FluidInference/verdict-coreml

1{2  "architecture_type": "uni-encoder",3  "architectures": [4    "GLiClassModel"5  ],6  "class_token_index": 50368,7  "class_token_pooling": "first",8  "contrastive_loss_coef": 0.0,9  "dropout": 0.1,10  "dtype": "float32",11  "embed_class_token": true,12  "encoder_config": {13    "_attn_implementation_autoset": false,14    "_name_or_path": "answerdotai/ModernBERT-base",15    "add_cross_attention": false,16    "architectures": [17      "ModernBertForMaskedLM"18    ],19    "attention_bias": false,20    "attention_dropout": 0.0,21    "bos_token_id": 50281,22    "classifier_activation": "gelu",23    "classifier_bias": false,24    "classifier_dropout": 0.0,25    "classifier_pooling": "mean",26    "cls_token_id": 50281,27    "cross_attention_hidden_size": null,28    "decoder_bias": true,29    "decoder_start_token_id": null,30    "deterministic_flash_attn": false,31    "dtype": "float32",32    "embedding_dropout": 0.0,33    "eos_token_id": 50282,34    "finetuning_task": null,35    "global_attn_every_n_layers": 3,36    "gradient_checkpointing": false,37    "hidden_activation": "gelu",38    "hidden_size": 768,39    "initializer_cutoff_factor": 2.0,40    "initializer_range": 0.02,41    "intermediate_size": 1152,42    "is_decoder": false,43    "layer_norm_eps": 1e-05,44    "layer_types": [45      "full_attention",46      "sliding_attention",47      "sliding_attention",48      "full_attention",49      "sliding_attention",50      "sliding_attention",51      "full_attention",52      "sliding_attention",53      "sliding_attention",54      "full_attention",55      "sliding_attention",56      "sliding_attention",57      "full_attention",58      "sliding_attention",59      "sliding_attention",60      "full_attention",61      "sliding_attention",62      "sliding_attention",63      "full_attention",64      "sliding_attention",65      "sliding_attention",66      "full_attention"67    ],68    "local_attention": 128,69    "max_position_embeddings": 8192,70    "mlp_bias": false,71    "mlp_dropout": 0.0,72    "model_type": "modernbert",73    "norm_bias": false,74    "norm_eps": 1e-05,75    "num_attention_heads": 12,76    "num_hidden_layers": 22,77    "pad_token_id": 50283,78    "position_embedding_type": "absolute",79    "prefix": null,80    "pruned_heads": {},81    "repad_logits_with_grad": false,82    "rope_parameters": {83      "full_attention": {84        "rope_theta": 160000.0,85        "rope_type": "default"86      },87      "sliding_attention": {88        "rope_theta": 10000.0,89        "rope_type": "default"90      }91    },92    "sep_token_id": 50282,93    "sparse_pred_ignore_index": -100,94    "sparse_prediction": false,95    "task_specific_params": null,96    "tf_legacy_loss": false,97    "tie_encoder_decoder": false,98    "tie_word_embeddings": true,99    "tokenizer_class": null,100    "torchscript": false,101    "use_bfloat16": false,102    "vocab_size": 50370103  },104  "encoder_layer_id": -1,105  "encoder_model_name": "answerdotai/ModernBERT-base",106  "example_token_index": 50372,107  "extract_text_features": false,108  "focal_loss_alpha": -1,109  "focal_loss_gamma": -1,110  "focal_loss_reduction": null,111  "hidden_size": 768,112  "ignore_index": -100,113  "initializer_range": 0.03,114  "label_model_config": null,115  "label_model_name": null,116  "layer_wise": false,117  "logit_scale_init_value": 2.6592,118  "max_labels_alloc": "dynamic",119  "max_num_classes": 25,120  "model_type": "GLiClass",121  "normalize_features": false,122  "pad_token_id": 50283,123  "pooling_strategy": "first",124  "problem_type": "single_label_classification",125  "projector_hidden_act": "gelu",126  "prompt_first": true,127  "scorer_attn_dropout": 0.1,128  "scorer_encoder_num_layers": 2,129  "scorer_mlp_hidden_size": 1024,130  "scorer_num_heads": 16,131  "scorer_type": "simple",132  "squeeze_layers": false,133  "text_token_index": 50369,134  "transformers_version": "5.17.0",135  "use_lstm": false,136  "use_segment_embeddings": false,137  "vocab_size": 50370138}139