optimum-intel-internal-testing/tiny-random-VisionEncoderDecoderModel-vit-gpt2
020k
1{2 "_commit_hash": null,3 "_name_or_path": "temp/dummy/vision-encoder-decoder/VisionEncoderDecoderModel-vit-gpt2",4 "architectures": [5 "VisionEncoderDecoderModel"6 ],7 "decoder": {8 "_name_or_path": "/tmp/tmp7f9pdu67/decoder/GPT2LMHeadModel",9 "activation_function": "gelu",10 "add_cross_attention": true,11 "architectures": [12 "GPT2LMHeadModel"13 ],14 "attn_pdrop": 0.1,15 "bad_words_ids": null,16 "begin_suppress_tokens": null,17 "bos_token_id": 0,18 "chunk_size_feed_forward": 0,19 "cross_attention_hidden_size": null,20 "decoder_start_token_id": null,21 "diversity_penalty": 0.0,22 "do_sample": false,23 "early_stopping": false,24 "embd_pdrop": 0.1,25 "encoder_no_repeat_ngram_size": 0,26 "eos_token_id": 0,27 "exponential_decay_length_penalty": null,28 "finetuning_task": null,29 "forced_bos_token_id": null,30 "forced_eos_token_id": null,31 "gradient_checkpointing": false,32 "id2label": {33 "0": "LABEL_0",34 "1": "LABEL_1"35 },36 "initializer_range": 0.02,37 "is_decoder": true,38 "is_encoder_decoder": false,39 "label2id": {40 "LABEL_0": 0,41 "LABEL_1": 142 },43 "layer_norm_epsilon": 1e-05,44 "length_penalty": 1.0,45 "max_length": 20,46 "min_length": 0,47 "model_type": "gpt2",48 "n_embd": 32,49 "n_head": 4,50 "n_inner": 37,51 "n_layer": 5,52 "n_positions": 512,53 "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": 1023,61 "prefix": null,62 "problem_type": null,63 "pruned_heads": {},64 "remove_invalid_values": false,65 "reorder_and_upcast_attn": false,66 "repetition_penalty": 1.0,67 "resid_pdrop": 0.1,68 "return_dict": true,69 "return_dict_in_generate": false,70 "scale_attn_by_inverse_layer_idx": false,71 "scale_attn_weights": true,72 "sep_token_id": null,73 "summary_activation": null,74 "summary_first_dropout": 0.1,75 "summary_proj_to_labels": true,76 "summary_type": "cls_index",77 "summary_use_proj": true,78 "suppress_tokens": null,79 "task_specific_params": null,80 "temperature": 1.0,81 "tf_legacy_loss": false,82 "tie_encoder_decoder": false,83 "tie_word_embeddings": true,84 "tokenizer_class": null,85 "top_k": 50,86 "top_p": 1.0,87 "torch_dtype": "float32",88 "torchscript": false,89 "transformers_version": "4.25.0.dev0",90 "type_vocab_size": 16,91 "typical_p": 1.0,92 "use_bfloat16": false,93 "use_cache": true,94 "vocab_size": 102495 },96 "encoder": {97 "_name_or_path": "/tmp/tmp7f9pdu67/encoder/ViTModel",98 "add_cross_attention": false,99 "architectures": [100 "ViTModel"101 ],102 "attention_probs_dropout_prob": 0.1,103 "bad_words_ids": null,104 "begin_suppress_tokens": null,105 "bos_token_id": null,106 "chunk_size_feed_forward": 0,107 "cross_attention_hidden_size": null,108 "decoder_start_token_id": null,109 "diversity_penalty": 0.0,110 "do_sample": false,111 "early_stopping": false,112 "encoder_no_repeat_ngram_size": 0,113 "encoder_stride": 2,114 "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_act": "gelu",120 "hidden_dropout_prob": 0.1,121 "hidden_size": 32,122 "id2label": {123 "0": "LABEL_0",124 "1": "LABEL_1"125 },126 "image_size": 30,127 "initializer_range": 0.02,128 "intermediate_size": 37,129 "is_decoder": false,130 "is_encoder_decoder": false,131 "label2id": {132 "LABEL_0": 0,133 "LABEL_1": 1134 },135 "layer_norm_eps": 1e-12,136 "length_penalty": 1.0,137 "max_length": 20,138 "min_length": 0,139 "model_type": "vit",140 "no_repeat_ngram_size": 0,141 "num_attention_heads": 4,142 "num_beam_groups": 1,143 "num_beams": 1,144 "num_channels": 3,145 "num_hidden_layers": 5,146 "num_return_sequences": 1,147 "output_attentions": false,148 "output_hidden_states": false,149 "output_scores": false,150 "pad_token_id": null,151 "patch_size": 2,152 "prefix": null,153 "problem_type": null,154 "pruned_heads": {},155 "qkv_bias": true,156 "remove_invalid_values": false,157 "repetition_penalty": 1.0,158 "return_dict": true,159 "return_dict_in_generate": false,160 "sep_token_id": null,161 "suppress_tokens": null,162 "task_specific_params": null,163 "temperature": 1.0,164 "tf_legacy_loss": false,165 "tie_encoder_decoder": false,166 "tie_word_embeddings": true,167 "tokenizer_class": null,168 "top_k": 50,169 "top_p": 1.0,170 "torch_dtype": "float32",171 "torchscript": false,172 "transformers_version": "4.25.0.dev0",173 "typical_p": 1.0,174 "use_bfloat16": false175 },176 "is_encoder_decoder": true,177 "model_type": "vision-encoder-decoder",178 "tie_word_embeddings": false,179 "torch_dtype": "float32",180 "transformers_version": null181}182 