| { |
| "architectures": [ |
| "ParakeetForTDT" |
| ], |
| "blank_token_id": 8192, |
| "decoder_hidden_size": 32, |
| "dtype": "float32", |
| "durations": [ |
| 0, |
| 1, |
| 2, |
| 3, |
| 4 |
| ], |
| "encoder_config": { |
| "activation_dropout": 0.0, |
| "attention_bias": true, |
| "attention_dropout": 0.0, |
| "conv_kernel_size": 9, |
| "convolution_bias": true, |
| "dropout": 0.0, |
| "dropout_positions": 0.0, |
| "hidden_act": "silu", |
| "hidden_size": 64, |
| "initializer_range": 0.02, |
| "intermediate_size": 256, |
| "layerdrop": 0.0, |
| "max_position_embeddings": 5000, |
| "model_type": "parakeet_encoder", |
| "num_attention_heads": 4, |
| "num_hidden_layers": 2, |
| "num_key_value_heads": 4, |
| "num_mel_bins": 80, |
| "scale_input": true, |
| "subsampling_conv_channels": 32, |
| "subsampling_conv_kernel_size": 3, |
| "subsampling_conv_stride": 2, |
| "subsampling_factor": 8 |
| }, |
| "hidden_act": "relu", |
| "initializer_range": 0.02, |
| "is_encoder_decoder": true, |
| "max_symbols_per_step": 5, |
| "model_type": "parakeet_tdt", |
| "num_decoder_layers": 1, |
| "pad_token_id": 2, |
| "transformers_version": "5.16.0.dev0", |
| "vocab_size": 8193 |
| } |
|
|
