.4468,
|
| 1091 |
+
"step": 352
|
| 1092 |
+
},
|
| 1093 |
+
{
|
| 1094 |
+
"epoch": 1.21,
|
| 1095 |
+
"learning_rate": 3.019406392694064e-05,
|
| 1096 |
+
"loss": 6.6031,
|
| 1097 |
+
"step": 354
|
| 1098 |
+
},
|
| 1099 |
+
{
|
| 1100 |
+
"epoch": 1.22,
|
| 1101 |
+
"learning_rate": 3.007990867579909e-05,
|
| 1102 |
+
"loss": 6.3145,
|
| 1103 |
+
"step": 356
|
| 1104 |
+
},
|
| 1105 |
+
{
|
| 1106 |
+
"epoch": 1.23,
|
| 1107 |
+
"learning_rate": 2.9965753424657534e-05,
|
| 1108 |
+
"loss": 6.3375,
|
| 1109 |
+
"step": 358
|
| 1110 |
+
},
|
| 1111 |
+
{
|
| 1112 |
+
"epoch": 1.23,
|
| 1113 |
+
"learning_rate": 2.9851598173515982e-05,
|
| 1114 |
+
"loss": 6.4769,
|
| 1115 |
+
"step": 360
|
| 1116 |
+
},
|
| 1117 |
+
{
|
| 1118 |
+
"epoch": 1.24,
|
| 1119 |
+
"learning_rate": 2.973744292237443e-05,
|
| 1120 |
+
"loss": 6.0631,
|
| 1121 |
+
"step": 362
|
| 1122 |
+
},
|
| 1123 |
+
{
|
| 1124 |
+
"epoch": 1.25,
|
| 1125 |
+
"learning_rate": 2.962328767123288e-05,
|
| 1126 |
+
"loss": 6.2347,
|
| 1127 |
+
"step": 364
|
| 1128 |
+
},
|
| 1129 |
+
{
|
| 1130 |
+
"epoch": 1.25,
|
| 1131 |
+
"learning_rate": 2.9509132420091322e-05,
|
| 1132 |
+
"loss": 6.0945,
|
| 1133 |
+
"step": 366
|
| 1134 |
+
},
|
| 1135 |
+
{
|
| 1136 |
+
"epoch": 1.26,
|
| 1137 |
+
"learning_rate": 2.9394977168949774e-05,
|
| 1138 |
+
"loss": 6.2858,
|
| 1139 |
+
"step": 368
|
| 1140 |
+
},
|
| 1141 |
+
{
|
| 1142 |
+
"epoch": 1.27,
|
| 1143 |
+
"learning_rate": 2.9280821917808222e-05,
|
| 1144 |
+
"loss": 6.3601,
|
| 1145 |
+
"step": 370
|
| 1146 |
+
},
|
| 1147 |
+
{
|
| 1148 |
+
"epoch": 1.27,
|
| 1149 |
+
"learning_rate": 2.916666666666667e-05,
|
| 1150 |
+
"loss": 6.3449,
|
| 1151 |
+
"step": 372
|
| 1152 |
+
},
|
| 1153 |
+
{
|
| 1154 |
+
"epoch": 1.28,
|
| 1155 |
+
"learning_rate": 2.9052511415525114e-05,
|
| 1156 |
+
"loss": 6.3585,
|
| 1157 |
+
"step": 374
|
| 1158 |
+
},
|
| 1159 |
+
{
|
| 1160 |
+
"epoch": 1.29,
|
| 1161 |
+
"learning_rate": 2.8938356164383562e-05,
|
| 1162 |
+
"loss": 6.0235,
|
| 1163 |
+
"step": 376
|
| 1164 |
+
},
|
| 1165 |
+
{
|
| 1166 |
+
"epoch": 1.29,
|
| 1167 |
+
"learning_rate": 2.882420091324201e-05,
|
| 1168 |
+
"loss": 6.1927,
|
| 1169 |
+
"step": 378
|
| 1170 |
+
},
|
| 1171 |
+
{
|
| 1172 |
+
"epoch": 1.3,
|
| 1173 |
+
"learning_rate": 2.871004566210046e-05,
|
| 1174 |
+
"loss": 6.4136,
|
| 1175 |
+
"step": 380
|
| 1176 |
+
},
|
| 1177 |
+
{
|
| 1178 |
+
"epoch": 1.31,
|
| 1179 |
+
"learning_rate": 2.8595890410958903e-05,
|
| 1180 |
+
"loss": 6.4073,
|
| 1181 |
+
"step": 382
|
| 1182 |
+
},
|
| 1183 |
+
{
|
| 1184 |
+
"epoch": 1.32,
|
| 1185 |
+
"learning_rate": 2.848173515981735e-05,
|
| 1186 |
+
"loss": 6.4599,
|
| 1187 |
+
"step": 384
|
| 1188 |
+
},
|
| 1189 |
+
{
|
| 1190 |
+
"epoch": 1.32,
|
| 1191 |
+
"learning_rate": 2.8367579908675802e-05,
|
| 1192 |
+
"loss": 6.1341,
|
| 1193 |
+
"step": 386
|
| 1194 |
+
},
|
| 1195 |
+
{
|
| 1196 |
+
"epoch": 1.33,
|
| 1197 |
+
"learning_rate": 2.825342465753425e-05,
|
| 1198 |
+
"loss": 6.0782,
|
| 1199 |
+
"step": 388
|
| 1200 |
+
},
|
| 1201 |
+
{
|
| 1202 |
+
"epoch": 1.34,
|
| 1203 |
+
"learning_rate": 2.813926940639269e-05,
|
| 1204 |
+
"loss": 6.1504,
|
| 1205 |
+
"step": 390
|
| 1206 |
+
},
|
| 1207 |
+
{
|
| 1208 |
+
"epoch": 1.34,
|
| 1209 |
+
"learning_rate": 2.8025114155251143e-05,
|
| 1210 |
+
"loss": 6.3133,
|
| 1211 |
+
"step": 392
|
| 1212 |
+
},
|
| 1213 |
+
{
|
| 1214 |
+
"epoch": 1.35,
|
| 1215 |
+
"learning_rate": 2.791095890410959e-05,
|
| 1216 |
+
"loss": 6.4356,
|
| 1217 |
+
"step": 394
|
| 1218 |
+
},
|
| 1219 |
+
{
|
| 1220 |
+
"epoch": 1.36,
|
| 1221 |
+
"learning_rate": 2.779680365296804e-05,
|
| 1222 |
+
"loss": 6.4857,
|
| 1223 |
+
"step": 396
|
| 1224 |
+
},
|
| 1225 |
+
{
|
| 1226 |
+
"epoch": 1.36,
|
| 1227 |
+
"learning_rate": 2.7682648401826484e-05,
|
| 1228 |
+
"loss": 6.4011,
|
| 1229 |
+
"step": 398
|
| 1230 |
+
},
|
| 1231 |
+
{
|
| 1232 |
+
"epoch": 1.37,
|
| 1233 |
+
"learning_rate": 2.756849315068493e-05,
|
| 1234 |
+
"loss": 6.1637,
|
| 1235 |
+
"step": 400
|
| 1236 |
+
},
|
| 1237 |
+
{
|
| 1238 |
+
"epoch": 1.37,
|
| 1239 |
+
"eval_cer": 0.9765208110992529,
|
| 1240 |
+
"eval_loss": 6.528327941894531,
|
| 1241 |
+
"eval_runtime": 37.1998,
|
| 1242 |
+
"eval_samples_per_second": 15.672,
|
| 1243 |
+
"eval_steps_per_second": 1.962,
|
| 1244 |
+
"step": 400
|
| 1245 |
+
},
|
| 1246 |
+
{
|
| 1247 |
+
"epoch": 1.38,
|
| 1248 |
+
"learning_rate": 2.745433789954338e-05,
|
| 1249 |
+
"loss": 6.3984,
|
| 1250 |
+
"step": 402
|
| 1251 |
+
},
|
| 1252 |
+
{
|
| 1253 |
+
"epoch": 1.38,
|
| 1254 |
+
"learning_rate": 2.734018264840183e-05,
|
| 1255 |
+
"loss": 6.0221,
|
| 1256 |
+
"step": 404
|
| 1257 |
+
},
|
| 1258 |
+
{
|
| 1259 |
+
"epoch": 1.39,
|
| 1260 |
+
"learning_rate": 2.7226027397260272e-05,
|
| 1261 |
+
"loss": 6.2472,
|
| 1262 |
+
"step": 406
|
| 1263 |
+
},
|
| 1264 |
+
{
|
| 1265 |
+
"epoch": 1.4,
|
| 1266 |
+
"learning_rate": 2.7111872146118723e-05,
|
| 1267 |
+
"loss": 6.333,
|
| 1268 |
+
"step": 408
|
| 1269 |
+
},
|
| 1270 |
+
{
|
| 1271 |
+
"epoch": 1.4,
|
| 1272 |
+
"learning_rate": 2.699771689497717e-05,
|
| 1273 |
+
"loss": 6.2516,
|
| 1274 |
+
"step": 410
|
| 1275 |
+
},
|
| 1276 |
+
{
|
| 1277 |
+
"epoch": 1.41,
|
| 1278 |
+
"learning_rate": 2.688356164383562e-05,
|
| 1279 |
+
"loss": 6.2314,
|
| 1280 |
+
"step": 412
|
| 1281 |
+
},
|
| 1282 |
+
{
|
| 1283 |
+
"epoch": 1.42,
|
| 1284 |
+
"learning_rate": 2.6769406392694064e-05,
|
| 1285 |
+
"loss": 6.1203,
|
| 1286 |
+
"step": 414
|
| 1287 |
+
},
|
| 1288 |
+
{
|
| 1289 |
+
"epoch": 1.42,
|
| 1290 |
+
"learning_rate": 2.6655251141552512e-05,
|
| 1291 |
+
"loss": 6.311,
|
| 1292 |
+
"step": 416
|
| 1293 |
+
},
|
| 1294 |
+
{
|
| 1295 |
+
"epoch": 1.43,
|
| 1296 |
+
"learning_rate": 2.654109589041096e-05,
|
| 1297 |
+
"loss": 6.1741,
|
| 1298 |
+
"step": 418
|
| 1299 |
+
},
|
| 1300 |
+
{
|
| 1301 |
+
"epoch": 1.44,
|
| 1302 |
+
"learning_rate": 2.6426940639269408e-05,
|
| 1303 |
+
"loss": 6.0226,
|
| 1304 |
+
"step": 420
|
| 1305 |
+
},
|
| 1306 |
+
{
|
| 1307 |
+
"epoch": 1.45,
|
| 1308 |
+
"learning_rate": 2.6312785388127853e-05,
|
| 1309 |
+
"loss": 6.1573,
|
| 1310 |
+
"step": 422
|
| 1311 |
+
},
|
| 1312 |
+
{
|
| 1313 |
+
"epoch": 1.45,
|
| 1314 |
+
"learning_rate": 2.61986301369863e-05,
|
| 1315 |
+
"loss": 6.3298,
|
| 1316 |
+
"step": 424
|
| 1317 |
+
},
|
| 1318 |
+
{
|
| 1319 |
+
"epoch": 1.46,
|
| 1320 |
+
"learning_rate": 2.6084474885844752e-05,
|
| 1321 |
+
"loss": 6.3197,
|
| 1322 |
+
"step": 426
|
| 1323 |
+
},
|
| 1324 |
+
{
|
| 1325 |
+
"epoch": 1.47,
|
| 1326 |
+
"learning_rate": 2.59703196347032e-05,
|
| 1327 |
+
"loss": 6.3265,
|
| 1328 |
+
"step": 428
|
| 1329 |
+
},
|
| 1330 |
+
{
|
| 1331 |
+
"epoch": 1.47,
|
| 1332 |
+
"learning_rate": 2.5856164383561645e-05,
|
| 1333 |
+
"loss": 6.5904,
|
| 1334 |
+
"step": 430
|
| 1335 |
+
},
|
| 1336 |
+
{
|
| 1337 |
+
"epoch": 1.48,
|
| 1338 |
+
"learning_rate": 2.5742009132420093e-05,
|
| 1339 |
+
"loss": 6.0295,
|
| 1340 |
+
"step": 432
|
| 1341 |
+
},
|
| 1342 |
+
{
|
| 1343 |
+
"epoch": 1.49,
|
| 1344 |
+
"learning_rate": 2.562785388127854e-05,
|
| 1345 |
+
"loss": 6.2887,
|
| 1346 |
+
"step": 434
|
| 1347 |
+
},
|
| 1348 |
+
{
|
| 1349 |
+
"epoch": 1.49,
|
| 1350 |
+
"learning_rate": 2.551369863013699e-05,
|
| 1351 |
+
"loss": 6.2315,
|
| 1352 |
+
"step": 436
|
| 1353 |
+
},
|
| 1354 |
+
{
|
| 1355 |
+
"epoch": 1.5,
|
| 1356 |
+
"learning_rate": 2.5399543378995433e-05,
|
| 1357 |
+
"loss": 6.1131,
|
| 1358 |
+
"step": 438
|
| 1359 |
+
},
|
| 1360 |
+
{
|
| 1361 |
+
"epoch": 1.51,
|
| 1362 |
+
"learning_rate": 2.528538812785388e-05,
|
| 1363 |
+
"loss": 6.0897,
|
| 1364 |
+
"step": 440
|
| 1365 |
+
},
|
| 1366 |
+
{
|
| 1367 |
+
"epoch": 1.51,
|
| 1368 |
+
"learning_rate": 2.517123287671233e-05,
|
| 1369 |
+
"loss": 6.2543,
|
| 1370 |
+
"step": 442
|
| 1371 |
+
},
|
| 1372 |
+
{
|
| 1373 |
+
"epoch": 1.52,
|
| 1374 |
+
"learning_rate": 2.505707762557078e-05,
|
| 1375 |
+
"loss": 6.2021,
|
| 1376 |
+
"step": 444
|
| 1377 |
+
},
|
| 1378 |
+
{
|
| 1379 |
+
"epoch": 1.53,
|
| 1380 |
+
"learning_rate": 2.4942922374429225e-05,
|
| 1381 |
+
"loss": 6.2414,
|
| 1382 |
+
"step": 446
|
| 1383 |
+
},
|
| 1384 |
+
{
|
| 1385 |
+
"epoch": 1.53,
|
| 1386 |
+
"learning_rate": 2.4828767123287673e-05,
|
| 1387 |
+
"loss": 6.3754,
|
| 1388 |
+
"step": 448
|
| 1389 |
+
},
|
| 1390 |
+
{
|
| 1391 |
+
"epoch": 1.54,
|
| 1392 |
+
"learning_rate": 2.471461187214612e-05,
|
| 1393 |
+
"loss": 6.4199,
|
| 1394 |
+
"step": 450
|
| 1395 |
+
},
|
| 1396 |
+
{
|
| 1397 |
+
"epoch": 1.55,
|
| 1398 |
+
"learning_rate": 2.4600456621004566e-05,
|
| 1399 |
+
"loss": 6.1184,
|
| 1400 |
+
"step": 452
|
| 1401 |
+
},
|
| 1402 |
+
{
|
| 1403 |
+
"epoch": 1.55,
|
| 1404 |
+
"learning_rate": 2.4486301369863017e-05,
|
| 1405 |
+
"loss": 6.308,
|
| 1406 |
+
"step": 454
|
| 1407 |
+
},
|
| 1408 |
+
{
|
| 1409 |
+
"epoch": 1.56,
|
| 1410 |
+
"learning_rate": 2.4372146118721462e-05,
|
| 1411 |
+
"loss": 6.4365,
|
| 1412 |
+
"step": 456
|
| 1413 |
+
},
|
| 1414 |
+
{
|
| 1415 |
+
"epoch": 1.57,
|
| 1416 |
+
"learning_rate": 2.425799086757991e-05,
|
| 1417 |
+
"loss": 6.0473,
|
| 1418 |
+
"step": 458
|
| 1419 |
+
},
|
| 1420 |
+
{
|
| 1421 |
+
"epoch": 1.58,
|
| 1422 |
+
"learning_rate": 2.4143835616438358e-05,
|
| 1423 |
+
"loss": 6.055,
|
| 1424 |
+
"step": 460
|
| 1425 |
+
},
|
| 1426 |
+
{
|
| 1427 |
+
"epoch": 1.58,
|
| 1428 |
+
"learning_rate": 2.4029680365296806e-05,
|
| 1429 |
+
"loss": 6.0705,
|
| 1430 |
+
"step": 462
|
| 1431 |
+
},
|
| 1432 |
+
{
|
| 1433 |
+
"epoch": 1.59,
|
| 1434 |
+
"learning_rate": 2.391552511415525e-05,
|
| 1435 |
+
"loss": 6.0425,
|
| 1436 |
+
"step": 464
|
| 1437 |
+
},
|
| 1438 |
+
{
|
| 1439 |
+
"epoch": 1.6,
|
| 1440 |
+
"learning_rate": 2.3801369863013702e-05,
|
| 1441 |
+
"loss": 5.9382,
|
| 1442 |
+
"step": 466
|
| 1443 |
+
},
|
| 1444 |
+
{
|
| 1445 |
+
"epoch": 1.6,
|
| 1446 |
+
"learning_rate": 2.3687214611872146e-05,
|
| 1447 |
+
"loss": 6.0501,
|
| 1448 |
+
"step": 468
|
| 1449 |
+
},
|
| 1450 |
+
{
|
| 1451 |
+
"epoch": 1.61,
|
| 1452 |
+
"learning_rate": 2.3573059360730594e-05,
|
| 1453 |
+
"loss": 6.1394,
|
| 1454 |
+
"step": 470
|
| 1455 |
+
},
|
| 1456 |
+
{
|
| 1457 |
+
"epoch": 1.62,
|
| 1458 |
+
"learning_rate": 2.3458904109589042e-05,
|
| 1459 |
+
"loss": 6.4409,
|
| 1460 |
+
"step": 472
|
| 1461 |
+
},
|
| 1462 |
+
{
|
| 1463 |
+
"epoch": 1.62,
|
| 1464 |
+
"learning_rate": 2.334474885844749e-05,
|
| 1465 |
+
"loss": 6.1642,
|
| 1466 |
+
"step": 474
|
| 1467 |
+
},
|
| 1468 |
+
{
|
| 1469 |
+
"epoch": 1.63,
|
| 1470 |
+
"learning_rate": 2.3230593607305935e-05,
|
| 1471 |
+
"loss": 6.1099,
|
| 1472 |
+
"step": 476
|
| 1473 |
+
},
|
| 1474 |
+
{
|
| 1475 |
+
"epoch": 1.64,
|
| 1476 |
+
"learning_rate": 2.3116438356164386e-05,
|
| 1477 |
+
"loss": 6.1933,
|
| 1478 |
+
"step": 478
|
| 1479 |
+
},
|
| 1480 |
+
{
|
| 1481 |
+
"epoch": 1.64,
|
| 1482 |
+
"learning_rate": 2.300228310502283e-05,
|
| 1483 |
+
"loss": 6.2047,
|
| 1484 |
+
"step": 480
|
| 1485 |
+
},
|
| 1486 |
+
{
|
| 1487 |
+
"epoch": 1.65,
|
| 1488 |
+
"learning_rate": 2.288812785388128e-05,
|
| 1489 |
+
"loss": 6.0044,
|
| 1490 |
+
"step": 482
|
| 1491 |
+
},
|
| 1492 |
+
{
|
| 1493 |
+
"epoch": 1.66,
|
| 1494 |
+
"learning_rate": 2.2773972602739727e-05,
|
| 1495 |
+
"loss": 6.0892,
|
| 1496 |
+
"step": 484
|
| 1497 |
+
},
|
| 1498 |
+
{
|
| 1499 |
+
"epoch": 1.66,
|
| 1500 |
+
"learning_rate": 2.2659817351598175e-05,
|
| 1501 |
+
"loss": 6.245,
|
| 1502 |
+
"step": 486
|
| 1503 |
+
},
|
| 1504 |
+
{
|
| 1505 |
+
"epoch": 1.67,
|
| 1506 |
+
"learning_rate": 2.2545662100456623e-05,
|
| 1507 |
+
"loss": 6.1838,
|
| 1508 |
+
"step": 488
|
| 1509 |
+
},
|
| 1510 |
+
{
|
| 1511 |
+
"epoch": 1.68,
|
| 1512 |
+
"learning_rate": 2.243150684931507e-05,
|
| 1513 |
+
"loss": 5.7336,
|
| 1514 |
+
"step": 490
|
| 1515 |
+
},
|
| 1516 |
+
{
|
| 1517 |
+
"epoch": 1.68,
|
| 1518 |
+
"learning_rate": 2.2317351598173515e-05,
|
| 1519 |
+
"loss": 6.1649,
|
| 1520 |
+
"step": 492
|
| 1521 |
+
},
|
| 1522 |
+
{
|
| 1523 |
+
"epoch": 1.69,
|
| 1524 |
+
"learning_rate": 2.2203196347031967e-05,
|
| 1525 |
+
"loss": 6.001,
|
| 1526 |
+
"step": 494
|
| 1527 |
+
},
|
| 1528 |
+
{
|
| 1529 |
+
"epoch": 1.7,
|
| 1530 |
+
"learning_rate": 2.208904109589041e-05,
|
| 1531 |
+
"loss": 6.6607,
|
| 1532 |
+
"step": 496
|
| 1533 |
+
},
|
| 1534 |
+
{
|
| 1535 |
+
"epoch": 1.71,
|
| 1536 |
+
"learning_rate": 2.197488584474886e-05,
|
| 1537 |
+
"loss": 5.9693,
|
| 1538 |
+
"step": 498
|
| 1539 |
+
},
|
| 1540 |
+
{
|
| 1541 |
+
"epoch": 1.71,
|
| 1542 |
+
"learning_rate": 2.1860730593607307e-05,
|
| 1543 |
+
"loss": 6.2159,
|
| 1544 |
+
"step": 500
|
| 1545 |
+
},
|
| 1546 |
+
{
|
| 1547 |
+
"epoch": 1.71,
|
| 1548 |
+
"eval_cer": 0.8505459321894754,
|
| 1549 |
+
"eval_loss": 6.468786716461182,
|
| 1550 |
+
"eval_runtime": 38.2571,
|
| 1551 |
+
"eval_samples_per_second": 15.239,
|
| 1552 |
+
"eval_steps_per_second": 1.908,
|
| 1553 |
+
"step": 500
|
| 1554 |
+
},
|
| 1555 |
+
{
|
| 1556 |
+
"epoch": 1.72,
|
| 1557 |
+
"learning_rate": 2.1746575342465755e-05,
|
| 1558 |
+
"loss": 6.5233,
|
| 1559 |
+
"step": 502
|
| 1560 |
+
},
|
| 1561 |
+
{
|
| 1562 |
+
"epoch": 1.73,
|
| 1563 |
+
"learning_rate": 2.16324200913242e-05,
|
| 1564 |
+
"loss": 6.0243,
|
| 1565 |
+
"step": 504
|
| 1566 |
+
},
|
| 1567 |
+
{
|
| 1568 |
+
"epoch": 1.73,
|
| 1569 |
+
"learning_rate": 2.151826484018265e-05,
|
| 1570 |
+
"loss": 6.0025,
|
| 1571 |
+
"step": 506
|
| 1572 |
+
},
|
| 1573 |
+
{
|
| 1574 |
+
"epoch": 1.74,
|
| 1575 |
+
"learning_rate": 2.1404109589041096e-05,
|
| 1576 |
+
"loss": 6.19,
|
| 1577 |
+
"step": 508
|
| 1578 |
+
},
|
| 1579 |
+
{
|
| 1580 |
+
"epoch": 1.75,
|
| 1581 |
+
"learning_rate": 2.1289954337899544e-05,
|
| 1582 |
+
"loss": 6.1283,
|
| 1583 |
+
"step": 510
|
| 1584 |
+
},
|
| 1585 |
+
{
|
| 1586 |
+
"epoch": 1.75,
|
| 1587 |
+
"learning_rate": 2.1175799086757992e-05,
|
| 1588 |
+
"loss": 5.9461,
|
| 1589 |
+
"step": 512
|
| 1590 |
+
},
|
| 1591 |
+
{
|
| 1592 |
+
"epoch": 1.76,
|
| 1593 |
+
"learning_rate": 2.106164383561644e-05,
|
| 1594 |
+
"loss": 6.163,
|
| 1595 |
+
"step": 514
|
| 1596 |
+
},
|
| 1597 |
+
{
|
| 1598 |
+
"epoch": 1.77,
|
| 1599 |
+
"learning_rate": 2.0947488584474888e-05,
|
| 1600 |
+
"loss": 5.9075,
|
| 1601 |
+
"step": 516
|
| 1602 |
+
},
|
| 1603 |
+
{
|
| 1604 |
+
"epoch": 1.77,
|
| 1605 |
+
"learning_rate": 2.0833333333333336e-05,
|
| 1606 |
+
"loss": 6.061,
|
| 1607 |
+
"step": 518
|
| 1608 |
+
},
|
| 1609 |
+
{
|
| 1610 |
+
"epoch": 1.78,
|
| 1611 |
+
"learning_rate": 2.071917808219178e-05,
|
| 1612 |
+
"loss": 5.8001,
|
| 1613 |
+
"step": 520
|
| 1614 |
+
},
|
| 1615 |
+
{
|
| 1616 |
+
"epoch": 1.79,
|
| 1617 |
+
"learning_rate": 2.060502283105023e-05,
|
| 1618 |
+
"loss": 6.3543,
|
| 1619 |
+
"step": 522
|
| 1620 |
+
},
|
| 1621 |
+
{
|
| 1622 |
+
"epoch": 1.79,
|
| 1623 |
+
"learning_rate": 2.0490867579908677e-05,
|
| 1624 |
+
"loss": 6.1044,
|
| 1625 |
+
"step": 524
|
| 1626 |
+
},
|
| 1627 |
+
{
|
| 1628 |
+
"epoch": 1.8,
|
| 1629 |
+
"learning_rate": 2.0376712328767125e-05,
|
| 1630 |
+
"loss": 6.0392,
|
| 1631 |
+
"step": 526
|
| 1632 |
+
},
|
| 1633 |
+
{
|
| 1634 |
+
"epoch": 1.81,
|
| 1635 |
+
"learning_rate": 2.0262557077625573e-05,
|
| 1636 |
+
"loss": 6.3613,
|
| 1637 |
+
"step": 528
|
| 1638 |
+
},
|
| 1639 |
+
{
|
| 1640 |
+
"epoch": 1.82,
|
| 1641 |
+
"learning_rate": 2.014840182648402e-05,
|
| 1642 |
+
"loss": 6.1969,
|
| 1643 |
+
"step": 530
|
| 1644 |
+
},
|
| 1645 |
+
{
|
| 1646 |
+
"epoch": 1.82,
|
| 1647 |
+
"learning_rate": 2.0034246575342465e-05,
|
| 1648 |
+
"loss": 6.2186,
|
| 1649 |
+
"step": 532
|
| 1650 |
+
},
|
| 1651 |
+
{
|
| 1652 |
+
"epoch": 1.83,
|
| 1653 |
+
"learning_rate": 1.9920091324200917e-05,
|
| 1654 |
+
"loss": 6.5516,
|
| 1655 |
+
"step": 534
|
| 1656 |
+
},
|
| 1657 |
+
{
|
| 1658 |
+
"epoch": 1.84,
|
| 1659 |
+
"learning_rate": 1.980593607305936e-05,
|
| 1660 |
+
"loss": 6.2319,
|
| 1661 |
+
"step": 536
|
| 1662 |
+
},
|
| 1663 |
+
{
|
| 1664 |
+
"epoch": 1.84,
|
| 1665 |
+
"learning_rate": 1.969178082191781e-05,
|
| 1666 |
+
"loss": 6.1139,
|
| 1667 |
+
"step": 538
|
| 1668 |
+
},
|
| 1669 |
+
{
|
| 1670 |
+
"epoch": 1.85,
|
| 1671 |
+
"learning_rate": 1.9577625570776257e-05,
|
| 1672 |
+
"loss": 6.6036,
|
| 1673 |
+
"step": 540
|
| 1674 |
+
},
|
| 1675 |
+
{
|
| 1676 |
+
"epoch": 1.86,
|
| 1677 |
+
"learning_rate": 1.9463470319634705e-05,
|
| 1678 |
+
"loss": 6.3099,
|
| 1679 |
+
"step": 542
|
| 1680 |
+
},
|
| 1681 |
+
{
|
| 1682 |
+
"epoch": 1.86,
|
| 1683 |
+
"learning_rate": 1.934931506849315e-05,
|
| 1684 |
+
"loss": 6.2986,
|
| 1685 |
+
"step": 544
|
| 1686 |
+
},
|
| 1687 |
+
{
|
| 1688 |
+
"epoch": 1.87,
|
| 1689 |
+
"learning_rate": 1.92351598173516e-05,
|
| 1690 |
+
"loss": 6.0614,
|
| 1691 |
+
"step": 546
|
| 1692 |
+
},
|
| 1693 |
+
{
|
| 1694 |
+
"epoch": 1.88,
|
| 1695 |
+
"learning_rate": 1.9121004566210046e-05,
|
| 1696 |
+
"loss": 5.8707,
|
| 1697 |
+
"step": 548
|
| 1698 |
+
},
|
| 1699 |
+
{
|
| 1700 |
+
"epoch": 1.88,
|
| 1701 |
+
"learning_rate": 1.9006849315068494e-05,
|
| 1702 |
+
"loss": 6.4741,
|
| 1703 |
+
"step": 550
|
| 1704 |
+
},
|
| 1705 |
+
{
|
| 1706 |
+
"epoch": 1.89,
|
| 1707 |
+
"learning_rate": 1.8892694063926942e-05,
|
| 1708 |
+
"loss": 5.9351,
|
| 1709 |
+
"step": 552
|
| 1710 |
+
},
|
| 1711 |
+
{
|
| 1712 |
+
"epoch": 1.9,
|
| 1713 |
+
"learning_rate": 1.877853881278539e-05,
|
| 1714 |
+
"loss": 6.3108,
|
| 1715 |
+
"step": 554
|
| 1716 |
+
},
|
| 1717 |
+
{
|
| 1718 |
+
"epoch": 1.9,
|
| 1719 |
+
"learning_rate": 1.8664383561643838e-05,
|
| 1720 |
+
"loss": 5.8719,
|
| 1721 |
+
"step": 556
|
| 1722 |
+
},
|
| 1723 |
+
{
|
| 1724 |
+
"epoch": 1.91,
|
| 1725 |
+
"learning_rate": 1.8550228310502286e-05,
|
| 1726 |
+
"loss": 5.9978,
|
| 1727 |
+
"step": 558
|
| 1728 |
+
},
|
| 1729 |
+
{
|
| 1730 |
+
"epoch": 1.92,
|
| 1731 |
+
"learning_rate": 1.843607305936073e-05,
|
| 1732 |
+
"loss": 5.9883,
|
| 1733 |
+
"step": 560
|
| 1734 |
+
},
|
| 1735 |
+
{
|
| 1736 |
+
"epoch": 1.92,
|
| 1737 |
+
"learning_rate": 1.832191780821918e-05,
|
| 1738 |
+
"loss": 5.7932,
|
| 1739 |
+
"step": 562
|
| 1740 |
+
},
|
| 1741 |
+
{
|
| 1742 |
+
"epoch": 1.93,
|
| 1743 |
+
"learning_rate": 1.8207762557077626e-05,
|
| 1744 |
+
"loss": 6.1634,
|
| 1745 |
+
"step": 564
|
| 1746 |
+
},
|
| 1747 |
+
{
|
| 1748 |
+
"epoch": 1.94,
|
| 1749 |
+
"learning_rate": 1.8093607305936074e-05,
|
| 1750 |
+
"loss": 6.2713,
|
| 1751 |
+
"step": 566
|
| 1752 |
+
},
|
| 1753 |
+
{
|
| 1754 |
+
"epoch": 1.95,
|
| 1755 |
+
"learning_rate": 1.7979452054794522e-05,
|
| 1756 |
+
"loss": 6.2679,
|
| 1757 |
+
"step": 568
|
| 1758 |
+
},
|
| 1759 |
+
{
|
| 1760 |
+
"epoch": 1.95,
|
| 1761 |
+
"learning_rate": 1.786529680365297e-05,
|
| 1762 |
+
"loss": 6.2969,
|
| 1763 |
+
"step": 570
|
| 1764 |
+
},
|
| 1765 |
+
{
|
| 1766 |
+
"epoch": 1.96,
|
| 1767 |
+
"learning_rate": 1.7751141552511415e-05,
|
| 1768 |
+
"loss": 6.0192,
|
| 1769 |
+
"step": 572
|
| 1770 |
+
},
|
| 1771 |
+
{
|
| 1772 |
+
"epoch": 1.97,
|
| 1773 |
+
"learning_rate": 1.7636986301369866e-05,
|
| 1774 |
+
"loss": 5.9912,
|
| 1775 |
+
"step": 574
|
| 1776 |
+
},
|
| 1777 |
+
{
|
| 1778 |
+
"epoch": 1.97,
|
| 1779 |
+
"learning_rate": 1.752283105022831e-05,
|
| 1780 |
+
"loss": 6.2804,
|
| 1781 |
+
"step": 576
|
| 1782 |
+
},
|
| 1783 |
+
{
|
| 1784 |
+
"epoch": 1.98,
|
| 1785 |
+
"learning_rate": 1.740867579908676e-05,
|
| 1786 |
+
"loss": 6.1972,
|
| 1787 |
+
"step": 578
|
| 1788 |
+
},
|
| 1789 |
+
{
|
| 1790 |
+
"epoch": 1.99,
|
| 1791 |
+
"learning_rate": 1.7294520547945207e-05,
|
| 1792 |
+
"loss": 6.0236,
|
| 1793 |
+
"step": 580
|
| 1794 |
+
},
|
| 1795 |
+
{
|
| 1796 |
+
"epoch": 1.99,
|
| 1797 |
+
"learning_rate": 1.7180365296803655e-05,
|
| 1798 |
+
"loss": 5.9499,
|
| 1799 |
+
"step": 582
|
| 1800 |
+
},
|
| 1801 |
+
{
|
| 1802 |
+
"epoch": 2.0,
|
| 1803 |
+
"learning_rate": 1.70662100456621e-05,
|
| 1804 |
+
"loss": 6.1688,
|
| 1805 |
+
"step": 584
|
| 1806 |
+
},
|
| 1807 |
+
{
|
| 1808 |
+
"epoch": 2.01,
|
| 1809 |
+
"learning_rate": 1.695205479452055e-05,
|
| 1810 |
+
"loss": 5.9289,
|
| 1811 |
+
"step": 586
|
| 1812 |
+
},
|
| 1813 |
+
{
|
| 1814 |
+
"epoch": 2.01,
|
| 1815 |
+
"learning_rate": 1.6837899543378995e-05,
|
| 1816 |
+
"loss": 5.5866,
|
| 1817 |
+
"step": 588
|
| 1818 |
+
},
|
| 1819 |
+
{
|
| 1820 |
+
"epoch": 2.02,
|
| 1821 |
+
"learning_rate": 1.6723744292237443e-05,
|
| 1822 |
+
"loss": 6.2259,
|
| 1823 |
+
"step": 590
|
| 1824 |
+
},
|
| 1825 |
+
{
|
| 1826 |
+
"epoch": 2.03,
|
| 1827 |
+
"learning_rate": 1.660958904109589e-05,
|
| 1828 |
+
"loss": 5.9349,
|
| 1829 |
+
"step": 592
|
| 1830 |
+
},
|
| 1831 |
+
{
|
| 1832 |
+
"epoch": 2.03,
|
| 1833 |
+
"learning_rate": 1.649543378995434e-05,
|
| 1834 |
+
"loss": 6.0916,
|
| 1835 |
+
"step": 594
|
| 1836 |
+
},
|
| 1837 |
+
{
|
| 1838 |
+
"epoch": 2.04,
|
| 1839 |
+
"learning_rate": 1.6381278538812787e-05,
|
| 1840 |
+
"loss": 5.8322,
|
| 1841 |
+
"step": 596
|
| 1842 |
+
},
|
| 1843 |
+
{
|
| 1844 |
+
"epoch": 2.05,
|
| 1845 |
+
"learning_rate": 1.6267123287671235e-05,
|
| 1846 |
+
"loss": 6.1475,
|
| 1847 |
+
"step": 598
|
| 1848 |
+
},
|
| 1849 |
+
{
|
| 1850 |
+
"epoch": 2.05,
|
| 1851 |
+
"learning_rate": 1.615296803652968e-05,
|
| 1852 |
+
"loss": 6.1342,
|
| 1853 |
+
"step": 600
|
| 1854 |
+
},
|
| 1855 |
+
{
|
| 1856 |
+
"epoch": 2.05,
|
| 1857 |
+
"eval_cer": 0.8386831951399721,
|
| 1858 |
+
"eval_loss": 6.42148494720459,
|
| 1859 |
+
"eval_runtime": 43.5835,
|
| 1860 |
+
"eval_samples_per_second": 13.377,
|
| 1861 |
+
"eval_steps_per_second": 1.675,
|
| 1862 |
+
"step": 600
|
| 1863 |
+
},
|
| 1864 |
+
{
|
| 1865 |
+
"epoch": 2.06,
|
| 1866 |
+
"learning_rate": 1.603881278538813e-05,
|
| 1867 |
+
"loss": 5.7958,
|
| 1868 |
+
"step": 602
|
| 1869 |
+
},
|
| 1870 |
+
{
|
| 1871 |
+
"epoch": 2.07,
|
| 1872 |
+
"learning_rate": 1.5924657534246576e-05,
|
| 1873 |
+
"loss": 6.3319,
|
| 1874 |
+
"step": 604
|
| 1875 |
+
},
|
| 1876 |
+
{
|
| 1877 |
+
"epoch": 2.08,
|
| 1878 |
+
"learning_rate": 1.5810502283105024e-05,
|
| 1879 |
+
"loss": 6.181,
|
| 1880 |
+
"step": 606
|
| 1881 |
+
},
|
| 1882 |
+
{
|
| 1883 |
+
"epoch": 2.08,
|
| 1884 |
+
"learning_rate": 1.5696347031963472e-05,
|
| 1885 |
+
"loss": 6.1645,
|
| 1886 |
+
"step": 608
|
| 1887 |
+
},
|
| 1888 |
+
{
|
| 1889 |
+
"epoch": 2.09,
|
| 1890 |
+
"learning_rate": 1.558219178082192e-05,
|
| 1891 |
+
"loss": 5.8628,
|
| 1892 |
+
"step": 610
|
| 1893 |
+
},
|
| 1894 |
+
{
|
| 1895 |
+
"epoch": 2.1,
|
| 1896 |
+
"learning_rate": 1.5468036529680365e-05,
|
| 1897 |
+
"loss": 6.1039,
|
| 1898 |
+
"step": 612
|
| 1899 |
+
},
|
| 1900 |
+
{
|
| 1901 |
+
"epoch": 2.1,
|
| 1902 |
+
"learning_rate": 1.5353881278538816e-05,
|
| 1903 |
+
"loss": 6.196,
|
| 1904 |
+
"step": 614
|
| 1905 |
+
},
|
| 1906 |
+
{
|
| 1907 |
+
"epoch": 2.11,
|
| 1908 |
+
"learning_rate": 1.523972602739726e-05,
|
| 1909 |
+
"loss": 6.0791,
|
| 1910 |
+
"step": 616
|
| 1911 |
+
},
|
| 1912 |
+
{
|
| 1913 |
+
"epoch": 2.12,
|
| 1914 |
+
"learning_rate": 1.512557077625571e-05,
|
| 1915 |
+
"loss": 6.2871,
|
| 1916 |
+
"step": 618
|
| 1917 |
+
},
|
| 1918 |
+
{
|
| 1919 |
+
"epoch": 2.12,
|
| 1920 |
+
"learning_rate": 1.5011415525114157e-05,
|
| 1921 |
+
"loss": 5.8731,
|
| 1922 |
+
"step": 620
|
| 1923 |
+
},
|
| 1924 |
+
{
|
| 1925 |
+
"epoch": 2.13,
|
| 1926 |
+
"learning_rate": 1.4897260273972605e-05,
|
| 1927 |
+
"loss": 6.1108,
|
| 1928 |
+
"step": 622
|
| 1929 |
+
},
|
| 1930 |
+
{
|
| 1931 |
+
"epoch": 2.14,
|
| 1932 |
+
"learning_rate": 1.4783105022831051e-05,
|
| 1933 |
+
"loss": 5.9225,
|
| 1934 |
+
"step": 624
|
| 1935 |
+
},
|
| 1936 |
+
{
|
| 1937 |
+
"epoch": 2.14,
|
| 1938 |
+
"learning_rate": 1.4668949771689499e-05,
|
| 1939 |
+
"loss": 6.1894,
|
| 1940 |
+
"step": 626
|
| 1941 |
+
},
|
| 1942 |
+
{
|
| 1943 |
+
"epoch": 2.15,
|
| 1944 |
+
"learning_rate": 1.4554794520547945e-05,
|
| 1945 |
+
"loss": 5.7433,
|
| 1946 |
+
"step": 628
|
| 1947 |
+
},
|
| 1948 |
+
{
|
| 1949 |
+
"epoch": 2.16,
|
| 1950 |
+
"learning_rate": 1.4440639269406395e-05,
|
| 1951 |
+
"loss": 6.0826,
|
| 1952 |
+
"step": 630
|
| 1953 |
+
},
|
| 1954 |
+
{
|
| 1955 |
+
"epoch": 2.16,
|
| 1956 |
+
"learning_rate": 1.4326484018264841e-05,
|
| 1957 |
+
"loss": 6.0097,
|
| 1958 |
+
"step": 632
|
| 1959 |
+
},
|
| 1960 |
+
{
|
| 1961 |
+
"epoch": 2.17,
|
| 1962 |
+
"learning_rate": 1.421232876712329e-05,
|
| 1963 |
+
"loss": 5.8454,
|
| 1964 |
+
"step": 634
|
| 1965 |
+
},
|
| 1966 |
+
{
|
| 1967 |
+
"epoch": 2.18,
|
| 1968 |
+
"learning_rate": 1.4098173515981735e-05,
|
| 1969 |
+
"loss": 6.1941,
|
| 1970 |
+
"step": 636
|
| 1971 |
+
},
|
| 1972 |
+
{
|
| 1973 |
+
"epoch": 2.18,
|
| 1974 |
+
"learning_rate": 1.3984018264840185e-05,
|
| 1975 |
+
"loss": 5.8484,
|
| 1976 |
+
"step": 638
|
| 1977 |
+
},
|
| 1978 |
+
{
|
| 1979 |
+
"epoch": 2.19,
|
| 1980 |
+
"learning_rate": 1.3869863013698631e-05,
|
| 1981 |
+
"loss": 6.328,
|
| 1982 |
+
"step": 640
|
| 1983 |
+
},
|
| 1984 |
+
{
|
| 1985 |
+
"epoch": 2.2,
|
| 1986 |
+
"learning_rate": 1.375570776255708e-05,
|
| 1987 |
+
"loss": 6.3417,
|
| 1988 |
+
"step": 642
|
| 1989 |
+
},
|
| 1990 |
+
{
|
| 1991 |
+
"epoch": 2.21,
|
| 1992 |
+
"learning_rate": 1.3641552511415526e-05,
|
| 1993 |
+
"loss": 6.1993,
|
| 1994 |
+
"step": 644
|
| 1995 |
+
},
|
| 1996 |
+
{
|
| 1997 |
+
"epoch": 2.21,
|
| 1998 |
+
"learning_rate": 1.3527397260273974e-05,
|
| 1999 |
+
"loss": 6.0958,
|
| 2000 |
+
"step": 646
|
| 2001 |
+
},
|
| 2002 |
+
{
|
| 2003 |
+
"epoch": 2.22,
|
| 2004 |
+
"learning_rate": 1.341324200913242e-05,
|
| 2005 |
+
"loss": 6.0639,
|
| 2006 |
+
"step": 648
|
| 2007 |
+
},
|
| 2008 |
+
{
|
| 2009 |
+
"epoch": 2.23,
|
| 2010 |
+
"learning_rate": 1.329908675799087e-05,
|
| 2011 |
+
"loss": 6.0299,
|
| 2012 |
+
"step": 650
|
| 2013 |
+
},
|
| 2014 |
+
{
|
| 2015 |
+
"epoch": 2.23,
|
| 2016 |
+
"learning_rate": 1.3184931506849316e-05,
|
| 2017 |
+
"loss": 6.1234,
|
| 2018 |
+
"step": 652
|
| 2019 |
+
},
|
| 2020 |
+
{
|
| 2021 |
+
"epoch": 2.24,
|
| 2022 |
+
"learning_rate": 1.3070776255707764e-05,
|
| 2023 |
+
"loss": 6.0303,
|
| 2024 |
+
"step": 654
|
| 2025 |
+
},
|
| 2026 |
+
{
|
| 2027 |
+
"epoch": 2.25,
|
| 2028 |
+
"learning_rate": 1.295662100456621e-05,
|
| 2029 |
+
"loss": 6.0098,
|
| 2030 |
+
"step": 656
|
| 2031 |
+
},
|
| 2032 |
+
{
|
| 2033 |
+
"epoch": 2.25,
|
| 2034 |
+
"learning_rate": 1.284246575342466e-05,
|
| 2035 |
+
"loss": 5.9058,
|
| 2036 |
+
"step": 658
|
| 2037 |
+
},
|
| 2038 |
+
{
|
| 2039 |
+
"epoch": 2.26,
|
| 2040 |
+
"learning_rate": 1.2728310502283106e-05,
|
| 2041 |
+
"loss": 6.2086,
|
| 2042 |
+
"step": 660
|
| 2043 |
+
},
|
| 2044 |
+
{
|
| 2045 |
+
"epoch": 2.27,
|
| 2046 |
+
"learning_rate": 1.2614155251141554e-05,
|
| 2047 |
+
"loss": 6.1667,
|
| 2048 |
+
"step": 662
|
| 2049 |
+
},
|
| 2050 |
+
{
|
| 2051 |
+
"epoch": 2.27,
|
| 2052 |
+
"learning_rate": 1.25e-05,
|
| 2053 |
+
"loss": 6.26,
|
| 2054 |
+
"step": 664
|
| 2055 |
+
},
|
| 2056 |
+
{
|
| 2057 |
+
"epoch": 2.28,
|
| 2058 |
+
"learning_rate": 1.2385844748858449e-05,
|
| 2059 |
+
"loss": 5.698,
|
| 2060 |
+
"step": 666
|
| 2061 |
+
},
|
| 2062 |
+
{
|
| 2063 |
+
"epoch": 2.29,
|
| 2064 |
+
"learning_rate": 1.2271689497716895e-05,
|
| 2065 |
+
"loss": 6.0933,
|
| 2066 |
+
"step": 668
|
| 2067 |
+
},
|
| 2068 |
+
{
|
| 2069 |
+
"epoch": 2.29,
|
| 2070 |
+
"learning_rate": 1.2157534246575343e-05,
|
| 2071 |
+
"loss": 6.0448,
|
| 2072 |
+
"step": 670
|
| 2073 |
+
},
|
| 2074 |
+
{
|
| 2075 |
+
"epoch": 2.3,
|
| 2076 |
+
"learning_rate": 1.2043378995433791e-05,
|
| 2077 |
+
"loss": 6.1017,
|
| 2078 |
+
"step": 672
|
| 2079 |
+
},
|
| 2080 |
+
{
|
| 2081 |
+
"epoch": 2.31,
|
| 2082 |
+
"learning_rate": 1.1929223744292239e-05,
|
| 2083 |
+
"loss": 6.2529,
|
| 2084 |
+
"step": 674
|
| 2085 |
+
},
|
| 2086 |
+
{
|
| 2087 |
+
"epoch": 2.32,
|
| 2088 |
+
"learning_rate": 1.1815068493150685e-05,
|
| 2089 |
+
"loss": 5.9125,
|
| 2090 |
+
"step": 676
|
| 2091 |
+
},
|
| 2092 |
+
{
|
| 2093 |
+
"epoch": 2.32,
|
| 2094 |
+
"learning_rate": 1.1700913242009133e-05,
|
| 2095 |
+
"loss": 6.0414,
|
| 2096 |
+
"step": 678
|
| 2097 |
+
},
|
| 2098 |
+
{
|
| 2099 |
+
"epoch": 2.33,
|
| 2100 |
+
"learning_rate": 1.1586757990867581e-05,
|
| 2101 |
+
"loss": 6.1012,
|
| 2102 |
+
"step": 680
|
| 2103 |
+
},
|
| 2104 |
+
{
|
| 2105 |
+
"epoch": 2.34,
|
| 2106 |
+
"learning_rate": 1.1472602739726027e-05,
|
| 2107 |
+
"loss": 6.1578,
|
| 2108 |
+
"step": 682
|
| 2109 |
+
},
|
| 2110 |
+
{
|
| 2111 |
+
"epoch": 2.34,
|
| 2112 |
+
"learning_rate": 1.1358447488584475e-05,
|
| 2113 |
+
"loss": 6.1924,
|
| 2114 |
+
"step": 684
|
| 2115 |
+
},
|
| 2116 |
+
{
|
| 2117 |
+
"epoch": 2.35,
|
| 2118 |
+
"learning_rate": 1.1244292237442923e-05,
|
| 2119 |
+
"loss": 5.9418,
|
| 2120 |
+
"step": 686
|
| 2121 |
+
},
|
| 2122 |
+
{
|
| 2123 |
+
"epoch": 2.36,
|
| 2124 |
+
"learning_rate": 1.113013698630137e-05,
|
| 2125 |
+
"loss": 6.0778,
|
| 2126 |
+
"step": 688
|
| 2127 |
+
},
|
| 2128 |
+
{
|
| 2129 |
+
"epoch": 2.36,
|
| 2130 |
+
"learning_rate": 1.1015981735159818e-05,
|
| 2131 |
+
"loss": 6.1784,
|
| 2132 |
+
"step": 690
|
| 2133 |
+
},
|
| 2134 |
+
{
|
| 2135 |
+
"epoch": 2.37,
|
| 2136 |
+
"learning_rate": 1.0901826484018266e-05,
|
| 2137 |
+
"loss": 6.2346,
|
| 2138 |
+
"step": 692
|
| 2139 |
+
},
|
| 2140 |
+
{
|
| 2141 |
+
"epoch": 2.38,
|
| 2142 |
+
"learning_rate": 1.0787671232876714e-05,
|
| 2143 |
+
"loss": 6.0849,
|
| 2144 |
+
"step": 694
|
| 2145 |
+
},
|
| 2146 |
+
{
|
| 2147 |
+
"epoch": 2.38,
|
| 2148 |
+
"learning_rate": 1.067351598173516e-05,
|
| 2149 |
+
"loss": 6.0025,
|
| 2150 |
+
"step": 696
|
| 2151 |
+
},
|
| 2152 |
+
{
|
| 2153 |
+
"epoch": 2.39,
|
| 2154 |
+
"learning_rate": 1.0559360730593608e-05,
|
| 2155 |
+
"loss": 6.2221,
|
| 2156 |
+
"step": 698
|
| 2157 |
+
},
|
| 2158 |
+
{
|
| 2159 |
+
"epoch": 2.4,
|
| 2160 |
+
"learning_rate": 1.0445205479452056e-05,
|
| 2161 |
+
"loss": 5.8295,
|
| 2162 |
+
"step": 700
|
| 2163 |
+
},
|
| 2164 |
+
{
|
| 2165 |
+
"epoch": 2.4,
|
| 2166 |
+
"eval_cer": 0.7918069124045645,
|
| 2167 |
+
"eval_loss": 6.3999199867248535,
|
| 2168 |
+
"eval_runtime": 40.7886,
|
| 2169 |
+
"eval_samples_per_second": 14.293,
|
| 2170 |
+
"eval_steps_per_second": 1.79,
|
| 2171 |
+
"step": 700
|
| 2172 |
+
},
|
| 2173 |
+
{
|
| 2174 |
+
"epoch": 2.4,
|
| 2175 |
+
"learning_rate": 1.0331050228310502e-05,
|
| 2176 |
+
"loss": 5.8848,
|
| 2177 |
+
"step": 702
|
| 2178 |
+
},
|
| 2179 |
+
{
|
| 2180 |
+
"epoch": 2.41,
|
| 2181 |
+
"learning_rate": 1.021689497716895e-05,
|
| 2182 |
+
"loss": 6.0501,
|
| 2183 |
+
"step": 704
|
| 2184 |
+
},
|
| 2185 |
+
{
|
| 2186 |
+
"epoch": 2.42,
|
| 2187 |
+
"learning_rate": 1.0102739726027398e-05,
|
| 2188 |
+
"loss": 6.1242,
|
| 2189 |
+
"step": 706
|
| 2190 |
+
},
|
| 2191 |
+
{
|
| 2192 |
+
"epoch": 2.42,
|
| 2193 |
+
"learning_rate": 9.988584474885845e-06,
|
| 2194 |
+
"loss": 5.9263,
|
| 2195 |
+
"step": 708
|
| 2196 |
+
},
|
| 2197 |
+
{
|
| 2198 |
+
"epoch": 2.43,
|
| 2199 |
+
"learning_rate": 9.874429223744293e-06,
|
| 2200 |
+
"loss": 5.897,
|
| 2201 |
+
"step": 710
|
| 2202 |
+
},
|
| 2203 |
+
{
|
| 2204 |
+
"epoch": 2.44,
|
| 2205 |
+
"learning_rate": 9.76027397260274e-06,
|
| 2206 |
+
"loss": 6.3139,
|
| 2207 |
+
"step": 712
|
| 2208 |
+
},
|
| 2209 |
+
{
|
| 2210 |
+
"epoch": 2.45,
|
| 2211 |
+
"learning_rate": 9.646118721461189e-06,
|
| 2212 |
+
"loss": 5.9507,
|
| 2213 |
+
"step": 714
|
| 2214 |
+
},
|
| 2215 |
+
{
|
| 2216 |
+
"epoch": 2.45,
|
| 2217 |
+
"learning_rate": 9.531963470319635e-06,
|
| 2218 |
+
"loss": 5.9616,
|
| 2219 |
+
"step": 716
|
| 2220 |
+
},
|
| 2221 |
+
{
|
| 2222 |
+
"epoch": 2.46,
|
| 2223 |
+
"learning_rate": 9.417808219178083e-06,
|
| 2224 |
+
"loss": 6.0359,
|
| 2225 |
+
"step": 718
|
| 2226 |
+
},
|
| 2227 |
+
{
|
| 2228 |
+
"epoch": 2.47,
|
| 2229 |
+
"learning_rate": 9.303652968036531e-06,
|
| 2230 |
+
"loss": 5.8949,
|
| 2231 |
+
"step": 720
|
| 2232 |
+
},
|
| 2233 |
+
{
|
| 2234 |
+
"epoch": 2.47,
|
| 2235 |
+
"learning_rate": 9.189497716894977e-06,
|
| 2236 |
+
"loss": 6.1793,
|
| 2237 |
+
"step": 722
|
| 2238 |
+
},
|
| 2239 |
+
{
|
| 2240 |
+
"epoch": 2.48,
|
| 2241 |
+
"learning_rate": 9.075342465753425e-06,
|
| 2242 |
+
"loss": 5.9515,
|
| 2243 |
+
"step": 724
|
| 2244 |
+
},
|
| 2245 |
+
{
|
| 2246 |
+
"epoch": 2.49,
|
| 2247 |
+
"learning_rate": 8.961187214611873e-06,
|
| 2248 |
+
"loss": 5.8504,
|
| 2249 |
+
"step": 726
|
| 2250 |
+
},
|
| 2251 |
+
{
|
| 2252 |
+
"epoch": 2.49,
|
| 2253 |
+
"learning_rate": 8.84703196347032e-06,
|
| 2254 |
+
"loss": 6.0409,
|
| 2255 |
+
"step": 728
|
| 2256 |
+
},
|
| 2257 |
+
{
|
| 2258 |
+
"epoch": 2.5,
|
| 2259 |
+
"learning_rate": 8.732876712328767e-06,
|
| 2260 |
+
"loss": 6.2968,
|
| 2261 |
+
"step": 730
|
| 2262 |
+
},
|
| 2263 |
+
{
|
| 2264 |
+
"epoch": 2.51,
|
| 2265 |
+
"learning_rate": 8.618721461187215e-06,
|
| 2266 |
+
"loss": 5.8559,
|
| 2267 |
+
"step": 732
|
| 2268 |
+
},
|
| 2269 |
+
{
|
| 2270 |
+
"epoch": 2.51,
|
| 2271 |
+
"learning_rate": 8.504566210045663e-06,
|
| 2272 |
+
"loss": 5.7911,
|
| 2273 |
+
"step": 734
|
| 2274 |
+
},
|
| 2275 |
+
{
|
| 2276 |
+
"epoch": 2.52,
|
| 2277 |
+
"learning_rate": 8.39041095890411e-06,
|
| 2278 |
+
"loss": 6.1524,
|
| 2279 |
+
"step": 736
|
| 2280 |
+
},
|
| 2281 |
+
{
|
| 2282 |
+
"epoch": 2.53,
|
| 2283 |
+
"learning_rate": 8.276255707762558e-06,
|
| 2284 |
+
"loss": 5.7204,
|
| 2285 |
+
"step": 738
|
| 2286 |
+
},
|
| 2287 |
+
{
|
| 2288 |
+
"epoch": 2.53,
|
| 2289 |
+
"learning_rate": 8.162100456621006e-06,
|
| 2290 |
+
"loss": 6.1016,
|
| 2291 |
+
"step": 740
|
| 2292 |
+
},
|
| 2293 |
+
{
|
| 2294 |
+
"epoch": 2.54,
|
| 2295 |
+
"learning_rate": 8.047945205479452e-06,
|
| 2296 |
+
"loss": 5.9837,
|
| 2297 |
+
"step": 742
|
| 2298 |
+
},
|
| 2299 |
+
{
|
| 2300 |
+
"epoch": 2.55,
|
| 2301 |
+
"learning_rate": 7.9337899543379e-06,
|
| 2302 |
+
"loss": 5.8908,
|
| 2303 |
+
"step": 744
|
| 2304 |
+
},
|
| 2305 |
+
{
|
| 2306 |
+
"epoch": 2.55,
|
| 2307 |
+
"learning_rate": 7.819634703196348e-06,
|
| 2308 |
+
"loss": 5.843,
|
| 2309 |
+
"step": 746
|
| 2310 |
+
},
|
| 2311 |
+
{
|
| 2312 |
+
"epoch": 2.56,
|
| 2313 |
+
"learning_rate": 7.705479452054794e-06,
|
| 2314 |
+
"loss": 5.9089,
|
| 2315 |
+
"step": 748
|
| 2316 |
+
},
|
| 2317 |
+
{
|
| 2318 |
+
"epoch": 2.57,
|
| 2319 |
+
"learning_rate": 7.591324200913242e-06,
|
| 2320 |
+
"loss": 6.0063,
|
| 2321 |
+
"step": 750
|
| 2322 |
+
},
|
| 2323 |
+
{
|
| 2324 |
+
"epoch": 2.58,
|
| 2325 |
+
"learning_rate": 7.47716894977169e-06,
|
| 2326 |
+
"loss": 5.9461,
|
| 2327 |
+
"step": 752
|
| 2328 |
+
},
|
| 2329 |
+
{
|
| 2330 |
+
"epoch": 2.58,
|
| 2331 |
+
"learning_rate": 7.3630136986301374e-06,
|
| 2332 |
+
"loss": 5.6393,
|
| 2333 |
+
"step": 754
|
| 2334 |
+
},
|
| 2335 |
+
{
|
| 2336 |
+
"epoch": 2.59,
|
| 2337 |
+
"learning_rate": 7.2488584474885854e-06,
|
| 2338 |
+
"loss": 6.0984,
|
| 2339 |
+
"step": 756
|
| 2340 |
+
},
|
| 2341 |
+
{
|
| 2342 |
+
"epoch": 2.6,
|
| 2343 |
+
"learning_rate": 7.134703196347033e-06,
|
| 2344 |
+
"loss": 5.7783,
|
| 2345 |
+
"step": 758
|
| 2346 |
+
},
|
| 2347 |
+
{
|
| 2348 |
+
"epoch": 2.6,
|
| 2349 |
+
"learning_rate": 7.02054794520548e-06,
|
| 2350 |
+
"loss": 5.8534,
|
| 2351 |
+
"step": 760
|
| 2352 |
+
},
|
| 2353 |
+
{
|
| 2354 |
+
"epoch": 2.61,
|
| 2355 |
+
"learning_rate": 6.906392694063928e-06,
|
| 2356 |
+
"loss": 5.9853,
|
| 2357 |
+
"step": 762
|
| 2358 |
+
},
|
| 2359 |
+
{
|
| 2360 |
+
"epoch": 2.62,
|
| 2361 |
+
"learning_rate": 6.792237442922375e-06,
|
| 2362 |
+
"loss": 6.2198,
|
| 2363 |
+
"step": 764
|
| 2364 |
+
},
|
| 2365 |
+
{
|
| 2366 |
+
"epoch": 2.62,
|
| 2367 |
+
"learning_rate": 6.678082191780823e-06,
|
| 2368 |
+
"loss": 6.1982,
|
| 2369 |
+
"step": 766
|
| 2370 |
+
},
|
| 2371 |
+
{
|
| 2372 |
+
"epoch": 2.63,
|
| 2373 |
+
"learning_rate": 6.56392694063927e-06,
|
| 2374 |
+
"loss": 5.9753,
|
| 2375 |
+
"step": 768
|
| 2376 |
+
},
|
| 2377 |
+
{
|
| 2378 |
+
"epoch": 2.64,
|
| 2379 |
+
"learning_rate": 6.449771689497717e-06,
|
| 2380 |
+
"loss": 5.806,
|
| 2381 |
+
"step": 770
|
| 2382 |
+
},
|
| 2383 |
+
{
|
| 2384 |
+
"epoch": 2.64,
|
| 2385 |
+
"learning_rate": 6.335616438356165e-06,
|
| 2386 |
+
"loss": 5.8563,
|
| 2387 |
+
"step": 772
|
| 2388 |
+
},
|
| 2389 |
+
{
|
| 2390 |
+
"epoch": 2.65,
|
| 2391 |
+
"learning_rate": 6.2214611872146115e-06,
|
| 2392 |
+
"loss": 6.0158,
|
| 2393 |
+
"step": 774
|
| 2394 |
+
},
|
| 2395 |
+
{
|
| 2396 |
+
"epoch": 2.66,
|
| 2397 |
+
"learning_rate": 6.1073059360730594e-06,
|
| 2398 |
+
"loss": 6.2569,
|
| 2399 |
+
"step": 776
|
| 2400 |
+
},
|
| 2401 |
+
{
|
| 2402 |
+
"epoch": 2.66,
|
| 2403 |
+
"learning_rate": 5.993150684931507e-06,
|
| 2404 |
+
"loss": 5.9379,
|
| 2405 |
+
"step": 778
|
| 2406 |
+
},
|
| 2407 |
+
{
|
| 2408 |
+
"epoch": 2.67,
|
| 2409 |
+
"learning_rate": 5.878995433789955e-06,
|
| 2410 |
+
"loss": 5.914,
|
| 2411 |
+
"step": 780
|
| 2412 |
+
},
|
| 2413 |
+
{
|
| 2414 |
+
"epoch": 2.68,
|
| 2415 |
+
"learning_rate": 5.764840182648402e-06,
|
| 2416 |
+
"loss": 5.847,
|
| 2417 |
+
"step": 782
|
| 2418 |
+
},
|
| 2419 |
+
{
|
| 2420 |
+
"epoch": 2.68,
|
| 2421 |
+
"learning_rate": 5.650684931506849e-06,
|
| 2422 |
+
"loss": 5.9925,
|
| 2423 |
+
"step": 784
|
| 2424 |
+
},
|
| 2425 |
+
{
|
| 2426 |
+
"epoch": 2.69,
|
| 2427 |
+
"learning_rate": 5.536529680365297e-06,
|
| 2428 |
+
"loss": 5.7515,
|
| 2429 |
+
"step": 786
|
| 2430 |
+
},
|
| 2431 |
+
{
|
| 2432 |
+
"epoch": 2.7,
|
| 2433 |
+
"learning_rate": 5.422374429223744e-06,
|
| 2434 |
+
"loss": 6.0283,
|
| 2435 |
+
"step": 788
|
| 2436 |
+
},
|
| 2437 |
+
{
|
| 2438 |
+
"epoch": 2.71,
|
| 2439 |
+
"learning_rate": 5.308219178082192e-06,
|
| 2440 |
+
"loss": 5.8207,
|
| 2441 |
+
"step": 790
|
| 2442 |
+
},
|
| 2443 |
+
{
|
| 2444 |
+
"epoch": 2.71,
|
| 2445 |
+
"learning_rate": 5.194063926940639e-06,
|
| 2446 |
+
"loss": 5.9529,
|
| 2447 |
+
"step": 792
|
| 2448 |
+
},
|
| 2449 |
+
{
|
| 2450 |
+
"epoch": 2.72,
|
| 2451 |
+
"learning_rate": 5.079908675799086e-06,
|
| 2452 |
+
"loss": 6.3816,
|
| 2453 |
+
"step": 794
|
| 2454 |
+
},
|
| 2455 |
+
{
|
| 2456 |
+
"epoch": 2.73,
|
| 2457 |
+
"learning_rate": 4.965753424657534e-06,
|
| 2458 |
+
"loss": 5.921,
|
| 2459 |
+
"step": 796
|
| 2460 |
+
},
|
| 2461 |
+
{
|
| 2462 |
+
"epoch": 2.73,
|
| 2463 |
+
"learning_rate": 4.8515981735159814e-06,
|
| 2464 |
+
"loss": 5.9369,
|
| 2465 |
+
"step": 798
|
| 2466 |
+
},
|
| 2467 |
+
{
|
| 2468 |
+
"epoch": 2.74,
|
| 2469 |
+
"learning_rate": 4.7374429223744294e-06,
|
| 2470 |
+
"loss": 6.0542,
|
| 2471 |
+
"step": 800
|
| 2472 |
+
},
|
| 2473 |
+
{
|
| 2474 |
+
"epoch": 2.74,
|
| 2475 |
+
"eval_cer": 0.7485838601100074,
|
| 2476 |
+
"eval_loss": 6.38665771484375,
|
| 2477 |
+
"eval_runtime": 43.0264,
|
| 2478 |
+
"eval_samples_per_second": 13.55,
|
| 2479 |
+
"eval_steps_per_second": 1.697,
|
| 2480 |
+
"step": 800
|
| 2481 |
+
},
|
| 2482 |
+
{
|
| 2483 |
+
"epoch": 2.75,
|
| 2484 |
+
"learning_rate": 4.623287671232877e-06,
|
| 2485 |
+
"loss": 5.6539,
|
| 2486 |
+
"step": 802
|
| 2487 |
+
},
|
| 2488 |
+
{
|
| 2489 |
+
"epoch": 2.75,
|
| 2490 |
+
"learning_rate": 4.509132420091325e-06,
|
| 2491 |
+
"loss": 6.0837,
|
| 2492 |
+
"step": 804
|
| 2493 |
+
},
|
| 2494 |
+
{
|
| 2495 |
+
"epoch": 2.76,
|
| 2496 |
+
"learning_rate": 4.394977168949772e-06,
|
| 2497 |
+
"loss": 5.5705,
|
| 2498 |
+
"step": 806
|
| 2499 |
+
},
|
| 2500 |
+
{
|
| 2501 |
+
"epoch": 2.77,
|
| 2502 |
+
"learning_rate": 4.280821917808219e-06,
|
| 2503 |
+
"loss": 5.797,
|
| 2504 |
+
"step": 808
|
| 2505 |
+
},
|
| 2506 |
+
{
|
| 2507 |
+
"epoch": 2.77,
|
| 2508 |
+
"learning_rate": 4.166666666666667e-06,
|
| 2509 |
+
"loss": 6.2038,
|
| 2510 |
+
"step": 810
|
| 2511 |
+
},
|
| 2512 |
+
{
|
| 2513 |
+
"epoch": 2.78,
|
| 2514 |
+
"learning_rate": 4.052511415525114e-06,
|
| 2515 |
+
"loss": 5.8418,
|
| 2516 |
+
"step": 812
|
| 2517 |
+
},
|
| 2518 |
+
{
|
| 2519 |
+
"epoch": 2.79,
|
| 2520 |
+
"learning_rate": 3.938356164383562e-06,
|
| 2521 |
+
"loss": 5.9495,
|
| 2522 |
+
"step": 814
|
| 2523 |
+
},
|
| 2524 |
+
{
|
| 2525 |
+
"epoch": 2.79,
|
| 2526 |
+
"learning_rate": 3.824200913242009e-06,
|
| 2527 |
+
"loss": 6.2109,
|
| 2528 |
+
"step": 816
|
| 2529 |
+
},
|
| 2530 |
+
{
|
| 2531 |
+
"epoch": 2.8,
|
| 2532 |
+
"learning_rate": 3.7100456621004567e-06,
|
| 2533 |
+
"loss": 5.8276,
|
| 2534 |
+
"step": 818
|
| 2535 |
+
},
|
| 2536 |
+
{
|
| 2537 |
+
"epoch": 2.81,
|
| 2538 |
+
"learning_rate": 3.5958904109589043e-06,
|
| 2539 |
+
"loss": 5.8859,
|
| 2540 |
+
"step": 820
|
| 2541 |
+
},
|
| 2542 |
+
{
|
| 2543 |
+
"epoch": 2.82,
|
| 2544 |
+
"learning_rate": 3.4817351598173514e-06,
|
| 2545 |
+
"loss": 5.9916,
|
| 2546 |
+
"step": 822
|
| 2547 |
+
},
|
| 2548 |
+
{
|
| 2549 |
+
"epoch": 2.82,
|
| 2550 |
+
"learning_rate": 3.367579908675799e-06,
|
| 2551 |
+
"loss": 6.2421,
|
| 2552 |
+
"step": 824
|
| 2553 |
+
},
|
| 2554 |
+
{
|
| 2555 |
+
"epoch": 2.83,
|
| 2556 |
+
"learning_rate": 3.2534246575342466e-06,
|
| 2557 |
+
"loss": 5.9774,
|
| 2558 |
+
"step": 826
|
| 2559 |
+
},
|
| 2560 |
+
{
|
| 2561 |
+
"epoch": 2.84,
|
| 2562 |
+
"learning_rate": 3.139269406392694e-06,
|
| 2563 |
+
"loss": 5.6032,
|
| 2564 |
+
"step": 828
|
| 2565 |
+
},
|
| 2566 |
+
{
|
| 2567 |
+
"epoch": 2.84,
|
| 2568 |
+
"learning_rate": 3.0251141552511417e-06,
|
| 2569 |
+
"loss": 6.0698,
|
| 2570 |
+
"step": 830
|
| 2571 |
+
},
|
| 2572 |
+
{
|
| 2573 |
+
"epoch": 2.85,
|
| 2574 |
+
"learning_rate": 2.910958904109589e-06,
|
| 2575 |
+
"loss": 6.1411,
|
| 2576 |
+
"step": 832
|
| 2577 |
+
},
|
| 2578 |
+
{
|
| 2579 |
+
"epoch": 2.86,
|
| 2580 |
+
"learning_rate": 2.7968036529680364e-06,
|
| 2581 |
+
"loss": 6.0411,
|
| 2582 |
+
"step": 834
|
| 2583 |
+
},
|
| 2584 |
+
{
|
| 2585 |
+
"epoch": 2.86,
|
| 2586 |
+
"learning_rate": 2.682648401826484e-06,
|
| 2587 |
+
"loss": 5.9259,
|
| 2588 |
+
"step": 836
|
| 2589 |
+
},
|
| 2590 |
+
{
|
| 2591 |
+
"epoch": 2.87,
|
| 2592 |
+
"learning_rate": 2.5684931506849316e-06,
|
| 2593 |
+
"loss": 6.1342,
|
| 2594 |
+
"step": 838
|
| 2595 |
+
},
|
| 2596 |
+
{
|
| 2597 |
+
"epoch": 2.88,
|
| 2598 |
+
"learning_rate": 2.454337899543379e-06,
|
| 2599 |
+
"loss": 5.9385,
|
| 2600 |
+
"step": 840
|
| 2601 |
+
},
|
| 2602 |
+
{
|
| 2603 |
+
"epoch": 2.88,
|
| 2604 |
+
"learning_rate": 2.3401826484018263e-06,
|
| 2605 |
+
"loss": 6.1046,
|
| 2606 |
+
"step": 842
|
| 2607 |
+
},
|
| 2608 |
+
{
|
| 2609 |
+
"epoch": 2.89,
|
| 2610 |
+
"learning_rate": 2.226027397260274e-06,
|
| 2611 |
+
"loss": 5.5534,
|
| 2612 |
+
"step": 844
|
| 2613 |
+
},
|
| 2614 |
+
{
|
| 2615 |
+
"epoch": 2.9,
|
| 2616 |
+
"learning_rate": 2.1118721461187214e-06,
|
| 2617 |
+
"loss": 6.0064,
|
| 2618 |
+
"step": 846
|
| 2619 |
+
},
|
| 2620 |
+
{
|
| 2621 |
+
"epoch": 2.9,
|
| 2622 |
+
"learning_rate": 1.997716894977169e-06,
|
| 2623 |
+
"loss": 5.9357,
|
| 2624 |
+
"step": 848
|
| 2625 |
+
},
|
| 2626 |
+
{
|
| 2627 |
+
"epoch": 2.91,
|
| 2628 |
+
"learning_rate": 1.8835616438356164e-06,
|
| 2629 |
+
"loss": 5.591,
|
| 2630 |
+
"step": 850
|
| 2631 |
+
},
|
| 2632 |
+
{
|
| 2633 |
+
"epoch": 2.92,
|
| 2634 |
+
"learning_rate": 1.769406392694064e-06,
|
| 2635 |
+
"loss": 6.0179,
|
| 2636 |
+
"step": 852
|
| 2637 |
+
},
|
| 2638 |
+
{
|
| 2639 |
+
"epoch": 2.92,
|
| 2640 |
+
"learning_rate": 1.6552511415525115e-06,
|
| 2641 |
+
"loss": 6.0487,
|
| 2642 |
+
"step": 854
|
| 2643 |
+
},
|
| 2644 |
+
{
|
| 2645 |
+
"epoch": 2.93,
|
| 2646 |
+
"learning_rate": 1.5410958904109589e-06,
|
| 2647 |
+
"loss": 6.0967,
|
| 2648 |
+
"step": 856
|
| 2649 |
+
},
|
| 2650 |
+
{
|
| 2651 |
+
"epoch": 2.94,
|
| 2652 |
+
"learning_rate": 1.4269406392694064e-06,
|
| 2653 |
+
"loss": 6.1703,
|
| 2654 |
+
"step": 858
|
| 2655 |
+
},
|
| 2656 |
+
{
|
| 2657 |
+
"epoch": 2.95,
|
| 2658 |
+
"learning_rate": 1.3127853881278538e-06,
|
| 2659 |
+
"loss": 5.8776,
|
| 2660 |
+
"step": 860
|
| 2661 |
+
},
|
| 2662 |
+
{
|
| 2663 |
+
"epoch": 2.95,
|
| 2664 |
+
"learning_rate": 1.1986301369863014e-06,
|
| 2665 |
+
"loss": 5.8851,
|
| 2666 |
+
"step": 862
|
| 2667 |
+
},
|
| 2668 |
+
{
|
| 2669 |
+
"epoch": 2.96,
|
| 2670 |
+
"learning_rate": 1.084474885844749e-06,
|
| 2671 |
+
"loss": 5.7484,
|
| 2672 |
+
"step": 864
|
| 2673 |
+
},
|
| 2674 |
+
{
|
| 2675 |
+
"epoch": 2.97,
|
| 2676 |
+
"learning_rate": 9.703196347031963e-07,
|
| 2677 |
+
"loss": 6.1194,
|
| 2678 |
+
"step": 866
|
| 2679 |
+
},
|
| 2680 |
+
{
|
| 2681 |
+
"epoch": 2.97,
|
| 2682 |
+
"learning_rate": 8.561643835616439e-07,
|
| 2683 |
+
"loss": 5.8983,
|
| 2684 |
+
"step": 868
|
| 2685 |
+
},
|
| 2686 |
+
{
|
| 2687 |
+
"epoch": 2.98,
|
| 2688 |
+
"learning_rate": 7.420091324200913e-07,
|
| 2689 |
+
"loss": 6.2007,
|
| 2690 |
+
"step": 870
|
| 2691 |
+
},
|
| 2692 |
+
{
|
| 2693 |
+
"epoch": 2.99,
|
| 2694 |
+
"learning_rate": 6.278538812785388e-07,
|
| 2695 |
+
"loss": 5.9643,
|
| 2696 |
+
"step": 872
|
| 2697 |
+
},
|
| 2698 |
+
{
|
| 2699 |
+
"epoch": 2.99,
|
| 2700 |
+
"learning_rate": 5.136986301369863e-07,
|
| 2701 |
+
"loss": 5.7904,
|
| 2702 |
+
"step": 874
|
| 2703 |
+
},
|
| 2704 |
+
{
|
| 2705 |
+
"epoch": 3.0,
|
| 2706 |
+
"learning_rate": 3.9954337899543377e-07,
|
| 2707 |
+
"loss": 5.9321,
|
| 2708 |
+
"step": 876
|
| 2709 |
+
},
|
| 2710 |
+
{
|
| 2711 |
+
"epoch": 3.0,
|
| 2712 |
+
"step": 876,
|
| 2713 |
+
"total_flos": 8.368943131860664e+17,
|
| 2714 |
+
"train_loss": 6.569673863720132,
|
| 2715 |
+
"train_runtime": 1471.7945,
|
| 2716 |
+
"train_samples_per_second": 4.753,
|
| 2717 |
+
"train_steps_per_second": 0.595
|
| 2718 |
+
}
|
| 2719 |
+
],
|
| 2720 |
+
"max_steps": 876,
|
| 2721 |
+
"num_train_epochs": 3,
|
| 2722 |
+
"total_flos": 8.368943131860664e+17,
|
| 2723 |
+
"trial_name": null,
|
| 2724 |
+
"trial_params": null
|
| 2725 |
+
}
|
training_args.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:33bde102253d7e820fef744fab4f7922c36508e4f38c4994d8a8022cc5871b9a
|
| 3 |
+
size 4155
|