proseph/ctrlv-wav2vec2-tokenizer
026
1{2 "_name_or_path": "facebook/wav2vec2-base",3 "activation_dropout": 0.0,4 "apply_spec_augment": true,5 "architectures": [6 "Wav2Vec2ForCTC"7 ],8 "attention_dropout": 0.1,9 "bos_token_id": 1,10 "classifier_proj_size": 256,11 "codevector_dim": 256,12 "contrastive_logits_temperature": 0.1,13 "conv_bias": false,14 "conv_dim": [15 512,16 512,17 512,18 512,19 512,20 512,21 51222 ],23 "conv_kernel": [24 10,25 3,26 3,27 3,28 3,29 2,30 231 ],32 "conv_stride": [33 5,34 2,35 2,36 2,37 2,38 2,39 240 ],41 "ctc_loss_reduction": "mean",42 "ctc_zero_infinity": false,43 "diversity_loss_weight": 0.1,44 "do_stable_layer_norm": false,45 "eos_token_id": 2,46 "feat_extract_activation": "gelu",47 "feat_extract_norm": "group",48 "feat_proj_dropout": 0.1,49 "feat_quantizer_dropout": 0.0,50 "final_dropout": 0.0,51 "freeze_feat_extract_train": true,52 "hidden_act": "gelu",53 "hidden_dropout": 0.1,54 "hidden_size": 768,55 "initializer_range": 0.02,56 "intermediate_size": 3072,57 "layer_norm_eps": 1e-05,58 "layerdrop": 0.0,59 "mask_channel_length": 10,60 "mask_channel_min_space": 1,61 "mask_channel_other": 0.0,62 "mask_channel_prob": 0.0,63 "mask_channel_selection": "static",64 "mask_feature_length": 10,65 "mask_feature_prob": 0.0,66 "mask_time_length": 10,67 "mask_time_min_space": 1,68 "mask_time_other": 0.0,69 "mask_time_prob": 0.05,70 "mask_time_selection": "static",71 "model_type": "wav2vec2",72 "no_mask_channel_overlap": false,73 "no_mask_time_overlap": false,74 "num_attention_heads": 12,75 "num_codevector_groups": 2,76 "num_codevectors_per_group": 320,77 "num_conv_pos_embedding_groups": 16,78 "num_conv_pos_embeddings": 128,79 "num_feat_extract_layers": 7,80 "num_hidden_layers": 12,81 "num_negatives": 100,82 "pad_token_id": 29,83 "proj_codevector_dim": 256,84 "torch_dtype": "float32",85 "transformers_version": "4.11.3",86 "use_weighted_layer_sum": false,87 "vocab_size": 3288}89 