Training in progress, epoch 2, checkpoint

checkpoint-562/adapter_config.json

@@ -1,6 +1,6 @@
 {
   "auto_mapping": null,
-  "base_model_name_or_path": "Qwen/Qwen1.5-4B",
+  "base_model_name_or_path": "meta-llama/Llama-2-7b-hf",
   "bias": "none",
   "fan_in_fan_out": false,
   "inference_mode": true,

checkpoint-562/adapter_model.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e46bbbbc9c48f0013a5fea0221dd9f25cbb7dbfd59698d7cae0b7c08aa79846d
+size 143269386

checkpoint-562/added_tokens.json

@@ -1,5 +1,5 @@
 {
-  "<|endoftext|>": 151643,
-  "<|im_end|>": 151645,
-  "<|im_start|>": 151644
+  "</s>": 2,
+  "<s>": 1,
+  "<unk>": 0
 }

checkpoint-562/optimizer.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:ec5b7fa3d0085997ca2891e0a49552bf064aadfacd0b249987559f37ef0677a4
-size 224537202
+oid sha256:01578dc4c6d97c3592575b26289a56750f44fe496b933a065603b4eb77e20ac8
+size 286585234

checkpoint-562/rng_state_0.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:51d7fb2f8ab04126d48204957a0b3d84fda74bf7b90934e2d3fb4b78e886cc6d
+oid sha256:268d59d8605ab630eeaae0ae1db78c65c2a68bc190791a4aba79c03f74f8edd6
 size 15024

checkpoint-562/rng_state_1.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:3836002a513d3ab89f8633009eecda644fb6becb6ad8bda29beee1013c38549d
+oid sha256:c2c057368c845d021c9b861384914ffd6ab202ed4ba06e9fbe94c2fb6f85864e
 size 15024

checkpoint-562/rng_state_2.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7b39cf73c26392f71400531750c763aafed5df27bd2e82abaefe623a1fac3c17
+oid sha256:beb8c45fd53f284800880a68ff2f247360dff095821e097611889efc50ae287b
 size 15024

checkpoint-562/rng_state_3.pth

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:92defd84116d80894fbae874a63da49b544386aa20a08547afa9240b9e2748a6
+oid sha256:d0837e0522a0087a44ffdeb00e411b6f0e3579e541ffac79fc8352be29cbb296
 size 15024

checkpoint-562/special_tokens_map.json

@@ -1,14 +1,6 @@
 {
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>"
-  ],
-  "eos_token": {
-    "content": "<|endoftext|>",
-    "lstrip": false,
-    "normalized": false,
-    "rstrip": false,
-    "single_word": false
-  },
-  "pad_token": "<|endoftext|>"
+  "bos_token": "<s>",
+  "eos_token": "</s>",
+  "pad_token": "</s>",
+  "unk_token": "<unk>"
 }

checkpoint-562/tokenizer.model

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9e556afd44213b6bd1be2b850ebbbd98f5481437a8021afaf58ee7fb1818d347
+size 499723

checkpoint-562/tokenizer_config.json

@@ -1,24 +1,23 @@
 {
-  "add_prefix_space": false,
   "added_tokens_decoder": {
-    "151643": {
-      "content": "<|endoftext|>",
+    "0": {
+      "content": "<unk>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
       "single_word": false,
       "special": true
     },
-    "151644": {
-      "content": "<|im_start|>",
+    "1": {
+      "content": "<s>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
       "single_word": false,
       "special": true
     },
-    "151645": {
-      "content": "<|im_end|>",
+    "2": {
+      "content": "</s>",
       "lstrip": false,
       "normalized": false,
       "rstrip": false,
@@ -26,18 +25,16 @@
       "special": true
     }
   },
-  "additional_special_tokens": [
-    "<|im_start|>",
-    "<|im_end|>"
-  ],
-  "bos_token": null,
-  "chat_template": "{% for message in messages %}{% if loop.first and messages[0]['role'] != 'system' %}{{ '<|im_start|>system\nYou are a helpful assistant<|im_end|>\n' }}{% endif %}{{'<|im_start|>' + message['role'] + '\n' + message['content'] + '<|im_end|>' + '\n'}}{% endfor %}{% if add_generation_prompt %}{{ '<|im_start|>assistant\n' }}{% endif %}",
+  "additional_special_tokens": [],
+  "bos_token": "<s>",
   "clean_up_tokenization_spaces": false,
-  "eos_token": "<|endoftext|>",
-  "errors": "replace",
-  "model_max_length": 32768,
-  "pad_token": "<|endoftext|>",
-  "split_special_tokens": false,
-  "tokenizer_class": "Qwen2Tokenizer",
-  "unk_token": null
+  "eos_token": "</s>",
+  "legacy": false,
+  "model_max_length": 1000000000000000019884624838656,
+  "pad_token": "</s>",
+  "padding_side": "left",
+  "sp_model_kwargs": {},
+  "tokenizer_class": "LlamaTokenizer",
+  "unk_token": "<unk>",
+  "use_default_system_prompt": true
 }

checkpoint-562/trainer_state.json

@@ -9,117 +9,98 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.5333333333333333,
-      "grad_norm": 0.2081739604473114,
+      "epoch": 0.53,
       "learning_rate": 0.0005,
-      "loss": 1.7461,
+      "loss": 1.2462,
       "step": 100
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_accuracy": 0.6093004484304932,
-      "eval_loss": 1.6695144176483154,
-      "eval_runtime": 6.0389,
-      "eval_samples_per_second": 82.796,
-      "eval_steps_per_second": 10.432,
+      "epoch": 1.0,
+      "eval_accuracy": 0.6696470588235294,
+      "eval_loss": 1.1816030740737915,
+      "eval_runtime": 7.5359,
+      "eval_samples_per_second": 66.349,
+      "eval_steps_per_second": 8.36,
       "step": 187
     },
     {
-      "epoch": 0.9973333333333333,
-      "eval_exact_match": 15.8,
-      "eval_f1": 24.64418803418805,
-      "eval_qa_bleu": 12.64258947770803,
-      "eval_qa_exact_match": 0.116,
-      "eval_recite_bleu": 12.676034356605994,
+      "epoch": 1.0,
+      "eval_exact_match": 28.8,
+      "eval_f1": 38.19841269841268,
+      "eval_qa_bleu": 24.052327144564227,
+      "eval_qa_exact_match": 0.216,
+      "eval_recite_bleu": 16.828353493471358,
       "eval_recite_exact_match": 0.002,
       "step": 187
     },
     {
-      "epoch": 1.0666666666666667,
-      "grad_norm": 0.2944607138633728,
+      "epoch": 1.07,
       "learning_rate": 0.0005,
-      "loss": 1.6853,
+      "loss": 1.1798,
       "step": 200
     },
     {
       "epoch": 1.6,
-      "grad_norm": 0.40688613057136536,
       "learning_rate": 0.0005,
-      "loss": 1.4286,
+      "loss": 0.9978,
       "step": 300
     },
     {
       "epoch": 2.0,
-      "eval_accuracy": 0.6086547085201793,
-      "eval_loss": 1.6994106769561768,
-      "eval_runtime": 6.3284,
-      "eval_samples_per_second": 79.009,
-      "eval_steps_per_second": 9.955,
+      "eval_accuracy": 0.6694823529411764,
+      "eval_loss": 1.196727991104126,
+      "eval_runtime": 6.6634,
+      "eval_samples_per_second": 75.037,
+      "eval_steps_per_second": 9.455,
       "step": 375
     },
     {
       "epoch": 2.0,
-      "eval_exact_match": 13.6,
-      "eval_f1": 22.022914862914874,
-      "eval_qa_bleu": 11.191170451488254,
-      "eval_qa_exact_match": 0.098,
-      "eval_recite_bleu": 13.062440471503999,
+      "eval_exact_match": 26.4,
+      "eval_f1": 35.01285714285714,
+      "eval_qa_bleu": 22.554598366442242,
+      "eval_qa_exact_match": 0.202,
+      "eval_recite_bleu": 16.66380860754827,
       "eval_recite_exact_match": 0.008,
       "step": 375
     },
     {
-      "epoch": 2.1333333333333333,
-      "grad_norm": 0.4661327600479126,
+      "epoch": 2.13,
       "learning_rate": 0.0005,
-      "loss": 1.3391,
+      "loss": 0.8986,
       "step": 400
     },
     {
-      "epoch": 2.6666666666666665,
-      "grad_norm": 0.5310370326042175,
+      "epoch": 2.67,
       "learning_rate": 0.0005,
-      "loss": 1.0309,
+      "loss": 0.6981,
       "step": 500
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_accuracy": 0.6053452914798206,
-      "eval_loss": 1.812237024307251,
-      "eval_runtime": 5.6113,
-      "eval_samples_per_second": 89.107,
-      "eval_steps_per_second": 11.227,
+      "epoch": 3.0,
+      "eval_accuracy": 0.6646196078431372,
+      "eval_loss": 1.2825865745544434,
+      "eval_runtime": 6.6803,
+      "eval_samples_per_second": 74.847,
+      "eval_steps_per_second": 9.431,
       "step": 562
     },
     {
-      "epoch": 2.997333333333333,
-      "eval_exact_match": 12.8,
-      "eval_f1": 21.852727272727286,
-      "eval_qa_bleu": 11.536967687478375,
-      "eval_qa_exact_match": 0.1,
-      "eval_recite_bleu": 13.977710455440388,
-      "eval_recite_exact_match": 0.008,
+      "epoch": 3.0,
+      "eval_exact_match": 25.6,
+      "eval_f1": 34.62047619047618,
+      "eval_qa_bleu": 19.019126060056248,
+      "eval_qa_exact_match": 0.202,
+      "eval_recite_bleu": 15.866754500529721,
+      "eval_recite_exact_match": 0.01,
       "step": 562
     }
   ],
   "logging_steps": 100,
   "max_steps": 9350,
-  "num_input_tokens_seen": 0,
   "num_train_epochs": 50,
   "save_steps": 500,
-  "stateful_callbacks": {
-    "TrainerControl": {
-      "args": {
-        "should_epoch_stop": false,
-        "should_evaluate": false,
-        "should_log": false,
-        "should_save": true,
-        "should_training_stop": false
-      },
-      "attributes": {}
-    }
-  },
-  "total_flos": 5.665963591807795e+16,
-  "train_batch_size": 1,
+  "total_flos": 1.3031708836233216e+17,
   "trial_name": null,
   "trial_params": null
 }

checkpoint-562/training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:7b6061b8d6135a3f759c6b956a7afa194ea4a0f719fc2e4eb293c4d6c15c2590
-size 5304
+oid sha256:a28079dc7535e79a5e324b19b9af4575c5cbbdcf35a6916d09f5d38aff3423a1
+size 4728