epoch 220-235

config.json

@@ -1,11 +1,11 @@
 {
-  "_commit_hash": "86f68fe78ff0811773d8233a005f1746702d4799",
-  "_name_or_path": "tt1225/finetuned-trocr-small-vietnamese-nom",
+  "_commit_hash": "620d114582d85d027705435e967dca3667a79825",
+  "_name_or_path": "tt1225/finetuned-trocr-tiny-vietnamese-nom",
   "architectures": [
     "VisionEncoderDecoderModel"
   ],
   "decoder": {
-    "_name_or_path": "minhtoan/roberta-masked-lm-vietnamese-nom",
+    "_name_or_path": "nxquang-al/roberta-small-masked-lm-vietnamese-nom",
     "add_cross_attention": true,
     "architectures": [
       "RobertaForMaskedLM"
@@ -27,10 +27,9 @@
     "finetuning_task": null,
     "forced_bos_token_id": null,
     "forced_eos_token_id": null,
-    "gradient_checkpointing": false,
     "hidden_act": "gelu",
     "hidden_dropout_prob": 0.1,
-    "hidden_size": 768,
+    "hidden_size": 256,
     "id2label": {
       "0": "LABEL_0",
       "1": "LABEL_1"
@@ -46,14 +45,14 @@
     "layer_norm_eps": 1e-12,
     "length_penalty": 1.0,
     "max_length": 20,
-    "max_position_embeddings": 514,
+    "max_position_embeddings": 512,
     "min_length": 0,
     "model_type": "roberta",
     "no_repeat_ngram_size": 0,
-    "num_attention_heads": 12,
+    "num_attention_heads": 4,
     "num_beam_groups": 1,
     "num_beams": 1,
-    "num_hidden_layers": 6,
+    "num_hidden_layers": 12,
     "num_return_sequences": 1,
     "output_attentions": false,
     "output_hidden_states": false,
@@ -79,7 +78,7 @@
     "top_p": 1.0,
     "torch_dtype": "float32",
     "torchscript": false,
-    "transformers_version": "4.30.2",
+    "transformers_version": "4.30.1",
     "type_vocab_size": 1,
     "typical_p": 1.0,
     "use_bfloat16": false,
@@ -89,12 +88,12 @@
   "decoder_start_token_id": 5,
   "early_stopping": true,
   "encoder": {
-    "_name_or_path": "facebook/deit-tiny-distilled-patch16-224",
+    "_name_or_path": "facebook/deit-small-patch16-224",
     "add_cross_attention": false,
     "architectures": [
-      "DeiTForImageClassificationWithTeacher"
+      "ViTForImageClassification"
     ],
-    "attention_probs_dropout_prob": 0.5,
+    "attention_probs_dropout_prob": 0.1,
     "bad_words_ids": null,
     "begin_suppress_tokens": null,
     "bos_token_id": null,
@@ -113,7 +112,7 @@
     "forced_eos_token_id": null,
     "hidden_act": "gelu",
     "hidden_dropout_prob": 0.0,
-    "hidden_size": 192,
+    "hidden_size": 384,
     "id2label": {
       "0": "tench, Tinca tinca",
       "1": "goldfish, Carassius auratus",
@@ -1118,7 +1117,7 @@
     },
     "image_size": 224,
     "initializer_range": 0.02,
-    "intermediate_size": 768,
+    "intermediate_size": 1536,
     "is_decoder": false,
     "is_encoder_decoder": false,
     "label2id": {
@@ -2126,9 +2125,9 @@
     "length_penalty": 1.0,
     "max_length": 20,
     "min_length": 0,
-    "model_type": "deit",
+    "model_type": "vit",
     "no_repeat_ngram_size": 0,
-    "num_attention_heads": 3,
+    "num_attention_heads": 6,
     "num_beam_groups": 1,
     "num_beams": 1,
     "num_channels": 3,
@@ -2159,7 +2158,7 @@
     "top_p": 1.0,
     "torch_dtype": null,
     "torchscript": false,
-    "transformers_version": "4.30.2",
+    "transformers_version": "4.30.1",
     "typical_p": 1.0,
     "use_bfloat16": false
   },

generation_config.json

@@ -9,5 +9,5 @@
   "no_repeat_ngram_size": 3,
   "num_beams": 4,
   "pad_token_id": 6198,
-  "transformers_version": "4.30.2"
+  "transformers_version": "4.30.1"
 }

pytorch_model.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:bb1591d78d4b0c5b0540b19f6880898f1f2c4cee765ff9818d53b8f86499db37
-size 349128597
+oid sha256:a95b2610553c7cffc1356ccd94814585f16dde224ea8b00754bbd2d58dd665df
+size 221492737