google/pix2struct-base

1{2  "_commit_hash": null,3  "_name_or_path": "/home/younes_huggingface_co/code/pix2struct-checkpoints/pix2struct-base/",4  "architectures": [5    "Pix2StructForConditionalGeneration"6  ],7  "decoder_start_token_id": 0,8  "initializer_factor": 1.0,9  "initializer_range": 0.02,10  "is_vqa": false,11  "model_type": "pix2struct",12  "pad_token_id": 0,13  "text_config": {14    "_name_or_path": "",15    "add_cross_attention": false,16    "architectures": null,17    "bad_words_ids": null,18    "begin_suppress_tokens": null,19    "bos_token_id": null,20    "chunk_size_feed_forward": 0,21    "cross_attention_hidden_size": null,22    "d_ff": 2048,23    "d_kv": 64,24    "decoder_start_token_id": 0,25    "dense_act_fn": "gelu_new",26    "diversity_penalty": 0.0,27    "do_sample": false,28    "dropout_rate": 0.2,29    "early_stopping": false,30    "encoder_hidden_size": 768,31    "encoder_no_repeat_ngram_size": 0,32    "eos_token_id": 1,33    "exponential_decay_length_penalty": null,34    "finetuning_task": null,35    "forced_bos_token_id": null,36    "forced_eos_token_id": null,37    "hidden_size": 768,38    "id2label": {39      "0": "LABEL_0",40      "1": "LABEL_1"41    },42    "initializer_factor": 1.0,43    "initializer_range": 0.02,44    "is_decoder": true,45    "is_encoder_decoder": false,46    "label2id": {47      "LABEL_0": 0,48      "LABEL_1": 149    },50    "layer_norm_epsilon": 1e-06,51    "length_penalty": 1.0,52    "max_length": 20,53    "min_length": 0,54    "model_type": "pix2struct_text_model",55    "no_repeat_ngram_size": 0,56    "num_beam_groups": 1,57    "num_beams": 1,58    "num_heads": 12,59    "num_layers": 12,60    "num_return_sequences": 1,61    "output_attentions": false,62    "output_hidden_states": false,63    "output_scores": false,64    "pad_token_id": 0,65    "prefix": null,66    "problem_type": null,67    "pruned_heads": {},68    "relative_attention_max_distance": 128,69    "relative_attention_num_buckets": 32,70    "remove_invalid_values": false,71    "repetition_penalty": 1.0,72    "return_dict": true,73    "return_dict_in_generate": false,74    "sep_token_id": null,75    "suppress_tokens": null,76    "task_specific_params": null,77    "temperature": 1.0,78    "tf_legacy_loss": false,79    "tie_encoder_decoder": false,80    "tie_word_embeddings": false,81    "tokenizer_class": null,82    "top_k": 50,83    "top_p": 1.0,84    "torch_dtype": null,85    "torchscript": false,86    "transformers_version": "4.27.0.dev0",87    "typical_p": 1.0,88    "use_bfloat16": false,89    "use_cache": false,90    "vocab_size": 5024491  },92  "torch_dtype": "float32",93  "tie_word_embeddings": false,94  "transformers_version": null,95  "vision_config": {96    "_name_or_path": "",97    "add_cross_attention": false,98    "architectures": null,99    "attention_dropout": 0.2,100    "bad_words_ids": null,101    "begin_suppress_tokens": null,102    "bos_token_id": null,103    "chunk_size_feed_forward": 0,104    "cross_attention_hidden_size": null,105    "d_ff": 2048,106    "d_kv": 64,107    "decoder_start_token_id": null,108    "dense_act_fn": "gelu_new",109    "diversity_penalty": 0.0,110    "do_sample": false,111    "dropout_rate": 0.2,112    "early_stopping": false,113    "encoder_no_repeat_ngram_size": 0,114    "eos_token_id": null,115    "exponential_decay_length_penalty": null,116    "finetuning_task": null,117    "forced_bos_token_id": null,118    "forced_eos_token_id": null,119    "hidden_dropout_prob": 0.2,120    "hidden_size": 768,121    "id2label": {122      "0": "LABEL_0",123      "1": "LABEL_1"124    },125    "initializer_factor": 1.0,126    "initializer_range": 0.02,127    "is_decoder": false,128    "is_encoder_decoder": false,129    "label2id": {130      "LABEL_0": 0,131      "LABEL_1": 1132    },133    "layer_norm_bias": false,134    "layer_norm_eps": 1e-06,135    "length_penalty": 1.0,136    "max_length": 20,137    "min_length": 0,138    "model_type": "pix2struct_vision_model",139    "no_repeat_ngram_size": 0,140    "num_attention_heads": 12,141    "num_beam_groups": 1,142    "num_beams": 1,143    "num_channels": 3,144    "num_hidden_layers": 12,145    "num_return_sequences": 1,146    "output_attentions": false,147    "output_hidden_states": false,148    "output_scores": false,149    "pad_token_id": null,150    "patch_embed_hidden_size": 768,151    "patch_size": 16,152    "prefix": null,153    "problem_type": null,154    "projection_dim": 768,155    "pruned_heads": {},156    "relative_attention_max_distance": 128,157    "relative_attention_num_buckets": 32,158    "remove_invalid_values": false,159    "repetition_penalty": 1.0,160    "return_dict": true,161    "return_dict_in_generate": false,162    "sep_token_id": null,163    "seq_len": 4096,164    "suppress_tokens": null,165    "task_specific_params": null,166    "temperature": 1.0,167    "tf_legacy_loss": false,168    "tie_encoder_decoder": false,169    "tie_word_embeddings": true,170    "tokenizer_class": null,171    "top_k": 50,172    "top_p": 1.0,173    "torch_dtype": null,174    "torchscript": false,175    "transformers_version": "4.27.0.dev0",176    "typical_p": 1.0,177    "use_bfloat16": false178  }179}180