fleonce/iter-scierc-deberta-large
0256
1{2 "_name_or_path": "models/fleonce/iter-scierc-deberta-large",3 "activation_fn": "relu",4 "architectures": [5 "ITERForRelationExtraction"6 ],7 "d_ff": 4096,8 "d_model": 1024,9 "dataset": "scierc",10 "dropout": 0.051157237106187456,11 "entity_types": [12 "Task",13 "Method",14 "Material",15 "OtherScientificTerm",16 "Metric",17 "Generic"18 ],19 "features": 74528,20 "link_types": [21 "Used-for",22 "Feature-of",23 "Hyponym-of",24 "Evaluate-for",25 "Part-of",26 "Compare",27 "Conjunction"28 ],29 "max_length": 512,30 "max_nest_depth": 3,31 "model_type": "iter",32 "num_links": 7,33 "num_types": 7,34 "threshold": 0.5,35 "torch_dtype": "float32",36 "transformer_config": {37 "_name_or_path": "microsoft/deberta-v3-large",38 "architectures": null,39 "attention_probs_dropout_prob": 0.1,40 "decoder_start_token_id": null,41 "eos_token_id": null,42 "hidden_act": "gelu",43 "hidden_dropout_prob": 0.1,44 "hidden_size": 1024,45 "initializer_range": 0.02,46 "intermediate_size": 4096,47 "is_encoder_decoder": false,48 "layer_norm_eps": 1e-07,49 "max_length": 512,50 "max_position_embeddings": 512,51 "max_relative_positions": -1,52 "model_type": "deberta-v2",53 "norm_rel_ebd": "layer_norm",54 "num_attention_heads": 16,55 "num_hidden_layers": 24,56 "pooler_dropout": 0,57 "pooler_hidden_act": "gelu",58 "pooler_hidden_size": 1024,59 "pos_att_type": [60 "p2c",61 "c2p"62 ],63 "position_biased_input": false,64 "position_buckets": 256,65 "relative_attention": true,66 "share_att_key": true,67 "task_specific_params": null,68 "type_vocab_size": 0,69 "vocab_size": 12810070 },71 "transformers_version": "4.37.0",72 "use_bias": false,73 "use_gate": true,74 "use_mlp": true,75 "use_scale": false76}77 