| { |
| "_name_or_path": "efficient-speech/lite-whisper-medium", |
| "activation_dropout": 0.0, |
| "activation_function": "gelu", |
| "apply_spec_augment": false, |
| "architectures": [ |
| "LiteWhisperForConditionalGeneration" |
| ], |
| "attention_dropout": 0.0, |
| "auto_map": { |
| "AutoConfig": "configuration_lite_whisper.LiteWhisperConfig", |
| "AutoModel": "modeling_lite_whisper.LiteWhisperForConditionalGeneration" |
| }, |
| "begin_suppress_tokens": null, |
| "bos_token_id": 50257, |
| "classifier_proj_size": 256, |
| "d_model": 1024, |
| "decoder_attention_heads": 16, |
| "decoder_ffn_dim": 4096, |
| "decoder_layerdrop": 0.0, |
| "decoder_layers": 24, |
| "decoder_start_token_id": 50258, |
| "dropout": 0.0, |
| "encoder_attention_heads": 16, |
| "encoder_ffn_dim": 4096, |
| "encoder_layerdrop": 0.0, |
| "encoder_layers": 24, |
| "eos_token_id": 50257, |
| "forced_decoder_ids": [ |
| [ |
| 1, |
| 50259 |
| ], |
| [ |
| 2, |
| 50359 |
| ], |
| [ |
| 3, |
| 50363 |
| ] |
| ], |
| "init_std": 0.02, |
| "is_encoder_decoder": true, |
| "low_rank_config": [ |
| { |
| "fc1": 304, |
| "fc2": 432, |
| "k_proj": 48, |
| "out_proj": 80, |
| "q_proj": 48, |
| "v_proj": 128 |
| }, |
| { |
| "fc1": 384, |
| "fc2": 512, |
| "k_proj": 80, |
| "out_proj": 176, |
| "q_proj": 96, |
| "v_proj": 240 |
| }, |
| { |
| "fc1": 416, |
| "fc2": 768, |
| "k_proj": 80, |
| "out_proj": 224, |
| "q_proj": 96, |
| "v_proj": 208 |
| }, |
| { |
| "fc1": 464, |
| "fc2": 768, |
| "k_proj": 80, |
| "out_proj": 240, |
| "q_proj": 96, |
| "v_proj": 256 |
| }, |
| { |
| "fc1": 512, |
| "k_proj": 112, |
| "out_proj": 224, |
| "q_proj": 144, |
| "v_proj": 288 |
| }, |
| { |
| "fc1": 672, |
| "k_proj": 128, |
| "out_proj": 304, |
| "q_proj": 160, |
| "v_proj": 368 |
| }, |
| { |
| "fc1": 736, |
| "fc2": 752, |
| "k_proj": 176, |
| "out_proj": 320, |
| "q_proj": 240, |
| "v_proj": 416 |
| }, |
| { |
| "fc1": 512, |
| "fc2": 16, |
| "k_proj": 160, |
| "out_proj": 320, |
| "q_proj": 224, |
| "v_proj": 400 |
| }, |
| { |
| "fc1": 768, |
| "k_proj": 144, |
| "out_proj": 336, |
| "q_proj": 208, |
| "v_proj": 416 |
| }, |
| { |
| "fc1": 800, |
| "k_proj": 176, |
| "out_proj": 304, |
| "q_proj": 240, |
| "v_proj": 400 |
| }, |
| { |
| "fc2": 768, |
| "k_proj": 144, |
| "out_proj": 240, |
| "q_proj": 208, |
| "v_proj": 368 |
| }, |
| { |
| "k_proj": 208, |
| "out_proj": 288, |
| "q_proj": 272, |
| "v_proj": 416 |
| }, |
| { |
| "k_proj": 240, |
| "out_proj": 336, |
| "q_proj": 288, |
| "v_proj": 464 |
| }, |
| { |
| "k_proj": 224, |
| "out_proj": 320, |
| "q_proj": 288, |
| "v_proj": 448 |
| }, |
| { |
| "k_proj": 256, |
| "out_proj": 336, |
| "q_proj": 320, |
| "v_proj": 480 |
| }, |
| { |
| "k_proj": 272, |
| "out_proj": 352, |
| "q_proj": 336, |
| "v_proj": 480 |
| }, |
| { |
| "k_proj": 272, |
| "out_proj": 320, |
| "q_proj": 320, |
| "v_proj": 496 |
| }, |
| { |
| "k_proj": 368, |
| "out_proj": 400, |
| "q_proj": 432 |
| }, |
| { |
| "k_proj": 224, |
| "out_proj": 384, |
| "q_proj": 320 |
| }, |
| { |
| "k_proj": 272, |
| "out_proj": 432, |
| "q_proj": 352 |
| }, |
| { |
| "k_proj": 368, |
| "out_proj": 368, |
| "q_proj": 432, |
| "v_proj": 512 |
| }, |
| { |
| "k_proj": 336, |
| "out_proj": 368, |
| "q_proj": 416, |
| "v_proj": 480 |
| }, |
| { |
| "k_proj": 208, |
| "out_proj": 288, |
| "q_proj": 304, |
| "v_proj": 352 |
| }, |
| { |
| "k_proj": 288, |
| "out_proj": 320, |
| "q_proj": 384, |
| "v_proj": 400 |
| } |
| ], |
| "mask_feature_length": 10, |
| "mask_feature_min_masks": 0, |
| "mask_feature_prob": 0.0, |
| "mask_time_length": 10, |
| "mask_time_min_masks": 2, |
| "mask_time_prob": 0.05, |
| "max_length": null, |
| "max_source_positions": 1500, |
| "max_target_positions": 448, |
| "median_filter_width": 7, |
| "model_type": "lite-whisper", |
| "num_hidden_layers": 24, |
| "num_mel_bins": 80, |
| "pad_token_id": 50257, |
| "scale_embedding": false, |
| "torch_dtype": "float32", |
| "transformers_version": "4.49.0", |
| "use_cache": true, |
| "use_weighted_layer_sum": false, |
| "vocab_size": 51865 |
| } |
|
|