end-2-end reborn model for librispeech unsupervised phoneme recognition (iter5-stage1)
62e72b3
verified
{ | |
"architectures": [ | |
"RebornUASRModel" | |
], | |
"auto_map": { | |
"AutoConfig": "configuration_reborn.RebornUASRConfig", | |
"AutoModel": "modeling_reborn.RebornUASRModel" | |
}, | |
"discriminator_act_after_linear": false, | |
"discriminator_causal": true, | |
"discriminator_depth": 1, | |
"discriminator_dilation": 1, | |
"discriminator_dim": 256, | |
"discriminator_dropout": 0.0, | |
"discriminator_input_dim": 44, | |
"discriminator_kernel": 3, | |
"discriminator_linear_emb": false, | |
"discriminator_max_pool": false, | |
"discriminator_spectral_norm": false, | |
"discriminator_weight_norm": false, | |
"generator_bias": false, | |
"generator_bn_apply": false, | |
"generator_bn_init_weight": 30.0, | |
"generator_dilation": 1, | |
"generator_dropout": 0.0, | |
"generator_input_dim": 512, | |
"generator_kernel": 4, | |
"generator_output_dim": 44, | |
"generator_stride": 1, | |
"model_type": "reborn_uasr", | |
"phones": [ | |
"AH", | |
"N", | |
"S", | |
"IH", | |
"T", | |
"L", | |
"R", | |
"D", | |
"K", | |
"IY", | |
"Z", | |
"M", | |
"ER", | |
"EH", | |
"P", | |
"AE", | |
"B", | |
"AA", | |
"EY", | |
"F", | |
"OW", | |
"NG", | |
"G", | |
"V", | |
"AO", | |
"AY", | |
"SH", | |
"UW", | |
"W", | |
"HH", | |
"JH", | |
"Y", | |
"CH", | |
"TH", | |
"AW", | |
"UH", | |
"OY", | |
"DH", | |
"ZH", | |
"<SIL>" | |
], | |
"segmenter_dropout": 0.1, | |
"segmenter_hidden_dim": 512, | |
"segmenter_input_dim": 512, | |
"segmenter_kernel_size": 7, | |
"segmenter_type": "cnn", | |
"special_token_nums": 4, | |
"torch_dtype": "float32", | |
"transformers_version": "4.24.0" | |
} | |