trocr fine-tune, epoch 8, cer=0.016840019811788013

config.json

@@ -0,0 +1,98 @@
+{
+  "architectures": [
+    "VisionEncoderDecoderModel"
+  ],
+  "decoder": {
+    "_name_or_path": "",
+    "activation_dropout": 0.0,
+    "activation_function": "gelu",
+    "add_cross_attention": true,
+    "architectures": null,
+    "attention_dropout": 0.0,
+    "bos_token_id": 0,
+    "chunk_size_feed_forward": 0,
+    "classifier_dropout": 0.0,
+    "cross_attention_hidden_size": 1024,
+    "d_model": 1024,
+    "decoder_attention_heads": 16,
+    "decoder_ffn_dim": 4096,
+    "decoder_layerdrop": 0.0,
+    "decoder_layers": 12,
+    "decoder_start_token_id": 2,
+    "dropout": 0.1,
+    "dtype": "float32",
+    "eos_token_id": 2,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "init_std": 0.02,
+    "is_decoder": true,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layernorm_embedding": true,
+    "max_position_embeddings": 512,
+    "model_type": "trocr",
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "pad_token_id": 1,
+    "problem_type": null,
+    "return_dict": true,
+    "scale_embedding": false,
+    "tie_word_embeddings": true,
+    "use_cache": false,
+    "use_learned_position_embeddings": true,
+    "vocab_size": 30000
+  },
+  "decoder_layerdrop": 0.16239780813448107,
+  "decoder_start_token_id": 2,
+  "dtype": "float32",
+  "encoder": {
+    "_name_or_path": "",
+    "architectures": null,
+    "attention_probs_dropout_prob": 0.0,
+    "chunk_size_feed_forward": 0,
+    "dtype": "float32",
+    "encoder_stride": 16,
+    "hidden_act": "gelu",
+    "hidden_dropout_prob": 0.0,
+    "hidden_size": 1024,
+    "id2label": {
+      "0": "LABEL_0",
+      "1": "LABEL_1"
+    },
+    "image_size": 384,
+    "initializer_range": 0.02,
+    "intermediate_size": 4096,
+    "is_encoder_decoder": false,
+    "label2id": {
+      "LABEL_0": 0,
+      "LABEL_1": 1
+    },
+    "layer_norm_eps": 1e-12,
+    "model_type": "vit",
+    "num_attention_heads": 16,
+    "num_channels": 3,
+    "num_hidden_layers": 24,
+    "output_attentions": false,
+    "output_hidden_states": false,
+    "patch_size": 16,
+    "pooler_act": "tanh",
+    "pooler_output_size": 1024,
+    "problem_type": null,
+    "qkv_bias": false,
+    "return_dict": true
+  },
+  "encoder_layerdrop": 0.16239780813448107,
+  "eos_token_id": 3,
+  "is_encoder_decoder": true,
+  "model_type": "vision-encoder-decoder",
+  "pad_token_id": 0,
+  "tie_word_embeddings": false,
+  "transformers_version": "5.0.0",
+  "use_cache": false,
+  "vocab_size": 30000
+}

generation_config.json

@@ -0,0 +1,36 @@
+{
+  "_from_model_config": false,
+  "assistant_confidence_threshold": 0.4,
+  "assistant_lookbehind": 10,
+  "bos_token_id": 0,
+  "decoder_start_token_id": 2,
+  "diversity_penalty": 0.0,
+  "do_sample": false,
+  "early_stopping": true,
+  "encoder_no_repeat_ngram_size": 0,
+  "encoder_repetition_penalty": 1.0,
+  "eos_token_id": 3,
+  "epsilon_cutoff": 0.0,
+  "eta_cutoff": 0.0,
+  "length_penalty": 1.0,
+  "max_length": 32,
+  "min_length": 0,
+  "no_repeat_ngram_size": 0,
+  "num_assistant_tokens": 20,
+  "num_assistant_tokens_schedule": "constant",
+  "num_beam_groups": 1,
+  "num_beams": 4,
+  "num_return_sequences": 1,
+  "output_scores": false,
+  "pad_token_id": 0,
+  "remove_invalid_values": false,
+  "repetition_penalty": 1.0,
+  "return_dict_in_generate": false,
+  "target_lookbehind": 10,
+  "temperature": 1.0,
+  "top_k": 50,
+  "top_p": 1.0,
+  "transformers_version": "5.0.0",
+  "typical_p": 1.0,
+  "use_cache": false
+}

model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:9addc3d6ef7d29628933d35bbd1cbb7e7da24d346505971bc10791ca12b077af
+size 2149979832

processor_config.json

@@ -0,0 +1,26 @@
+{
+  "image_processor": {
+    "data_format": "channels_first",
+    "do_normalize": true,
+    "do_rescale": true,
+    "do_resize": true,
+    "image_mean": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "image_processor_type": "ViTImageProcessorFast",
+    "image_std": [
+      0.5,
+      0.5,
+      0.5
+    ],
+    "resample": 2,
+    "rescale_factor": 0.00392156862745098,
+    "size": {
+      "height": 384,
+      "width": 384
+    }
+  },
+  "processor_class": "TrOCRProcessor"
+}

tokenizer_config.json

@@ -0,0 +1,20 @@
+{
+  "backend": "tokenizers",
+  "clean_up_tokenization_spaces": true,
+  "cls_token": "[CLS]",
+  "do_basic_tokenize": true,
+  "do_lower_case": false,
+  "full_tokenizer_file": null,
+  "is_local": false,
+  "mask_token": "[MASK]",
+  "model_max_length": 1000000000000000019884624838656,
+  "model_specific_special_tokens": {},
+  "never_split": null,
+  "pad_token": "[PAD]",
+  "processor_class": "TrOCRProcessor",
+  "sep_token": "[SEP]",
+  "strip_accents": null,
+  "tokenize_chinese_chars": true,
+  "tokenizer_class": "TokenizersBackend",
+  "unk_token": "[UNK]"
+}