| { | |
| "_attn_implementation_autoset": true, | |
| "_name_or_path": "eustlb/moonshine", | |
| "apply_spec_augment": false, | |
| "architectures": [ | |
| "MoonshineModel" | |
| ], | |
| "attention_bias": false, | |
| "attention_dropout": 0.0, | |
| "bos_token_id": 1, | |
| "decoder_hidden_act": "silu", | |
| "decoder_start_token_id": 1, | |
| "encoder_hidden_act": "gelu", | |
| "eos_token_id": 2, | |
| "ff_mult": 4, | |
| "hidden_size": 288, | |
| "initializer_range": 0.02, | |
| "intermediate_size": null, | |
| "is_encoder_decoder": true, | |
| "layer_norm_eps": 1e-05, | |
| "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_position_embeddings": 2048, | |
| "model_type": "moonshine", | |
| "num_attention_heads": 8, | |
| "num_hidden_layers": 6, | |
| "num_key_value_heads": 8, | |
| "partial_rotary_factor": 0.5, | |
| "qk_layernorm": false, | |
| "rope_scaling": null, | |
| "rope_theta": 10000.0, | |
| "transformers_version": "4.47.0.dev0", | |
| "use_cache": true, | |
| "vocab_size": 32768 | |
| } | |