Add files using upload-large-folder tool
Browse files- qwen3-tts-12hz-1.7b-base/config/generation_config.json +12 -0
- qwen3-tts-12hz-1.7b-base/config/model.json +167 -0
- qwen3-tts-12hz-1.7b-base/config/speech_tokenizer.json +94 -0
- qwen3-tts-12hz-1.7b-base/manifest.json +66 -0
- qwen3-tts-12hz-1.7b-base/tokenizer/merges.txt +3 -0
- qwen3-tts-12hz-1.7b-base/tokenizer/preprocessor_config.json +6 -0
- qwen3-tts-12hz-1.7b-base/tokenizer/tokenizer_config.json +316 -0
- qwen3-tts-12hz-1.7b-base/tokenizer/vocab.json +3 -0
- qwen3-tts-12hz-1.7b-base/weights/model.safetensors +3 -0
- qwen3-tts-12hz-1.7b-base/weights/speech_tokenizer.safetensors +3 -0
- qwen3-tts-12hz-1.7b-customvoice/config/generation_config.json +12 -0
- qwen3-tts-12hz-1.7b-customvoice/config/model.json +183 -0
- qwen3-tts-12hz-1.7b-customvoice/config/speech_tokenizer.json +94 -0
- qwen3-tts-12hz-1.7b-customvoice/manifest.json +65 -0
- qwen3-tts-12hz-1.7b-customvoice/tokenizer/merges.txt +3 -0
- qwen3-tts-12hz-1.7b-customvoice/tokenizer/preprocessor_config.json +6 -0
- qwen3-tts-12hz-1.7b-customvoice/tokenizer/tokenizer_config.json +316 -0
- qwen3-tts-12hz-1.7b-customvoice/tokenizer/vocab.json +3 -0
- qwen3-tts-12hz-1.7b-customvoice/weights/model.safetensors +3 -0
- qwen3-tts-12hz-1.7b-customvoice/weights/speech_tokenizer.safetensors +3 -0
- qwen3-tts-12hz-1.7b-voicedesign/config/generation_config.json +12 -0
- qwen3-tts-12hz-1.7b-voicedesign/config/model.json +163 -0
- qwen3-tts-12hz-1.7b-voicedesign/config/speech_tokenizer.json +94 -0
- qwen3-tts-12hz-1.7b-voicedesign/manifest.json +65 -0
- qwen3-tts-12hz-1.7b-voicedesign/tokenizer/merges.txt +3 -0
- qwen3-tts-12hz-1.7b-voicedesign/tokenizer/preprocessor_config.json +6 -0
- qwen3-tts-12hz-1.7b-voicedesign/tokenizer/tokenizer_config.json +316 -0
- qwen3-tts-12hz-1.7b-voicedesign/tokenizer/vocab.json +3 -0
- qwen3-tts-12hz-1.7b-voicedesign/weights/model.safetensors +3 -0
- qwen3-tts-12hz-1.7b-voicedesign/weights/speech_tokenizer.safetensors +3 -0
qwen3-tts-12hz-1.7b-base/config/generation_config.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_sample": true,
|
| 3 |
+
"repetition_penalty": 1.05,
|
| 4 |
+
"temperature": 0.9,
|
| 5 |
+
"top_p": 1.0,
|
| 6 |
+
"top_k": 50,
|
| 7 |
+
"subtalker_dosample": true,
|
| 8 |
+
"subtalker_temperature": 0.9,
|
| 9 |
+
"subtalker_top_p": 1.0,
|
| 10 |
+
"subtalker_top_k": 50,
|
| 11 |
+
"max_new_tokens": 8192
|
| 12 |
+
}
|
qwen3-tts-12hz-1.7b-base/config/model.json
ADDED
|
@@ -0,0 +1,167 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3TTSForConditionalGeneration"
|
| 4 |
+
],
|
| 5 |
+
"assistant_token_id": 77091,
|
| 6 |
+
"im_end_token_id": 151645,
|
| 7 |
+
"im_start_token_id": 151644,
|
| 8 |
+
"tts_bos_token_id": 151672,
|
| 9 |
+
"tts_eos_token_id": 151673,
|
| 10 |
+
"tts_pad_token_id": 151671,
|
| 11 |
+
"model_type": "qwen3_tts",
|
| 12 |
+
"tokenizer_type": "qwen3_tts_tokenizer_12hz",
|
| 13 |
+
"tts_model_size": "1b7",
|
| 14 |
+
"tts_model_type": "base",
|
| 15 |
+
"speaker_encoder_config": {
|
| 16 |
+
"enc_dim": 2048,
|
| 17 |
+
"sample_rate": 24000
|
| 18 |
+
},
|
| 19 |
+
"talker_config": {
|
| 20 |
+
"attention_bias": false,
|
| 21 |
+
"attention_dropout": 0,
|
| 22 |
+
"code_predictor_config": {
|
| 23 |
+
"_name_or_path": "",
|
| 24 |
+
"add_cross_attention": false,
|
| 25 |
+
"architectures": null,
|
| 26 |
+
"attention_bias": false,
|
| 27 |
+
"attention_dropout": 0,
|
| 28 |
+
"bad_words_ids": null,
|
| 29 |
+
"begin_suppress_tokens": null,
|
| 30 |
+
"bos_token_id": null,
|
| 31 |
+
"chunk_size_feed_forward": 0,
|
| 32 |
+
"cross_attention_hidden_size": null,
|
| 33 |
+
"decoder_start_token_id": null,
|
| 34 |
+
"diversity_penalty": 0.0,
|
| 35 |
+
"do_sample": false,
|
| 36 |
+
"early_stopping": false,
|
| 37 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 38 |
+
"eos_token_id": null,
|
| 39 |
+
"exponential_decay_length_penalty": null,
|
| 40 |
+
"finetuning_task": null,
|
| 41 |
+
"forced_bos_token_id": null,
|
| 42 |
+
"forced_eos_token_id": null,
|
| 43 |
+
"head_dim": 128,
|
| 44 |
+
"hidden_act": "silu",
|
| 45 |
+
"hidden_size": 1024,
|
| 46 |
+
"id2label": {
|
| 47 |
+
"0": "LABEL_0",
|
| 48 |
+
"1": "LABEL_1"
|
| 49 |
+
},
|
| 50 |
+
"initializer_range": 0.02,
|
| 51 |
+
"intermediate_size": 3072,
|
| 52 |
+
"is_decoder": false,
|
| 53 |
+
"is_encoder_decoder": false,
|
| 54 |
+
"label2id": {
|
| 55 |
+
"LABEL_0": 0,
|
| 56 |
+
"LABEL_1": 1
|
| 57 |
+
},
|
| 58 |
+
"layer_types": [
|
| 59 |
+
"full_attention",
|
| 60 |
+
"full_attention",
|
| 61 |
+
"full_attention",
|
| 62 |
+
"full_attention",
|
| 63 |
+
"full_attention"
|
| 64 |
+
],
|
| 65 |
+
"length_penalty": 1.0,
|
| 66 |
+
"max_length": 20,
|
| 67 |
+
"max_position_embeddings": 65536,
|
| 68 |
+
"max_window_layers": 28,
|
| 69 |
+
"min_length": 0,
|
| 70 |
+
"model_type": "qwen3_tts_talker_code_predictor",
|
| 71 |
+
"no_repeat_ngram_size": 0,
|
| 72 |
+
"num_attention_heads": 16,
|
| 73 |
+
"num_beam_groups": 1,
|
| 74 |
+
"num_beams": 1,
|
| 75 |
+
"num_code_groups": 16,
|
| 76 |
+
"num_hidden_layers": 5,
|
| 77 |
+
"num_key_value_heads": 8,
|
| 78 |
+
"num_return_sequences": 1,
|
| 79 |
+
"output_attentions": false,
|
| 80 |
+
"output_hidden_states": false,
|
| 81 |
+
"output_scores": false,
|
| 82 |
+
"pad_token_id": null,
|
| 83 |
+
"prefix": null,
|
| 84 |
+
"problem_type": null,
|
| 85 |
+
"pruned_heads": {},
|
| 86 |
+
"remove_invalid_values": false,
|
| 87 |
+
"repetition_penalty": 1.0,
|
| 88 |
+
"return_dict": true,
|
| 89 |
+
"return_dict_in_generate": false,
|
| 90 |
+
"rms_norm_eps": 1e-06,
|
| 91 |
+
"rope_scaling": null,
|
| 92 |
+
"rope_theta": 1000000,
|
| 93 |
+
"sep_token_id": null,
|
| 94 |
+
"sliding_window": null,
|
| 95 |
+
"suppress_tokens": null,
|
| 96 |
+
"task_specific_params": null,
|
| 97 |
+
"temperature": 1.0,
|
| 98 |
+
"tf_legacy_loss": false,
|
| 99 |
+
"tie_encoder_decoder": false,
|
| 100 |
+
"tie_word_embeddings": false,
|
| 101 |
+
"tokenizer_class": null,
|
| 102 |
+
"top_k": 50,
|
| 103 |
+
"top_p": 1.0,
|
| 104 |
+
"dtype": null,
|
| 105 |
+
"torchscript": false,
|
| 106 |
+
"typical_p": 1.0,
|
| 107 |
+
"use_bfloat16": false,
|
| 108 |
+
"use_cache": true,
|
| 109 |
+
"use_sliding_window": false,
|
| 110 |
+
"vocab_size": 2048
|
| 111 |
+
},
|
| 112 |
+
"codec_bos_id": 2149,
|
| 113 |
+
"codec_eos_token_id": 2150,
|
| 114 |
+
"codec_think_id": 2154,
|
| 115 |
+
"codec_language_id": {
|
| 116 |
+
"chinese": 2055,
|
| 117 |
+
"english": 2050,
|
| 118 |
+
"german": 2053,
|
| 119 |
+
"italian": 2070,
|
| 120 |
+
"portuguese": 2071,
|
| 121 |
+
"spanish": 2054,
|
| 122 |
+
"japanese": 2058,
|
| 123 |
+
"korean": 2064,
|
| 124 |
+
"french": 2061,
|
| 125 |
+
"russian": 2069
|
| 126 |
+
},
|
| 127 |
+
"codec_nothink_id": 2155,
|
| 128 |
+
"codec_pad_id": 2148,
|
| 129 |
+
"codec_think_bos_id": 2156,
|
| 130 |
+
"codec_think_eos_id": 2157,
|
| 131 |
+
"spk_id": {
|
| 132 |
+
},
|
| 133 |
+
"spk_is_dialect": {
|
| 134 |
+
},
|
| 135 |
+
"head_dim": 128,
|
| 136 |
+
"hidden_act": "silu",
|
| 137 |
+
"hidden_size": 2048,
|
| 138 |
+
"initializer_range": 0.02,
|
| 139 |
+
"intermediate_size": 6144,
|
| 140 |
+
"max_position_embeddings": 32768,
|
| 141 |
+
"model_type": "qwen3_tts_talker",
|
| 142 |
+
"num_attention_heads": 16,
|
| 143 |
+
"num_code_groups": 16,
|
| 144 |
+
"num_hidden_layers": 28,
|
| 145 |
+
"num_key_value_heads": 8,
|
| 146 |
+
"position_id_per_seconds": 13,
|
| 147 |
+
"rms_norm_eps": 1e-06,
|
| 148 |
+
"rope_scaling": {
|
| 149 |
+
"interleaved": true,
|
| 150 |
+
"mrope_section": [
|
| 151 |
+
24,
|
| 152 |
+
20,
|
| 153 |
+
20
|
| 154 |
+
],
|
| 155 |
+
"rope_type": "default",
|
| 156 |
+
"type": "default"
|
| 157 |
+
},
|
| 158 |
+
"rope_theta": 1000000,
|
| 159 |
+
"sliding_window": null,
|
| 160 |
+
"text_hidden_size": 2048,
|
| 161 |
+
"text_vocab_size": 151936,
|
| 162 |
+
"use_cache": true,
|
| 163 |
+
"use_sliding_window": false,
|
| 164 |
+
"vocab_size": 3072
|
| 165 |
+
},
|
| 166 |
+
"transformers_version": "4.57.3"
|
| 167 |
+
}
|
qwen3-tts-12hz-1.7b-base/config/speech_tokenizer.json
ADDED
|
@@ -0,0 +1,94 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3TTSTokenizerV2Model"
|
| 4 |
+
],
|
| 5 |
+
"model_type": "qwen3_tts_tokenizer_12hz",
|
| 6 |
+
"encoder_valid_num_quantizers": 16,
|
| 7 |
+
"input_sample_rate": 24000,
|
| 8 |
+
"output_sample_rate": 24000,
|
| 9 |
+
"decode_upsample_rate": 1920,
|
| 10 |
+
"encode_downsample_rate": 1920,
|
| 11 |
+
"decoder_config": {
|
| 12 |
+
"attention_bias": false,
|
| 13 |
+
"attention_dropout": 0.0,
|
| 14 |
+
"latent_dim": 1024,
|
| 15 |
+
"codebook_dim": 512,
|
| 16 |
+
"codebook_size": 2048,
|
| 17 |
+
"decoder_dim": 1536,
|
| 18 |
+
"hidden_act": "silu",
|
| 19 |
+
"hidden_size": 512,
|
| 20 |
+
"intermediate_size": 1024,
|
| 21 |
+
"layer_scale_initial_scale": 0.01,
|
| 22 |
+
"max_position_embeddings": 8000,
|
| 23 |
+
"head_dim": 64,
|
| 24 |
+
"num_attention_heads": 16,
|
| 25 |
+
"num_hidden_layers": 8,
|
| 26 |
+
"num_key_value_heads": 16,
|
| 27 |
+
"num_quantizers": 16,
|
| 28 |
+
"num_semantic_quantizers": 1,
|
| 29 |
+
"rms_norm_eps": 1e-05,
|
| 30 |
+
"rope_theta": 10000,
|
| 31 |
+
"semantic_codebook_size": 4096,
|
| 32 |
+
"sliding_window": 72,
|
| 33 |
+
"upsample_rates": [
|
| 34 |
+
8,
|
| 35 |
+
5,
|
| 36 |
+
4,
|
| 37 |
+
3
|
| 38 |
+
],
|
| 39 |
+
"upsampling_ratios": [
|
| 40 |
+
2,
|
| 41 |
+
2
|
| 42 |
+
],
|
| 43 |
+
"vector_quantization_hidden_dimension": 512
|
| 44 |
+
},
|
| 45 |
+
"encoder_config": {
|
| 46 |
+
"_frame_rate": 12.5,
|
| 47 |
+
"attention_bias": false,
|
| 48 |
+
"attention_dropout": 0.0,
|
| 49 |
+
"audio_channels": 1,
|
| 50 |
+
"codebook_dim": 256,
|
| 51 |
+
"codebook_size": 2048,
|
| 52 |
+
"compress": 2,
|
| 53 |
+
"dilation_growth_rate": 2,
|
| 54 |
+
"dtype": "float32",
|
| 55 |
+
"head_dim": 64,
|
| 56 |
+
"hidden_act": "gelu",
|
| 57 |
+
"hidden_size": 512,
|
| 58 |
+
"initializer_range": 0.02,
|
| 59 |
+
"intermediate_size": 2048,
|
| 60 |
+
"kernel_size": 7,
|
| 61 |
+
"last_kernel_size": 3,
|
| 62 |
+
"layer_scale_initial_scale": 0.01,
|
| 63 |
+
"max_position_embeddings": 8000,
|
| 64 |
+
"norm_eps": 1e-05,
|
| 65 |
+
"normalize": false,
|
| 66 |
+
"num_attention_heads": 8,
|
| 67 |
+
"num_filters": 64,
|
| 68 |
+
"num_hidden_layers": 8,
|
| 69 |
+
"num_key_value_heads": 8,
|
| 70 |
+
"num_quantizers": 32,
|
| 71 |
+
"num_residual_layers": 1,
|
| 72 |
+
"num_semantic_quantizers": 1,
|
| 73 |
+
"pad_mode": "constant",
|
| 74 |
+
"residual_kernel_size": 3,
|
| 75 |
+
"rope_theta": 10000.0,
|
| 76 |
+
"sampling_rate": 24000,
|
| 77 |
+
"sliding_window": 250,
|
| 78 |
+
"transformers_version": "4.57.0.dev0",
|
| 79 |
+
"trim_right_ratio": 1.0,
|
| 80 |
+
"upsample_groups": 512,
|
| 81 |
+
"upsampling_ratios": [
|
| 82 |
+
8,
|
| 83 |
+
6,
|
| 84 |
+
5,
|
| 85 |
+
4
|
| 86 |
+
],
|
| 87 |
+
"use_cache": false,
|
| 88 |
+
"use_causal_conv": true,
|
| 89 |
+
"use_conv_shortcut": false,
|
| 90 |
+
"use_streaming": false,
|
| 91 |
+
"vector_quantization_hidden_dimension": 256
|
| 92 |
+
},
|
| 93 |
+
"transformers_version": "4.57.3"
|
| 94 |
+
}
|
qwen3-tts-12hz-1.7b-base/manifest.json
ADDED
|
@@ -0,0 +1,66 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"format": "mlx-bundle/1",
|
| 3 |
+
"components": {
|
| 4 |
+
"model": {
|
| 5 |
+
"class": "Qwen3TTSForConditionalGeneration",
|
| 6 |
+
"source": "/Volumes/work/project/mlx/downloads/Qwen3-TTS-12Hz-1.7B-Base",
|
| 7 |
+
"dtype_policy": "keep",
|
| 8 |
+
"num_tensors": 480,
|
| 9 |
+
"dtypes": {
|
| 10 |
+
"bfloat16": 480
|
| 11 |
+
},
|
| 12 |
+
"layout_rules": {
|
| 13 |
+
"identity": 442,
|
| 14 |
+
"permute(0, 2, 1)": 38
|
| 15 |
+
},
|
| 16 |
+
"files": [
|
| 17 |
+
{
|
| 18 |
+
"file": "weights/model.safetensors",
|
| 19 |
+
"tensors": 480,
|
| 20 |
+
"bytes": 3857413816,
|
| 21 |
+
"sha256": "ab4841b3d8d0d8f70ce5c130a9230b4b992560f3d851572798d9179b70976778"
|
| 22 |
+
},
|
| 23 |
+
{
|
| 24 |
+
"file": "config/model.json",
|
| 25 |
+
"tensors": 0,
|
| 26 |
+
"bytes": 4494,
|
| 27 |
+
"sha256": "b4f01752d15a488abde3e1ab44723ae4f4b9e68a4037257b098b3737893cc1f9"
|
| 28 |
+
},
|
| 29 |
+
{
|
| 30 |
+
"file": "config/generation_config.json",
|
| 31 |
+
"tensors": 0,
|
| 32 |
+
"bytes": 245,
|
| 33 |
+
"sha256": "f1b90b4513f3b34c62851049e2492d7b4c5940daf1276f89c82b8ef04127f3aa"
|
| 34 |
+
}
|
| 35 |
+
]
|
| 36 |
+
},
|
| 37 |
+
"speech_tokenizer": {
|
| 38 |
+
"class": "Qwen3TTSTokenizerV2Model",
|
| 39 |
+
"source": "/Volumes/work/project/mlx/downloads/Qwen3-TTS-12Hz-1.7B-Base/speech_tokenizer",
|
| 40 |
+
"dtype_policy": "keep",
|
| 41 |
+
"num_tensors": 496,
|
| 42 |
+
"dtypes": {
|
| 43 |
+
"float32": 496
|
| 44 |
+
},
|
| 45 |
+
"layout_rules": {
|
| 46 |
+
"identity": 438,
|
| 47 |
+
"permute(0, 2, 1)": 52,
|
| 48 |
+
"permute(1, 2, 0)": 6
|
| 49 |
+
},
|
| 50 |
+
"files": [
|
| 51 |
+
{
|
| 52 |
+
"file": "weights/speech_tokenizer.safetensors",
|
| 53 |
+
"tensors": 496,
|
| 54 |
+
"bytes": 682293164,
|
| 55 |
+
"sha256": "ee778743f0734e5e35f751f771ffe42a8943ca3ce4c9cc2e39b58719989c20c6"
|
| 56 |
+
},
|
| 57 |
+
{
|
| 58 |
+
"file": "config/speech_tokenizer.json",
|
| 59 |
+
"tensors": 0,
|
| 60 |
+
"bytes": 2336,
|
| 61 |
+
"sha256": "ee65bb901c876664ab8707c487157aa1a6ee57c65969b28fb5ec9dc211e68167"
|
| 62 |
+
}
|
| 63 |
+
]
|
| 64 |
+
}
|
| 65 |
+
}
|
| 66 |
+
}
|
qwen3-tts-12hz-1.7b-base/tokenizer/merges.txt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:599bab54075088774b1733fde865d5bd747cbcc7a547c5bc12610e874e26f5e3
|
| 3 |
+
size 1671839
|
qwen3-tts-12hz-1.7b-base/tokenizer/preprocessor_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"padding_side": "left",
|
| 3 |
+
"padding_value": 0.0,
|
| 4 |
+
"processor_class": "Qwen3TTSProcessor",
|
| 5 |
+
"return_attention_mask": true
|
| 6 |
+
}
|
qwen3-tts-12hz-1.7b-base/tokenizer/tokenizer_config.json
ADDED
|
@@ -0,0 +1,316 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": false,
|
| 3 |
+
"add_prefix_space": false,
|
| 4 |
+
"added_tokens_decoder": {
|
| 5 |
+
"151643": {
|
| 6 |
+
"content": "<|endoftext|>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false,
|
| 11 |
+
"special": true
|
| 12 |
+
},
|
| 13 |
+
"151644": {
|
| 14 |
+
"content": "<|im_start|>",
|
| 15 |
+
"lstrip": false,
|
| 16 |
+
"normalized": false,
|
| 17 |
+
"rstrip": false,
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"special": true
|
| 20 |
+
},
|
| 21 |
+
"151645": {
|
| 22 |
+
"content": "<|im_end|>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": false,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false,
|
| 27 |
+
"special": true
|
| 28 |
+
},
|
| 29 |
+
"151646": {
|
| 30 |
+
"content": "<|object_ref_start|>",
|
| 31 |
+
"lstrip": false,
|
| 32 |
+
"normalized": false,
|
| 33 |
+
"rstrip": false,
|
| 34 |
+
"single_word": false,
|
| 35 |
+
"special": true
|
| 36 |
+
},
|
| 37 |
+
"151647": {
|
| 38 |
+
"content": "<|object_ref_end|>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false,
|
| 43 |
+
"special": true
|
| 44 |
+
},
|
| 45 |
+
"151648": {
|
| 46 |
+
"content": "<|box_start|>",
|
| 47 |
+
"lstrip": false,
|
| 48 |
+
"normalized": false,
|
| 49 |
+
"rstrip": false,
|
| 50 |
+
"single_word": false,
|
| 51 |
+
"special": true
|
| 52 |
+
},
|
| 53 |
+
"151649": {
|
| 54 |
+
"content": "<|box_end|>",
|
| 55 |
+
"lstrip": false,
|
| 56 |
+
"normalized": false,
|
| 57 |
+
"rstrip": false,
|
| 58 |
+
"single_word": false,
|
| 59 |
+
"special": true
|
| 60 |
+
},
|
| 61 |
+
"151650": {
|
| 62 |
+
"content": "<|quad_start|>",
|
| 63 |
+
"lstrip": false,
|
| 64 |
+
"normalized": false,
|
| 65 |
+
"rstrip": false,
|
| 66 |
+
"single_word": false,
|
| 67 |
+
"special": true
|
| 68 |
+
},
|
| 69 |
+
"151651": {
|
| 70 |
+
"content": "<|quad_end|>",
|
| 71 |
+
"lstrip": false,
|
| 72 |
+
"normalized": false,
|
| 73 |
+
"rstrip": false,
|
| 74 |
+
"single_word": false,
|
| 75 |
+
"special": true
|
| 76 |
+
},
|
| 77 |
+
"151652": {
|
| 78 |
+
"content": "<|vision_start|>",
|
| 79 |
+
"lstrip": false,
|
| 80 |
+
"normalized": false,
|
| 81 |
+
"rstrip": false,
|
| 82 |
+
"single_word": false,
|
| 83 |
+
"special": true
|
| 84 |
+
},
|
| 85 |
+
"151653": {
|
| 86 |
+
"content": "<|vision_end|>",
|
| 87 |
+
"lstrip": false,
|
| 88 |
+
"normalized": false,
|
| 89 |
+
"rstrip": false,
|
| 90 |
+
"single_word": false,
|
| 91 |
+
"special": true
|
| 92 |
+
},
|
| 93 |
+
"151654": {
|
| 94 |
+
"content": "<|vision_pad|>",
|
| 95 |
+
"lstrip": false,
|
| 96 |
+
"normalized": false,
|
| 97 |
+
"rstrip": false,
|
| 98 |
+
"single_word": false,
|
| 99 |
+
"special": true
|
| 100 |
+
},
|
| 101 |
+
"151655": {
|
| 102 |
+
"content": "<|image_pad|>",
|
| 103 |
+
"lstrip": false,
|
| 104 |
+
"normalized": false,
|
| 105 |
+
"rstrip": false,
|
| 106 |
+
"single_word": false,
|
| 107 |
+
"special": true
|
| 108 |
+
},
|
| 109 |
+
"151656": {
|
| 110 |
+
"content": "<|video_pad|>",
|
| 111 |
+
"lstrip": false,
|
| 112 |
+
"normalized": false,
|
| 113 |
+
"rstrip": false,
|
| 114 |
+
"single_word": false,
|
| 115 |
+
"special": true
|
| 116 |
+
},
|
| 117 |
+
"151657": {
|
| 118 |
+
"content": "<tool_call>",
|
| 119 |
+
"lstrip": false,
|
| 120 |
+
"normalized": false,
|
| 121 |
+
"rstrip": false,
|
| 122 |
+
"single_word": false,
|
| 123 |
+
"special": false
|
| 124 |
+
},
|
| 125 |
+
"151658": {
|
| 126 |
+
"content": "</tool_call>",
|
| 127 |
+
"lstrip": false,
|
| 128 |
+
"normalized": false,
|
| 129 |
+
"rstrip": false,
|
| 130 |
+
"single_word": false,
|
| 131 |
+
"special": false
|
| 132 |
+
},
|
| 133 |
+
"151659": {
|
| 134 |
+
"content": "<|fim_prefix|>",
|
| 135 |
+
"lstrip": false,
|
| 136 |
+
"normalized": false,
|
| 137 |
+
"rstrip": false,
|
| 138 |
+
"single_word": false,
|
| 139 |
+
"special": false
|
| 140 |
+
},
|
| 141 |
+
"151660": {
|
| 142 |
+
"content": "<|fim_middle|>",
|
| 143 |
+
"lstrip": false,
|
| 144 |
+
"normalized": false,
|
| 145 |
+
"rstrip": false,
|
| 146 |
+
"single_word": false,
|
| 147 |
+
"special": false
|
| 148 |
+
},
|
| 149 |
+
"151661": {
|
| 150 |
+
"content": "<|fim_suffix|>",
|
| 151 |
+
"lstrip": false,
|
| 152 |
+
"normalized": false,
|
| 153 |
+
"rstrip": false,
|
| 154 |
+
"single_word": false,
|
| 155 |
+
"special": false
|
| 156 |
+
},
|
| 157 |
+
"151662": {
|
| 158 |
+
"content": "<|fim_pad|>",
|
| 159 |
+
"lstrip": false,
|
| 160 |
+
"normalized": false,
|
| 161 |
+
"rstrip": false,
|
| 162 |
+
"single_word": false,
|
| 163 |
+
"special": false
|
| 164 |
+
},
|
| 165 |
+
"151663": {
|
| 166 |
+
"content": "<|repo_name|>",
|
| 167 |
+
"lstrip": false,
|
| 168 |
+
"normalized": false,
|
| 169 |
+
"rstrip": false,
|
| 170 |
+
"single_word": false,
|
| 171 |
+
"special": false
|
| 172 |
+
},
|
| 173 |
+
"151664": {
|
| 174 |
+
"content": "<|file_sep|>",
|
| 175 |
+
"lstrip": false,
|
| 176 |
+
"normalized": false,
|
| 177 |
+
"rstrip": false,
|
| 178 |
+
"single_word": false,
|
| 179 |
+
"special": false
|
| 180 |
+
},
|
| 181 |
+
"151665": {
|
| 182 |
+
"content": "<tool_response>",
|
| 183 |
+
"lstrip": false,
|
| 184 |
+
"normalized": false,
|
| 185 |
+
"rstrip": false,
|
| 186 |
+
"single_word": false,
|
| 187 |
+
"special": false
|
| 188 |
+
},
|
| 189 |
+
"151666": {
|
| 190 |
+
"content": "</tool_response>",
|
| 191 |
+
"lstrip": false,
|
| 192 |
+
"normalized": false,
|
| 193 |
+
"rstrip": false,
|
| 194 |
+
"single_word": false,
|
| 195 |
+
"special": false
|
| 196 |
+
},
|
| 197 |
+
"151667": {
|
| 198 |
+
"content": "<think>",
|
| 199 |
+
"lstrip": false,
|
| 200 |
+
"normalized": false,
|
| 201 |
+
"rstrip": false,
|
| 202 |
+
"single_word": false,
|
| 203 |
+
"special": false
|
| 204 |
+
},
|
| 205 |
+
"151668": {
|
| 206 |
+
"content": "</think>",
|
| 207 |
+
"lstrip": false,
|
| 208 |
+
"normalized": false,
|
| 209 |
+
"rstrip": false,
|
| 210 |
+
"single_word": false,
|
| 211 |
+
"special": false
|
| 212 |
+
},
|
| 213 |
+
"151669": {
|
| 214 |
+
"content": "<|audio_start|>",
|
| 215 |
+
"lstrip": false,
|
| 216 |
+
"normalized": false,
|
| 217 |
+
"rstrip": false,
|
| 218 |
+
"single_word": false,
|
| 219 |
+
"special": true
|
| 220 |
+
},
|
| 221 |
+
"151670": {
|
| 222 |
+
"content": "<|audio_end|>",
|
| 223 |
+
"lstrip": false,
|
| 224 |
+
"normalized": false,
|
| 225 |
+
"rstrip": false,
|
| 226 |
+
"single_word": false,
|
| 227 |
+
"special": true
|
| 228 |
+
},
|
| 229 |
+
"151671": {
|
| 230 |
+
"content": "<tts_pad>",
|
| 231 |
+
"lstrip": false,
|
| 232 |
+
"normalized": false,
|
| 233 |
+
"rstrip": false,
|
| 234 |
+
"single_word": false,
|
| 235 |
+
"special": true
|
| 236 |
+
},
|
| 237 |
+
"151672": {
|
| 238 |
+
"content": "<tts_text_bos>",
|
| 239 |
+
"lstrip": false,
|
| 240 |
+
"normalized": false,
|
| 241 |
+
"rstrip": false,
|
| 242 |
+
"single_word": false,
|
| 243 |
+
"special": true
|
| 244 |
+
},
|
| 245 |
+
"151673": {
|
| 246 |
+
"content": "<tts_text_eod>",
|
| 247 |
+
"lstrip": false,
|
| 248 |
+
"normalized": false,
|
| 249 |
+
"rstrip": false,
|
| 250 |
+
"single_word": false,
|
| 251 |
+
"special": true
|
| 252 |
+
},
|
| 253 |
+
"151674": {
|
| 254 |
+
"content": "<tts_text_bos_single>",
|
| 255 |
+
"lstrip": false,
|
| 256 |
+
"normalized": false,
|
| 257 |
+
"rstrip": false,
|
| 258 |
+
"single_word": false,
|
| 259 |
+
"special": true
|
| 260 |
+
},
|
| 261 |
+
"151675": {
|
| 262 |
+
"content": "<|audio_pad|>",
|
| 263 |
+
"lstrip": false,
|
| 264 |
+
"normalized": false,
|
| 265 |
+
"rstrip": false,
|
| 266 |
+
"single_word": false,
|
| 267 |
+
"special": true
|
| 268 |
+
}
|
| 269 |
+
},
|
| 270 |
+
"additional_special_tokens": [
|
| 271 |
+
"<|im_start|>",
|
| 272 |
+
"<|im_end|>",
|
| 273 |
+
"<|object_ref_start|>",
|
| 274 |
+
"<|object_ref_end|>",
|
| 275 |
+
"<|box_start|>",
|
| 276 |
+
"<|box_end|>",
|
| 277 |
+
"<|quad_start|>",
|
| 278 |
+
"<|quad_end|>",
|
| 279 |
+
"<|vision_start|>",
|
| 280 |
+
"<|vision_end|>",
|
| 281 |
+
"<|vision_pad|>",
|
| 282 |
+
"<|image_pad|>",
|
| 283 |
+
"<|video_pad|>",
|
| 284 |
+
"<|audio_start|>",
|
| 285 |
+
"<|audio_end|>",
|
| 286 |
+
"<tts_pad>",
|
| 287 |
+
"<tts_text_bos>",
|
| 288 |
+
"<tts_text_bos_single>",
|
| 289 |
+
"<|audio_pad|>"
|
| 290 |
+
],
|
| 291 |
+
"extra_special_tokens": {
|
| 292 |
+
"image_token": "<|image_pad|>",
|
| 293 |
+
"audio_token": "<|audio_pad|>",
|
| 294 |
+
"video_token": "<|video_pad|>",
|
| 295 |
+
"vision_bos_token": "<|vision_start|>",
|
| 296 |
+
"vision_eos_token": "<|vision_end|>",
|
| 297 |
+
"audio_bos_token": "<|audio_start|>",
|
| 298 |
+
"audio_eos_token": "<|audio_end|>"
|
| 299 |
+
},
|
| 300 |
+
"bos_token": null,
|
| 301 |
+
"clean_up_tokenization_spaces": false,
|
| 302 |
+
"eos_token": "<|im_end|>",
|
| 303 |
+
"errors": "replace",
|
| 304 |
+
"model_max_length": 131072,
|
| 305 |
+
"pad_token": "<|endoftext|>",
|
| 306 |
+
"split_special_tokens": false,
|
| 307 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 308 |
+
"unk_token": null,
|
| 309 |
+
"image_token": "<|image_pad|>",
|
| 310 |
+
"audio_token": "<|audio_pad|>",
|
| 311 |
+
"video_token": "<|video_pad|>",
|
| 312 |
+
"vision_bos_token": "<|vision_start|>",
|
| 313 |
+
"vision_eos_token": "<|vision_end|>",
|
| 314 |
+
"audio_bos_token": "<|audio_start|>",
|
| 315 |
+
"audio_eos_token": "<|audio_end|>"
|
| 316 |
+
}
|
qwen3-tts-12hz-1.7b-base/tokenizer/vocab.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ca10d7e9fb3ed18575dd1e277a2579c16d108e32f27439684afa0e10b1440910
|
| 3 |
+
size 2776833
|
qwen3-tts-12hz-1.7b-base/weights/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ab4841b3d8d0d8f70ce5c130a9230b4b992560f3d851572798d9179b70976778
|
| 3 |
+
size 3857413816
|
qwen3-tts-12hz-1.7b-base/weights/speech_tokenizer.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ee778743f0734e5e35f751f771ffe42a8943ca3ce4c9cc2e39b58719989c20c6
|
| 3 |
+
size 682293164
|
qwen3-tts-12hz-1.7b-customvoice/config/generation_config.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_sample": true,
|
| 3 |
+
"repetition_penalty": 1.05,
|
| 4 |
+
"temperature": 0.9,
|
| 5 |
+
"top_p": 1.0,
|
| 6 |
+
"top_k": 50,
|
| 7 |
+
"subtalker_dosample": true,
|
| 8 |
+
"subtalker_temperature": 0.9,
|
| 9 |
+
"subtalker_top_p": 1.0,
|
| 10 |
+
"subtalker_top_k": 50,
|
| 11 |
+
"max_new_tokens": 8192
|
| 12 |
+
}
|
qwen3-tts-12hz-1.7b-customvoice/config/model.json
ADDED
|
@@ -0,0 +1,183 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3TTSForConditionalGeneration"
|
| 4 |
+
],
|
| 5 |
+
"assistant_token_id": 77091,
|
| 6 |
+
"im_end_token_id": 151645,
|
| 7 |
+
"im_start_token_id": 151644,
|
| 8 |
+
"tts_bos_token_id": 151672,
|
| 9 |
+
"tts_eos_token_id": 151673,
|
| 10 |
+
"tts_pad_token_id": 151671,
|
| 11 |
+
"model_type": "qwen3_tts",
|
| 12 |
+
"tokenizer_type": "qwen3_tts_tokenizer_12hz",
|
| 13 |
+
"tts_model_size": "1b7",
|
| 14 |
+
"tts_model_type": "custom_voice",
|
| 15 |
+
"talker_config": {
|
| 16 |
+
"attention_bias": false,
|
| 17 |
+
"attention_dropout": 0,
|
| 18 |
+
"code_predictor_config": {
|
| 19 |
+
"_name_or_path": "",
|
| 20 |
+
"add_cross_attention": false,
|
| 21 |
+
"architectures": null,
|
| 22 |
+
"attention_bias": false,
|
| 23 |
+
"attention_dropout": 0,
|
| 24 |
+
"bad_words_ids": null,
|
| 25 |
+
"begin_suppress_tokens": null,
|
| 26 |
+
"bos_token_id": null,
|
| 27 |
+
"chunk_size_feed_forward": 0,
|
| 28 |
+
"cross_attention_hidden_size": null,
|
| 29 |
+
"decoder_start_token_id": null,
|
| 30 |
+
"diversity_penalty": 0.0,
|
| 31 |
+
"do_sample": false,
|
| 32 |
+
"early_stopping": false,
|
| 33 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 34 |
+
"eos_token_id": null,
|
| 35 |
+
"exponential_decay_length_penalty": null,
|
| 36 |
+
"finetuning_task": null,
|
| 37 |
+
"forced_bos_token_id": null,
|
| 38 |
+
"forced_eos_token_id": null,
|
| 39 |
+
"head_dim": 128,
|
| 40 |
+
"hidden_act": "silu",
|
| 41 |
+
"hidden_size": 1024,
|
| 42 |
+
"id2label": {
|
| 43 |
+
"0": "LABEL_0",
|
| 44 |
+
"1": "LABEL_1"
|
| 45 |
+
},
|
| 46 |
+
"initializer_range": 0.02,
|
| 47 |
+
"intermediate_size": 3072,
|
| 48 |
+
"is_decoder": false,
|
| 49 |
+
"is_encoder_decoder": false,
|
| 50 |
+
"label2id": {
|
| 51 |
+
"LABEL_0": 0,
|
| 52 |
+
"LABEL_1": 1
|
| 53 |
+
},
|
| 54 |
+
"layer_types": [
|
| 55 |
+
"full_attention",
|
| 56 |
+
"full_attention",
|
| 57 |
+
"full_attention",
|
| 58 |
+
"full_attention",
|
| 59 |
+
"full_attention"
|
| 60 |
+
],
|
| 61 |
+
"length_penalty": 1.0,
|
| 62 |
+
"max_length": 20,
|
| 63 |
+
"max_position_embeddings": 65536,
|
| 64 |
+
"max_window_layers": 28,
|
| 65 |
+
"min_length": 0,
|
| 66 |
+
"model_type": "qwen3_tts_talker_code_predictor",
|
| 67 |
+
"no_repeat_ngram_size": 0,
|
| 68 |
+
"num_attention_heads": 16,
|
| 69 |
+
"num_beam_groups": 1,
|
| 70 |
+
"num_beams": 1,
|
| 71 |
+
"num_code_groups": 16,
|
| 72 |
+
"num_hidden_layers": 5,
|
| 73 |
+
"num_key_value_heads": 8,
|
| 74 |
+
"num_return_sequences": 1,
|
| 75 |
+
"output_attentions": false,
|
| 76 |
+
"output_hidden_states": false,
|
| 77 |
+
"output_scores": false,
|
| 78 |
+
"pad_token_id": null,
|
| 79 |
+
"prefix": null,
|
| 80 |
+
"problem_type": null,
|
| 81 |
+
"pruned_heads": {},
|
| 82 |
+
"remove_invalid_values": false,
|
| 83 |
+
"repetition_penalty": 1.0,
|
| 84 |
+
"return_dict": true,
|
| 85 |
+
"return_dict_in_generate": false,
|
| 86 |
+
"rms_norm_eps": 1e-06,
|
| 87 |
+
"rope_scaling": null,
|
| 88 |
+
"rope_theta": 1000000,
|
| 89 |
+
"sep_token_id": null,
|
| 90 |
+
"sliding_window": null,
|
| 91 |
+
"suppress_tokens": null,
|
| 92 |
+
"task_specific_params": null,
|
| 93 |
+
"temperature": 1.0,
|
| 94 |
+
"tf_legacy_loss": false,
|
| 95 |
+
"tie_encoder_decoder": false,
|
| 96 |
+
"tie_word_embeddings": false,
|
| 97 |
+
"tokenizer_class": null,
|
| 98 |
+
"top_k": 50,
|
| 99 |
+
"top_p": 1.0,
|
| 100 |
+
"dtype": null,
|
| 101 |
+
"torchscript": false,
|
| 102 |
+
"typical_p": 1.0,
|
| 103 |
+
"use_bfloat16": false,
|
| 104 |
+
"use_cache": true,
|
| 105 |
+
"use_sliding_window": false,
|
| 106 |
+
"vocab_size": 2048
|
| 107 |
+
},
|
| 108 |
+
"codec_bos_id": 2149,
|
| 109 |
+
"codec_eos_token_id": 2150,
|
| 110 |
+
"codec_think_id": 2154,
|
| 111 |
+
"codec_language_id": {
|
| 112 |
+
"chinese": 2055,
|
| 113 |
+
"english": 2050,
|
| 114 |
+
"german": 2053,
|
| 115 |
+
"italian": 2070,
|
| 116 |
+
"portuguese": 2071,
|
| 117 |
+
"spanish": 2054,
|
| 118 |
+
"japanese": 2058,
|
| 119 |
+
"korean": 2064,
|
| 120 |
+
"french": 2061,
|
| 121 |
+
"russian": 2069,
|
| 122 |
+
"beijing_dialect": 2074,
|
| 123 |
+
"sichuan_dialect": 2062
|
| 124 |
+
},
|
| 125 |
+
"codec_nothink_id": 2155,
|
| 126 |
+
"codec_pad_id": 2148,
|
| 127 |
+
"codec_think_bos_id": 2156,
|
| 128 |
+
"codec_think_eos_id": 2157,
|
| 129 |
+
"spk_id": {
|
| 130 |
+
"serena": 3066,
|
| 131 |
+
"vivian": 3065,
|
| 132 |
+
"uncle_fu": 3010,
|
| 133 |
+
"ryan": 3061,
|
| 134 |
+
"aiden": 2861,
|
| 135 |
+
"ono_anna": 2873,
|
| 136 |
+
"sohee": 2864,
|
| 137 |
+
"eric": 2875,
|
| 138 |
+
"dylan": 2878
|
| 139 |
+
},
|
| 140 |
+
"spk_is_dialect": {
|
| 141 |
+
"serena": false,
|
| 142 |
+
"vivian": false,
|
| 143 |
+
"uncle_fu": false,
|
| 144 |
+
"ryan": false,
|
| 145 |
+
"aiden": false,
|
| 146 |
+
"ono_anna": false,
|
| 147 |
+
"sohee": false,
|
| 148 |
+
"eric": "sichuan_dialect",
|
| 149 |
+
"dylan": "beijing_dialect"
|
| 150 |
+
},
|
| 151 |
+
"head_dim": 128,
|
| 152 |
+
"hidden_act": "silu",
|
| 153 |
+
"hidden_size": 2048,
|
| 154 |
+
"initializer_range": 0.02,
|
| 155 |
+
"intermediate_size": 6144,
|
| 156 |
+
"max_position_embeddings": 32768,
|
| 157 |
+
"model_type": "qwen3_tts_talker",
|
| 158 |
+
"num_attention_heads": 16,
|
| 159 |
+
"num_code_groups": 16,
|
| 160 |
+
"num_hidden_layers": 28,
|
| 161 |
+
"num_key_value_heads": 8,
|
| 162 |
+
"position_id_per_seconds": 13,
|
| 163 |
+
"rms_norm_eps": 1e-06,
|
| 164 |
+
"rope_scaling": {
|
| 165 |
+
"interleaved": true,
|
| 166 |
+
"mrope_section": [
|
| 167 |
+
24,
|
| 168 |
+
20,
|
| 169 |
+
20
|
| 170 |
+
],
|
| 171 |
+
"rope_type": "default",
|
| 172 |
+
"type": "default"
|
| 173 |
+
},
|
| 174 |
+
"rope_theta": 1000000,
|
| 175 |
+
"sliding_window": null,
|
| 176 |
+
"text_hidden_size": 2048,
|
| 177 |
+
"text_vocab_size": 151936,
|
| 178 |
+
"use_cache": true,
|
| 179 |
+
"use_sliding_window": false,
|
| 180 |
+
"vocab_size": 3072
|
| 181 |
+
},
|
| 182 |
+
"transformers_version": "4.57.3"
|
| 183 |
+
}
|
qwen3-tts-12hz-1.7b-customvoice/config/speech_tokenizer.json
ADDED
|
@@ -0,0 +1,94 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3TTSTokenizerV2Model"
|
| 4 |
+
],
|
| 5 |
+
"model_type": "qwen3_tts_tokenizer_12hz",
|
| 6 |
+
"encoder_valid_num_quantizers": 16,
|
| 7 |
+
"input_sample_rate": 24000,
|
| 8 |
+
"output_sample_rate": 24000,
|
| 9 |
+
"decode_upsample_rate": 1920,
|
| 10 |
+
"encode_downsample_rate": 1920,
|
| 11 |
+
"decoder_config": {
|
| 12 |
+
"attention_bias": false,
|
| 13 |
+
"attention_dropout": 0.0,
|
| 14 |
+
"latent_dim": 1024,
|
| 15 |
+
"codebook_dim": 512,
|
| 16 |
+
"codebook_size": 2048,
|
| 17 |
+
"decoder_dim": 1536,
|
| 18 |
+
"hidden_act": "silu",
|
| 19 |
+
"hidden_size": 512,
|
| 20 |
+
"intermediate_size": 1024,
|
| 21 |
+
"layer_scale_initial_scale": 0.01,
|
| 22 |
+
"max_position_embeddings": 8000,
|
| 23 |
+
"head_dim": 64,
|
| 24 |
+
"num_attention_heads": 16,
|
| 25 |
+
"num_hidden_layers": 8,
|
| 26 |
+
"num_key_value_heads": 16,
|
| 27 |
+
"num_quantizers": 16,
|
| 28 |
+
"num_semantic_quantizers": 1,
|
| 29 |
+
"rms_norm_eps": 1e-05,
|
| 30 |
+
"rope_theta": 10000,
|
| 31 |
+
"semantic_codebook_size": 4096,
|
| 32 |
+
"sliding_window": 72,
|
| 33 |
+
"upsample_rates": [
|
| 34 |
+
8,
|
| 35 |
+
5,
|
| 36 |
+
4,
|
| 37 |
+
3
|
| 38 |
+
],
|
| 39 |
+
"upsampling_ratios": [
|
| 40 |
+
2,
|
| 41 |
+
2
|
| 42 |
+
],
|
| 43 |
+
"vector_quantization_hidden_dimension": 512
|
| 44 |
+
},
|
| 45 |
+
"encoder_config": {
|
| 46 |
+
"_frame_rate": 12.5,
|
| 47 |
+
"attention_bias": false,
|
| 48 |
+
"attention_dropout": 0.0,
|
| 49 |
+
"audio_channels": 1,
|
| 50 |
+
"codebook_dim": 256,
|
| 51 |
+
"codebook_size": 2048,
|
| 52 |
+
"compress": 2,
|
| 53 |
+
"dilation_growth_rate": 2,
|
| 54 |
+
"dtype": "float32",
|
| 55 |
+
"head_dim": 64,
|
| 56 |
+
"hidden_act": "gelu",
|
| 57 |
+
"hidden_size": 512,
|
| 58 |
+
"initializer_range": 0.02,
|
| 59 |
+
"intermediate_size": 2048,
|
| 60 |
+
"kernel_size": 7,
|
| 61 |
+
"last_kernel_size": 3,
|
| 62 |
+
"layer_scale_initial_scale": 0.01,
|
| 63 |
+
"max_position_embeddings": 8000,
|
| 64 |
+
"norm_eps": 1e-05,
|
| 65 |
+
"normalize": false,
|
| 66 |
+
"num_attention_heads": 8,
|
| 67 |
+
"num_filters": 64,
|
| 68 |
+
"num_hidden_layers": 8,
|
| 69 |
+
"num_key_value_heads": 8,
|
| 70 |
+
"num_quantizers": 32,
|
| 71 |
+
"num_residual_layers": 1,
|
| 72 |
+
"num_semantic_quantizers": 1,
|
| 73 |
+
"pad_mode": "constant",
|
| 74 |
+
"residual_kernel_size": 3,
|
| 75 |
+
"rope_theta": 10000.0,
|
| 76 |
+
"sampling_rate": 24000,
|
| 77 |
+
"sliding_window": 250,
|
| 78 |
+
"transformers_version": "4.57.0.dev0",
|
| 79 |
+
"trim_right_ratio": 1.0,
|
| 80 |
+
"upsample_groups": 512,
|
| 81 |
+
"upsampling_ratios": [
|
| 82 |
+
8,
|
| 83 |
+
6,
|
| 84 |
+
5,
|
| 85 |
+
4
|
| 86 |
+
],
|
| 87 |
+
"use_cache": false,
|
| 88 |
+
"use_causal_conv": true,
|
| 89 |
+
"use_conv_shortcut": false,
|
| 90 |
+
"use_streaming": false,
|
| 91 |
+
"vector_quantization_hidden_dimension": 256
|
| 92 |
+
},
|
| 93 |
+
"transformers_version": "4.57.3"
|
| 94 |
+
}
|
qwen3-tts-12hz-1.7b-customvoice/manifest.json
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"format": "mlx-bundle/1",
|
| 3 |
+
"components": {
|
| 4 |
+
"model": {
|
| 5 |
+
"class": "Qwen3TTSForConditionalGeneration",
|
| 6 |
+
"source": "/Volumes/work/project/mlx/downloads/Qwen3-TTS-12Hz-1.7B-CustomVoice",
|
| 7 |
+
"dtype_policy": "keep",
|
| 8 |
+
"num_tensors": 404,
|
| 9 |
+
"dtypes": {
|
| 10 |
+
"bfloat16": 404
|
| 11 |
+
},
|
| 12 |
+
"layout_rules": {
|
| 13 |
+
"identity": 404
|
| 14 |
+
},
|
| 15 |
+
"files": [
|
| 16 |
+
{
|
| 17 |
+
"file": "weights/model.safetensors",
|
| 18 |
+
"tensors": 404,
|
| 19 |
+
"bytes": 3833402584,
|
| 20 |
+
"sha256": "1b528965ec1fe58e172fa4241ce7d1e225cdb71f2332b2ca387f97f3f0583b34"
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"file": "config/model.json",
|
| 24 |
+
"tensors": 0,
|
| 25 |
+
"bytes": 4908,
|
| 26 |
+
"sha256": "17a07f527a1c25ea30b4e023a184482a23d3e279d697b1dc81b1bde498d29cf9"
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"file": "config/generation_config.json",
|
| 30 |
+
"tensors": 0,
|
| 31 |
+
"bytes": 245,
|
| 32 |
+
"sha256": "f1b90b4513f3b34c62851049e2492d7b4c5940daf1276f89c82b8ef04127f3aa"
|
| 33 |
+
}
|
| 34 |
+
]
|
| 35 |
+
},
|
| 36 |
+
"speech_tokenizer": {
|
| 37 |
+
"class": "Qwen3TTSTokenizerV2Model",
|
| 38 |
+
"source": "/Volumes/work/project/mlx/downloads/Qwen3-TTS-12Hz-1.7B-CustomVoice/speech_tokenizer",
|
| 39 |
+
"dtype_policy": "keep",
|
| 40 |
+
"num_tensors": 496,
|
| 41 |
+
"dtypes": {
|
| 42 |
+
"float32": 496
|
| 43 |
+
},
|
| 44 |
+
"layout_rules": {
|
| 45 |
+
"identity": 438,
|
| 46 |
+
"permute(0, 2, 1)": 52,
|
| 47 |
+
"permute(1, 2, 0)": 6
|
| 48 |
+
},
|
| 49 |
+
"files": [
|
| 50 |
+
{
|
| 51 |
+
"file": "weights/speech_tokenizer.safetensors",
|
| 52 |
+
"tensors": 496,
|
| 53 |
+
"bytes": 682293164,
|
| 54 |
+
"sha256": "e6ab4ca6473fd948c20089a3e8abe9d7caaeaa41f7597c768c2b230a6bb6c995"
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"file": "config/speech_tokenizer.json",
|
| 58 |
+
"tensors": 0,
|
| 59 |
+
"bytes": 2336,
|
| 60 |
+
"sha256": "ee65bb901c876664ab8707c487157aa1a6ee57c65969b28fb5ec9dc211e68167"
|
| 61 |
+
}
|
| 62 |
+
]
|
| 63 |
+
}
|
| 64 |
+
}
|
| 65 |
+
}
|
qwen3-tts-12hz-1.7b-customvoice/tokenizer/merges.txt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:599bab54075088774b1733fde865d5bd747cbcc7a547c5bc12610e874e26f5e3
|
| 3 |
+
size 1671839
|
qwen3-tts-12hz-1.7b-customvoice/tokenizer/preprocessor_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"padding_side": "left",
|
| 3 |
+
"padding_value": 0.0,
|
| 4 |
+
"processor_class": "Qwen3TTSProcessor",
|
| 5 |
+
"return_attention_mask": true
|
| 6 |
+
}
|
qwen3-tts-12hz-1.7b-customvoice/tokenizer/tokenizer_config.json
ADDED
|
@@ -0,0 +1,316 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": false,
|
| 3 |
+
"add_prefix_space": false,
|
| 4 |
+
"added_tokens_decoder": {
|
| 5 |
+
"151643": {
|
| 6 |
+
"content": "<|endoftext|>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false,
|
| 11 |
+
"special": true
|
| 12 |
+
},
|
| 13 |
+
"151644": {
|
| 14 |
+
"content": "<|im_start|>",
|
| 15 |
+
"lstrip": false,
|
| 16 |
+
"normalized": false,
|
| 17 |
+
"rstrip": false,
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"special": true
|
| 20 |
+
},
|
| 21 |
+
"151645": {
|
| 22 |
+
"content": "<|im_end|>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": false,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false,
|
| 27 |
+
"special": true
|
| 28 |
+
},
|
| 29 |
+
"151646": {
|
| 30 |
+
"content": "<|object_ref_start|>",
|
| 31 |
+
"lstrip": false,
|
| 32 |
+
"normalized": false,
|
| 33 |
+
"rstrip": false,
|
| 34 |
+
"single_word": false,
|
| 35 |
+
"special": true
|
| 36 |
+
},
|
| 37 |
+
"151647": {
|
| 38 |
+
"content": "<|object_ref_end|>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false,
|
| 43 |
+
"special": true
|
| 44 |
+
},
|
| 45 |
+
"151648": {
|
| 46 |
+
"content": "<|box_start|>",
|
| 47 |
+
"lstrip": false,
|
| 48 |
+
"normalized": false,
|
| 49 |
+
"rstrip": false,
|
| 50 |
+
"single_word": false,
|
| 51 |
+
"special": true
|
| 52 |
+
},
|
| 53 |
+
"151649": {
|
| 54 |
+
"content": "<|box_end|>",
|
| 55 |
+
"lstrip": false,
|
| 56 |
+
"normalized": false,
|
| 57 |
+
"rstrip": false,
|
| 58 |
+
"single_word": false,
|
| 59 |
+
"special": true
|
| 60 |
+
},
|
| 61 |
+
"151650": {
|
| 62 |
+
"content": "<|quad_start|>",
|
| 63 |
+
"lstrip": false,
|
| 64 |
+
"normalized": false,
|
| 65 |
+
"rstrip": false,
|
| 66 |
+
"single_word": false,
|
| 67 |
+
"special": true
|
| 68 |
+
},
|
| 69 |
+
"151651": {
|
| 70 |
+
"content": "<|quad_end|>",
|
| 71 |
+
"lstrip": false,
|
| 72 |
+
"normalized": false,
|
| 73 |
+
"rstrip": false,
|
| 74 |
+
"single_word": false,
|
| 75 |
+
"special": true
|
| 76 |
+
},
|
| 77 |
+
"151652": {
|
| 78 |
+
"content": "<|vision_start|>",
|
| 79 |
+
"lstrip": false,
|
| 80 |
+
"normalized": false,
|
| 81 |
+
"rstrip": false,
|
| 82 |
+
"single_word": false,
|
| 83 |
+
"special": true
|
| 84 |
+
},
|
| 85 |
+
"151653": {
|
| 86 |
+
"content": "<|vision_end|>",
|
| 87 |
+
"lstrip": false,
|
| 88 |
+
"normalized": false,
|
| 89 |
+
"rstrip": false,
|
| 90 |
+
"single_word": false,
|
| 91 |
+
"special": true
|
| 92 |
+
},
|
| 93 |
+
"151654": {
|
| 94 |
+
"content": "<|vision_pad|>",
|
| 95 |
+
"lstrip": false,
|
| 96 |
+
"normalized": false,
|
| 97 |
+
"rstrip": false,
|
| 98 |
+
"single_word": false,
|
| 99 |
+
"special": true
|
| 100 |
+
},
|
| 101 |
+
"151655": {
|
| 102 |
+
"content": "<|image_pad|>",
|
| 103 |
+
"lstrip": false,
|
| 104 |
+
"normalized": false,
|
| 105 |
+
"rstrip": false,
|
| 106 |
+
"single_word": false,
|
| 107 |
+
"special": true
|
| 108 |
+
},
|
| 109 |
+
"151656": {
|
| 110 |
+
"content": "<|video_pad|>",
|
| 111 |
+
"lstrip": false,
|
| 112 |
+
"normalized": false,
|
| 113 |
+
"rstrip": false,
|
| 114 |
+
"single_word": false,
|
| 115 |
+
"special": true
|
| 116 |
+
},
|
| 117 |
+
"151657": {
|
| 118 |
+
"content": "<tool_call>",
|
| 119 |
+
"lstrip": false,
|
| 120 |
+
"normalized": false,
|
| 121 |
+
"rstrip": false,
|
| 122 |
+
"single_word": false,
|
| 123 |
+
"special": false
|
| 124 |
+
},
|
| 125 |
+
"151658": {
|
| 126 |
+
"content": "</tool_call>",
|
| 127 |
+
"lstrip": false,
|
| 128 |
+
"normalized": false,
|
| 129 |
+
"rstrip": false,
|
| 130 |
+
"single_word": false,
|
| 131 |
+
"special": false
|
| 132 |
+
},
|
| 133 |
+
"151659": {
|
| 134 |
+
"content": "<|fim_prefix|>",
|
| 135 |
+
"lstrip": false,
|
| 136 |
+
"normalized": false,
|
| 137 |
+
"rstrip": false,
|
| 138 |
+
"single_word": false,
|
| 139 |
+
"special": false
|
| 140 |
+
},
|
| 141 |
+
"151660": {
|
| 142 |
+
"content": "<|fim_middle|>",
|
| 143 |
+
"lstrip": false,
|
| 144 |
+
"normalized": false,
|
| 145 |
+
"rstrip": false,
|
| 146 |
+
"single_word": false,
|
| 147 |
+
"special": false
|
| 148 |
+
},
|
| 149 |
+
"151661": {
|
| 150 |
+
"content": "<|fim_suffix|>",
|
| 151 |
+
"lstrip": false,
|
| 152 |
+
"normalized": false,
|
| 153 |
+
"rstrip": false,
|
| 154 |
+
"single_word": false,
|
| 155 |
+
"special": false
|
| 156 |
+
},
|
| 157 |
+
"151662": {
|
| 158 |
+
"content": "<|fim_pad|>",
|
| 159 |
+
"lstrip": false,
|
| 160 |
+
"normalized": false,
|
| 161 |
+
"rstrip": false,
|
| 162 |
+
"single_word": false,
|
| 163 |
+
"special": false
|
| 164 |
+
},
|
| 165 |
+
"151663": {
|
| 166 |
+
"content": "<|repo_name|>",
|
| 167 |
+
"lstrip": false,
|
| 168 |
+
"normalized": false,
|
| 169 |
+
"rstrip": false,
|
| 170 |
+
"single_word": false,
|
| 171 |
+
"special": false
|
| 172 |
+
},
|
| 173 |
+
"151664": {
|
| 174 |
+
"content": "<|file_sep|>",
|
| 175 |
+
"lstrip": false,
|
| 176 |
+
"normalized": false,
|
| 177 |
+
"rstrip": false,
|
| 178 |
+
"single_word": false,
|
| 179 |
+
"special": false
|
| 180 |
+
},
|
| 181 |
+
"151665": {
|
| 182 |
+
"content": "<tool_response>",
|
| 183 |
+
"lstrip": false,
|
| 184 |
+
"normalized": false,
|
| 185 |
+
"rstrip": false,
|
| 186 |
+
"single_word": false,
|
| 187 |
+
"special": false
|
| 188 |
+
},
|
| 189 |
+
"151666": {
|
| 190 |
+
"content": "</tool_response>",
|
| 191 |
+
"lstrip": false,
|
| 192 |
+
"normalized": false,
|
| 193 |
+
"rstrip": false,
|
| 194 |
+
"single_word": false,
|
| 195 |
+
"special": false
|
| 196 |
+
},
|
| 197 |
+
"151667": {
|
| 198 |
+
"content": "<think>",
|
| 199 |
+
"lstrip": false,
|
| 200 |
+
"normalized": false,
|
| 201 |
+
"rstrip": false,
|
| 202 |
+
"single_word": false,
|
| 203 |
+
"special": false
|
| 204 |
+
},
|
| 205 |
+
"151668": {
|
| 206 |
+
"content": "</think>",
|
| 207 |
+
"lstrip": false,
|
| 208 |
+
"normalized": false,
|
| 209 |
+
"rstrip": false,
|
| 210 |
+
"single_word": false,
|
| 211 |
+
"special": false
|
| 212 |
+
},
|
| 213 |
+
"151669": {
|
| 214 |
+
"content": "<|audio_start|>",
|
| 215 |
+
"lstrip": false,
|
| 216 |
+
"normalized": false,
|
| 217 |
+
"rstrip": false,
|
| 218 |
+
"single_word": false,
|
| 219 |
+
"special": true
|
| 220 |
+
},
|
| 221 |
+
"151670": {
|
| 222 |
+
"content": "<|audio_end|>",
|
| 223 |
+
"lstrip": false,
|
| 224 |
+
"normalized": false,
|
| 225 |
+
"rstrip": false,
|
| 226 |
+
"single_word": false,
|
| 227 |
+
"special": true
|
| 228 |
+
},
|
| 229 |
+
"151671": {
|
| 230 |
+
"content": "<tts_pad>",
|
| 231 |
+
"lstrip": false,
|
| 232 |
+
"normalized": false,
|
| 233 |
+
"rstrip": false,
|
| 234 |
+
"single_word": false,
|
| 235 |
+
"special": true
|
| 236 |
+
},
|
| 237 |
+
"151672": {
|
| 238 |
+
"content": "<tts_text_bos>",
|
| 239 |
+
"lstrip": false,
|
| 240 |
+
"normalized": false,
|
| 241 |
+
"rstrip": false,
|
| 242 |
+
"single_word": false,
|
| 243 |
+
"special": true
|
| 244 |
+
},
|
| 245 |
+
"151673": {
|
| 246 |
+
"content": "<tts_text_eod>",
|
| 247 |
+
"lstrip": false,
|
| 248 |
+
"normalized": false,
|
| 249 |
+
"rstrip": false,
|
| 250 |
+
"single_word": false,
|
| 251 |
+
"special": true
|
| 252 |
+
},
|
| 253 |
+
"151674": {
|
| 254 |
+
"content": "<tts_text_bos_single>",
|
| 255 |
+
"lstrip": false,
|
| 256 |
+
"normalized": false,
|
| 257 |
+
"rstrip": false,
|
| 258 |
+
"single_word": false,
|
| 259 |
+
"special": true
|
| 260 |
+
},
|
| 261 |
+
"151675": {
|
| 262 |
+
"content": "<|audio_pad|>",
|
| 263 |
+
"lstrip": false,
|
| 264 |
+
"normalized": false,
|
| 265 |
+
"rstrip": false,
|
| 266 |
+
"single_word": false,
|
| 267 |
+
"special": true
|
| 268 |
+
}
|
| 269 |
+
},
|
| 270 |
+
"additional_special_tokens": [
|
| 271 |
+
"<|im_start|>",
|
| 272 |
+
"<|im_end|>",
|
| 273 |
+
"<|object_ref_start|>",
|
| 274 |
+
"<|object_ref_end|>",
|
| 275 |
+
"<|box_start|>",
|
| 276 |
+
"<|box_end|>",
|
| 277 |
+
"<|quad_start|>",
|
| 278 |
+
"<|quad_end|>",
|
| 279 |
+
"<|vision_start|>",
|
| 280 |
+
"<|vision_end|>",
|
| 281 |
+
"<|vision_pad|>",
|
| 282 |
+
"<|image_pad|>",
|
| 283 |
+
"<|video_pad|>",
|
| 284 |
+
"<|audio_start|>",
|
| 285 |
+
"<|audio_end|>",
|
| 286 |
+
"<tts_pad>",
|
| 287 |
+
"<tts_text_bos>",
|
| 288 |
+
"<tts_text_bos_single>",
|
| 289 |
+
"<|audio_pad|>"
|
| 290 |
+
],
|
| 291 |
+
"extra_special_tokens": {
|
| 292 |
+
"image_token": "<|image_pad|>",
|
| 293 |
+
"audio_token": "<|audio_pad|>",
|
| 294 |
+
"video_token": "<|video_pad|>",
|
| 295 |
+
"vision_bos_token": "<|vision_start|>",
|
| 296 |
+
"vision_eos_token": "<|vision_end|>",
|
| 297 |
+
"audio_bos_token": "<|audio_start|>",
|
| 298 |
+
"audio_eos_token": "<|audio_end|>"
|
| 299 |
+
},
|
| 300 |
+
"bos_token": null,
|
| 301 |
+
"clean_up_tokenization_spaces": false,
|
| 302 |
+
"eos_token": "<|im_end|>",
|
| 303 |
+
"errors": "replace",
|
| 304 |
+
"model_max_length": 131072,
|
| 305 |
+
"pad_token": "<|endoftext|>",
|
| 306 |
+
"split_special_tokens": false,
|
| 307 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 308 |
+
"unk_token": null,
|
| 309 |
+
"image_token": "<|image_pad|>",
|
| 310 |
+
"audio_token": "<|audio_pad|>",
|
| 311 |
+
"video_token": "<|video_pad|>",
|
| 312 |
+
"vision_bos_token": "<|vision_start|>",
|
| 313 |
+
"vision_eos_token": "<|vision_end|>",
|
| 314 |
+
"audio_bos_token": "<|audio_start|>",
|
| 315 |
+
"audio_eos_token": "<|audio_end|>"
|
| 316 |
+
}
|
qwen3-tts-12hz-1.7b-customvoice/tokenizer/vocab.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ca10d7e9fb3ed18575dd1e277a2579c16d108e32f27439684afa0e10b1440910
|
| 3 |
+
size 2776833
|
qwen3-tts-12hz-1.7b-customvoice/weights/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:1b528965ec1fe58e172fa4241ce7d1e225cdb71f2332b2ca387f97f3f0583b34
|
| 3 |
+
size 3833402584
|
qwen3-tts-12hz-1.7b-customvoice/weights/speech_tokenizer.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e6ab4ca6473fd948c20089a3e8abe9d7caaeaa41f7597c768c2b230a6bb6c995
|
| 3 |
+
size 682293164
|
qwen3-tts-12hz-1.7b-voicedesign/config/generation_config.json
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"do_sample": true,
|
| 3 |
+
"repetition_penalty": 1.05,
|
| 4 |
+
"temperature": 0.9,
|
| 5 |
+
"top_p": 1.0,
|
| 6 |
+
"top_k": 50,
|
| 7 |
+
"subtalker_dosample": true,
|
| 8 |
+
"subtalker_temperature": 0.9,
|
| 9 |
+
"subtalker_top_p": 1.0,
|
| 10 |
+
"subtalker_top_k": 50,
|
| 11 |
+
"max_new_tokens": 8192
|
| 12 |
+
}
|
qwen3-tts-12hz-1.7b-voicedesign/config/model.json
ADDED
|
@@ -0,0 +1,163 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3TTSForConditionalGeneration"
|
| 4 |
+
],
|
| 5 |
+
"assistant_token_id": 77091,
|
| 6 |
+
"im_end_token_id": 151645,
|
| 7 |
+
"im_start_token_id": 151644,
|
| 8 |
+
"tts_bos_token_id": 151672,
|
| 9 |
+
"tts_eos_token_id": 151673,
|
| 10 |
+
"tts_pad_token_id": 151671,
|
| 11 |
+
"model_type": "qwen3_tts",
|
| 12 |
+
"tokenizer_type": "qwen3_tts_tokenizer_12hz",
|
| 13 |
+
"tts_model_size": "1b7",
|
| 14 |
+
"tts_model_type": "voice_design",
|
| 15 |
+
"talker_config": {
|
| 16 |
+
"attention_bias": false,
|
| 17 |
+
"attention_dropout": 0,
|
| 18 |
+
"code_predictor_config": {
|
| 19 |
+
"_name_or_path": "",
|
| 20 |
+
"add_cross_attention": false,
|
| 21 |
+
"architectures": null,
|
| 22 |
+
"attention_bias": false,
|
| 23 |
+
"attention_dropout": 0,
|
| 24 |
+
"bad_words_ids": null,
|
| 25 |
+
"begin_suppress_tokens": null,
|
| 26 |
+
"bos_token_id": null,
|
| 27 |
+
"chunk_size_feed_forward": 0,
|
| 28 |
+
"cross_attention_hidden_size": null,
|
| 29 |
+
"decoder_start_token_id": null,
|
| 30 |
+
"diversity_penalty": 0.0,
|
| 31 |
+
"do_sample": false,
|
| 32 |
+
"early_stopping": false,
|
| 33 |
+
"encoder_no_repeat_ngram_size": 0,
|
| 34 |
+
"eos_token_id": null,
|
| 35 |
+
"exponential_decay_length_penalty": null,
|
| 36 |
+
"finetuning_task": null,
|
| 37 |
+
"forced_bos_token_id": null,
|
| 38 |
+
"forced_eos_token_id": null,
|
| 39 |
+
"head_dim": 128,
|
| 40 |
+
"hidden_act": "silu",
|
| 41 |
+
"hidden_size": 1024,
|
| 42 |
+
"id2label": {
|
| 43 |
+
"0": "LABEL_0",
|
| 44 |
+
"1": "LABEL_1"
|
| 45 |
+
},
|
| 46 |
+
"initializer_range": 0.02,
|
| 47 |
+
"intermediate_size": 3072,
|
| 48 |
+
"is_decoder": false,
|
| 49 |
+
"is_encoder_decoder": false,
|
| 50 |
+
"label2id": {
|
| 51 |
+
"LABEL_0": 0,
|
| 52 |
+
"LABEL_1": 1
|
| 53 |
+
},
|
| 54 |
+
"layer_types": [
|
| 55 |
+
"full_attention",
|
| 56 |
+
"full_attention",
|
| 57 |
+
"full_attention",
|
| 58 |
+
"full_attention",
|
| 59 |
+
"full_attention"
|
| 60 |
+
],
|
| 61 |
+
"length_penalty": 1.0,
|
| 62 |
+
"max_length": 20,
|
| 63 |
+
"max_position_embeddings": 65536,
|
| 64 |
+
"max_window_layers": 28,
|
| 65 |
+
"min_length": 0,
|
| 66 |
+
"model_type": "qwen3_tts_talker_code_predictor",
|
| 67 |
+
"no_repeat_ngram_size": 0,
|
| 68 |
+
"num_attention_heads": 16,
|
| 69 |
+
"num_beam_groups": 1,
|
| 70 |
+
"num_beams": 1,
|
| 71 |
+
"num_code_groups": 16,
|
| 72 |
+
"num_hidden_layers": 5,
|
| 73 |
+
"num_key_value_heads": 8,
|
| 74 |
+
"num_return_sequences": 1,
|
| 75 |
+
"output_attentions": false,
|
| 76 |
+
"output_hidden_states": false,
|
| 77 |
+
"output_scores": false,
|
| 78 |
+
"pad_token_id": null,
|
| 79 |
+
"prefix": null,
|
| 80 |
+
"problem_type": null,
|
| 81 |
+
"pruned_heads": {},
|
| 82 |
+
"remove_invalid_values": false,
|
| 83 |
+
"repetition_penalty": 1.0,
|
| 84 |
+
"return_dict": true,
|
| 85 |
+
"return_dict_in_generate": false,
|
| 86 |
+
"rms_norm_eps": 1e-06,
|
| 87 |
+
"rope_scaling": null,
|
| 88 |
+
"rope_theta": 1000000,
|
| 89 |
+
"sep_token_id": null,
|
| 90 |
+
"sliding_window": null,
|
| 91 |
+
"suppress_tokens": null,
|
| 92 |
+
"task_specific_params": null,
|
| 93 |
+
"temperature": 1.0,
|
| 94 |
+
"tf_legacy_loss": false,
|
| 95 |
+
"tie_encoder_decoder": false,
|
| 96 |
+
"tie_word_embeddings": false,
|
| 97 |
+
"tokenizer_class": null,
|
| 98 |
+
"top_k": 50,
|
| 99 |
+
"top_p": 1.0,
|
| 100 |
+
"dtype": null,
|
| 101 |
+
"torchscript": false,
|
| 102 |
+
"typical_p": 1.0,
|
| 103 |
+
"use_bfloat16": false,
|
| 104 |
+
"use_cache": true,
|
| 105 |
+
"use_sliding_window": false,
|
| 106 |
+
"vocab_size": 2048
|
| 107 |
+
},
|
| 108 |
+
"codec_bos_id": 2149,
|
| 109 |
+
"codec_eos_token_id": 2150,
|
| 110 |
+
"codec_think_id": 2154,
|
| 111 |
+
"codec_language_id": {
|
| 112 |
+
"chinese": 2055,
|
| 113 |
+
"english": 2050,
|
| 114 |
+
"german": 2053,
|
| 115 |
+
"italian": 2070,
|
| 116 |
+
"portuguese": 2071,
|
| 117 |
+
"spanish": 2054,
|
| 118 |
+
"japanese": 2058,
|
| 119 |
+
"korean": 2064,
|
| 120 |
+
"french": 2061,
|
| 121 |
+
"russian": 2069
|
| 122 |
+
},
|
| 123 |
+
"codec_nothink_id": 2155,
|
| 124 |
+
"codec_pad_id": 2148,
|
| 125 |
+
"codec_think_bos_id": 2156,
|
| 126 |
+
"codec_think_eos_id": 2157,
|
| 127 |
+
"spk_id": {
|
| 128 |
+
},
|
| 129 |
+
"spk_is_dialect": {
|
| 130 |
+
},
|
| 131 |
+
"head_dim": 128,
|
| 132 |
+
"hidden_act": "silu",
|
| 133 |
+
"hidden_size": 2048,
|
| 134 |
+
"initializer_range": 0.02,
|
| 135 |
+
"intermediate_size": 6144,
|
| 136 |
+
"max_position_embeddings": 32768,
|
| 137 |
+
"model_type": "qwen3_tts_talker",
|
| 138 |
+
"num_attention_heads": 16,
|
| 139 |
+
"num_code_groups": 16,
|
| 140 |
+
"num_hidden_layers": 28,
|
| 141 |
+
"num_key_value_heads": 8,
|
| 142 |
+
"position_id_per_seconds": 13,
|
| 143 |
+
"rms_norm_eps": 1e-06,
|
| 144 |
+
"rope_scaling": {
|
| 145 |
+
"interleaved": true,
|
| 146 |
+
"mrope_section": [
|
| 147 |
+
24,
|
| 148 |
+
20,
|
| 149 |
+
20
|
| 150 |
+
],
|
| 151 |
+
"rope_type": "default",
|
| 152 |
+
"type": "default"
|
| 153 |
+
},
|
| 154 |
+
"rope_theta": 1000000,
|
| 155 |
+
"sliding_window": null,
|
| 156 |
+
"text_hidden_size": 2048,
|
| 157 |
+
"text_vocab_size": 151936,
|
| 158 |
+
"use_cache": true,
|
| 159 |
+
"use_sliding_window": false,
|
| 160 |
+
"vocab_size": 3072
|
| 161 |
+
},
|
| 162 |
+
"transformers_version": "4.57.3"
|
| 163 |
+
}
|
qwen3-tts-12hz-1.7b-voicedesign/config/speech_tokenizer.json
ADDED
|
@@ -0,0 +1,94 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen3TTSTokenizerV2Model"
|
| 4 |
+
],
|
| 5 |
+
"model_type": "qwen3_tts_tokenizer_12hz",
|
| 6 |
+
"encoder_valid_num_quantizers": 16,
|
| 7 |
+
"input_sample_rate": 24000,
|
| 8 |
+
"output_sample_rate": 24000,
|
| 9 |
+
"decode_upsample_rate": 1920,
|
| 10 |
+
"encode_downsample_rate": 1920,
|
| 11 |
+
"decoder_config": {
|
| 12 |
+
"attention_bias": false,
|
| 13 |
+
"attention_dropout": 0.0,
|
| 14 |
+
"latent_dim": 1024,
|
| 15 |
+
"codebook_dim": 512,
|
| 16 |
+
"codebook_size": 2048,
|
| 17 |
+
"decoder_dim": 1536,
|
| 18 |
+
"hidden_act": "silu",
|
| 19 |
+
"hidden_size": 512,
|
| 20 |
+
"intermediate_size": 1024,
|
| 21 |
+
"layer_scale_initial_scale": 0.01,
|
| 22 |
+
"max_position_embeddings": 8000,
|
| 23 |
+
"head_dim": 64,
|
| 24 |
+
"num_attention_heads": 16,
|
| 25 |
+
"num_hidden_layers": 8,
|
| 26 |
+
"num_key_value_heads": 16,
|
| 27 |
+
"num_quantizers": 16,
|
| 28 |
+
"num_semantic_quantizers": 1,
|
| 29 |
+
"rms_norm_eps": 1e-05,
|
| 30 |
+
"rope_theta": 10000,
|
| 31 |
+
"semantic_codebook_size": 4096,
|
| 32 |
+
"sliding_window": 72,
|
| 33 |
+
"upsample_rates": [
|
| 34 |
+
8,
|
| 35 |
+
5,
|
| 36 |
+
4,
|
| 37 |
+
3
|
| 38 |
+
],
|
| 39 |
+
"upsampling_ratios": [
|
| 40 |
+
2,
|
| 41 |
+
2
|
| 42 |
+
],
|
| 43 |
+
"vector_quantization_hidden_dimension": 512
|
| 44 |
+
},
|
| 45 |
+
"encoder_config": {
|
| 46 |
+
"_frame_rate": 12.5,
|
| 47 |
+
"attention_bias": false,
|
| 48 |
+
"attention_dropout": 0.0,
|
| 49 |
+
"audio_channels": 1,
|
| 50 |
+
"codebook_dim": 256,
|
| 51 |
+
"codebook_size": 2048,
|
| 52 |
+
"compress": 2,
|
| 53 |
+
"dilation_growth_rate": 2,
|
| 54 |
+
"dtype": "float32",
|
| 55 |
+
"head_dim": 64,
|
| 56 |
+
"hidden_act": "gelu",
|
| 57 |
+
"hidden_size": 512,
|
| 58 |
+
"initializer_range": 0.02,
|
| 59 |
+
"intermediate_size": 2048,
|
| 60 |
+
"kernel_size": 7,
|
| 61 |
+
"last_kernel_size": 3,
|
| 62 |
+
"layer_scale_initial_scale": 0.01,
|
| 63 |
+
"max_position_embeddings": 8000,
|
| 64 |
+
"norm_eps": 1e-05,
|
| 65 |
+
"normalize": false,
|
| 66 |
+
"num_attention_heads": 8,
|
| 67 |
+
"num_filters": 64,
|
| 68 |
+
"num_hidden_layers": 8,
|
| 69 |
+
"num_key_value_heads": 8,
|
| 70 |
+
"num_quantizers": 32,
|
| 71 |
+
"num_residual_layers": 1,
|
| 72 |
+
"num_semantic_quantizers": 1,
|
| 73 |
+
"pad_mode": "constant",
|
| 74 |
+
"residual_kernel_size": 3,
|
| 75 |
+
"rope_theta": 10000.0,
|
| 76 |
+
"sampling_rate": 24000,
|
| 77 |
+
"sliding_window": 250,
|
| 78 |
+
"transformers_version": "4.57.0.dev0",
|
| 79 |
+
"trim_right_ratio": 1.0,
|
| 80 |
+
"upsample_groups": 512,
|
| 81 |
+
"upsampling_ratios": [
|
| 82 |
+
8,
|
| 83 |
+
6,
|
| 84 |
+
5,
|
| 85 |
+
4
|
| 86 |
+
],
|
| 87 |
+
"use_cache": false,
|
| 88 |
+
"use_causal_conv": true,
|
| 89 |
+
"use_conv_shortcut": false,
|
| 90 |
+
"use_streaming": false,
|
| 91 |
+
"vector_quantization_hidden_dimension": 256
|
| 92 |
+
},
|
| 93 |
+
"transformers_version": "4.57.3"
|
| 94 |
+
}
|
qwen3-tts-12hz-1.7b-voicedesign/manifest.json
ADDED
|
@@ -0,0 +1,65 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"format": "mlx-bundle/1",
|
| 3 |
+
"components": {
|
| 4 |
+
"model": {
|
| 5 |
+
"class": "Qwen3TTSForConditionalGeneration",
|
| 6 |
+
"source": "/Volumes/work/project/mlx/downloads/Qwen3-TTS-12Hz-1.7B-VoiceDesign",
|
| 7 |
+
"dtype_policy": "keep",
|
| 8 |
+
"num_tensors": 404,
|
| 9 |
+
"dtypes": {
|
| 10 |
+
"bfloat16": 404
|
| 11 |
+
},
|
| 12 |
+
"layout_rules": {
|
| 13 |
+
"identity": 404
|
| 14 |
+
},
|
| 15 |
+
"files": [
|
| 16 |
+
{
|
| 17 |
+
"file": "weights/model.safetensors",
|
| 18 |
+
"tensors": 404,
|
| 19 |
+
"bytes": 3833402584,
|
| 20 |
+
"sha256": "c0240a5d663f431683c63606770665321ef21594dbd7559fd4bd2154b26091ea"
|
| 21 |
+
},
|
| 22 |
+
{
|
| 23 |
+
"file": "config/model.json",
|
| 24 |
+
"tensors": 0,
|
| 25 |
+
"bytes": 4421,
|
| 26 |
+
"sha256": "aecd2cc4c1fe9edef1cb7ca7c401685a43879ad43f3f9e883f1c6760b61731e0"
|
| 27 |
+
},
|
| 28 |
+
{
|
| 29 |
+
"file": "config/generation_config.json",
|
| 30 |
+
"tensors": 0,
|
| 31 |
+
"bytes": 245,
|
| 32 |
+
"sha256": "f1b90b4513f3b34c62851049e2492d7b4c5940daf1276f89c82b8ef04127f3aa"
|
| 33 |
+
}
|
| 34 |
+
]
|
| 35 |
+
},
|
| 36 |
+
"speech_tokenizer": {
|
| 37 |
+
"class": "Qwen3TTSTokenizerV2Model",
|
| 38 |
+
"source": "/Volumes/work/project/mlx/downloads/Qwen3-TTS-12Hz-1.7B-VoiceDesign/speech_tokenizer",
|
| 39 |
+
"dtype_policy": "keep",
|
| 40 |
+
"num_tensors": 496,
|
| 41 |
+
"dtypes": {
|
| 42 |
+
"float32": 496
|
| 43 |
+
},
|
| 44 |
+
"layout_rules": {
|
| 45 |
+
"identity": 438,
|
| 46 |
+
"permute(0, 2, 1)": 52,
|
| 47 |
+
"permute(1, 2, 0)": 6
|
| 48 |
+
},
|
| 49 |
+
"files": [
|
| 50 |
+
{
|
| 51 |
+
"file": "weights/speech_tokenizer.safetensors",
|
| 52 |
+
"tensors": 496,
|
| 53 |
+
"bytes": 682293164,
|
| 54 |
+
"sha256": "ee778743f0734e5e35f751f771ffe42a8943ca3ce4c9cc2e39b58719989c20c6"
|
| 55 |
+
},
|
| 56 |
+
{
|
| 57 |
+
"file": "config/speech_tokenizer.json",
|
| 58 |
+
"tensors": 0,
|
| 59 |
+
"bytes": 2336,
|
| 60 |
+
"sha256": "ee65bb901c876664ab8707c487157aa1a6ee57c65969b28fb5ec9dc211e68167"
|
| 61 |
+
}
|
| 62 |
+
]
|
| 63 |
+
}
|
| 64 |
+
}
|
| 65 |
+
}
|
qwen3-tts-12hz-1.7b-voicedesign/tokenizer/merges.txt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:599bab54075088774b1733fde865d5bd747cbcc7a547c5bc12610e874e26f5e3
|
| 3 |
+
size 1671839
|
qwen3-tts-12hz-1.7b-voicedesign/tokenizer/preprocessor_config.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"padding_side": "left",
|
| 3 |
+
"padding_value": 0.0,
|
| 4 |
+
"processor_class": "Qwen3TTSProcessor",
|
| 5 |
+
"return_attention_mask": true
|
| 6 |
+
}
|
qwen3-tts-12hz-1.7b-voicedesign/tokenizer/tokenizer_config.json
ADDED
|
@@ -0,0 +1,316 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": false,
|
| 3 |
+
"add_prefix_space": false,
|
| 4 |
+
"added_tokens_decoder": {
|
| 5 |
+
"151643": {
|
| 6 |
+
"content": "<|endoftext|>",
|
| 7 |
+
"lstrip": false,
|
| 8 |
+
"normalized": false,
|
| 9 |
+
"rstrip": false,
|
| 10 |
+
"single_word": false,
|
| 11 |
+
"special": true
|
| 12 |
+
},
|
| 13 |
+
"151644": {
|
| 14 |
+
"content": "<|im_start|>",
|
| 15 |
+
"lstrip": false,
|
| 16 |
+
"normalized": false,
|
| 17 |
+
"rstrip": false,
|
| 18 |
+
"single_word": false,
|
| 19 |
+
"special": true
|
| 20 |
+
},
|
| 21 |
+
"151645": {
|
| 22 |
+
"content": "<|im_end|>",
|
| 23 |
+
"lstrip": false,
|
| 24 |
+
"normalized": false,
|
| 25 |
+
"rstrip": false,
|
| 26 |
+
"single_word": false,
|
| 27 |
+
"special": true
|
| 28 |
+
},
|
| 29 |
+
"151646": {
|
| 30 |
+
"content": "<|object_ref_start|>",
|
| 31 |
+
"lstrip": false,
|
| 32 |
+
"normalized": false,
|
| 33 |
+
"rstrip": false,
|
| 34 |
+
"single_word": false,
|
| 35 |
+
"special": true
|
| 36 |
+
},
|
| 37 |
+
"151647": {
|
| 38 |
+
"content": "<|object_ref_end|>",
|
| 39 |
+
"lstrip": false,
|
| 40 |
+
"normalized": false,
|
| 41 |
+
"rstrip": false,
|
| 42 |
+
"single_word": false,
|
| 43 |
+
"special": true
|
| 44 |
+
},
|
| 45 |
+
"151648": {
|
| 46 |
+
"content": "<|box_start|>",
|
| 47 |
+
"lstrip": false,
|
| 48 |
+
"normalized": false,
|
| 49 |
+
"rstrip": false,
|
| 50 |
+
"single_word": false,
|
| 51 |
+
"special": true
|
| 52 |
+
},
|
| 53 |
+
"151649": {
|
| 54 |
+
"content": "<|box_end|>",
|
| 55 |
+
"lstrip": false,
|
| 56 |
+
"normalized": false,
|
| 57 |
+
"rstrip": false,
|
| 58 |
+
"single_word": false,
|
| 59 |
+
"special": true
|
| 60 |
+
},
|
| 61 |
+
"151650": {
|
| 62 |
+
"content": "<|quad_start|>",
|
| 63 |
+
"lstrip": false,
|
| 64 |
+
"normalized": false,
|
| 65 |
+
"rstrip": false,
|
| 66 |
+
"single_word": false,
|
| 67 |
+
"special": true
|
| 68 |
+
},
|
| 69 |
+
"151651": {
|
| 70 |
+
"content": "<|quad_end|>",
|
| 71 |
+
"lstrip": false,
|
| 72 |
+
"normalized": false,
|
| 73 |
+
"rstrip": false,
|
| 74 |
+
"single_word": false,
|
| 75 |
+
"special": true
|
| 76 |
+
},
|
| 77 |
+
"151652": {
|
| 78 |
+
"content": "<|vision_start|>",
|
| 79 |
+
"lstrip": false,
|
| 80 |
+
"normalized": false,
|
| 81 |
+
"rstrip": false,
|
| 82 |
+
"single_word": false,
|
| 83 |
+
"special": true
|
| 84 |
+
},
|
| 85 |
+
"151653": {
|
| 86 |
+
"content": "<|vision_end|>",
|
| 87 |
+
"lstrip": false,
|
| 88 |
+
"normalized": false,
|
| 89 |
+
"rstrip": false,
|
| 90 |
+
"single_word": false,
|
| 91 |
+
"special": true
|
| 92 |
+
},
|
| 93 |
+
"151654": {
|
| 94 |
+
"content": "<|vision_pad|>",
|
| 95 |
+
"lstrip": false,
|
| 96 |
+
"normalized": false,
|
| 97 |
+
"rstrip": false,
|
| 98 |
+
"single_word": false,
|
| 99 |
+
"special": true
|
| 100 |
+
},
|
| 101 |
+
"151655": {
|
| 102 |
+
"content": "<|image_pad|>",
|
| 103 |
+
"lstrip": false,
|
| 104 |
+
"normalized": false,
|
| 105 |
+
"rstrip": false,
|
| 106 |
+
"single_word": false,
|
| 107 |
+
"special": true
|
| 108 |
+
},
|
| 109 |
+
"151656": {
|
| 110 |
+
"content": "<|video_pad|>",
|
| 111 |
+
"lstrip": false,
|
| 112 |
+
"normalized": false,
|
| 113 |
+
"rstrip": false,
|
| 114 |
+
"single_word": false,
|
| 115 |
+
"special": true
|
| 116 |
+
},
|
| 117 |
+
"151657": {
|
| 118 |
+
"content": "<tool_call>",
|
| 119 |
+
"lstrip": false,
|
| 120 |
+
"normalized": false,
|
| 121 |
+
"rstrip": false,
|
| 122 |
+
"single_word": false,
|
| 123 |
+
"special": false
|
| 124 |
+
},
|
| 125 |
+
"151658": {
|
| 126 |
+
"content": "</tool_call>",
|
| 127 |
+
"lstrip": false,
|
| 128 |
+
"normalized": false,
|
| 129 |
+
"rstrip": false,
|
| 130 |
+
"single_word": false,
|
| 131 |
+
"special": false
|
| 132 |
+
},
|
| 133 |
+
"151659": {
|
| 134 |
+
"content": "<|fim_prefix|>",
|
| 135 |
+
"lstrip": false,
|
| 136 |
+
"normalized": false,
|
| 137 |
+
"rstrip": false,
|
| 138 |
+
"single_word": false,
|
| 139 |
+
"special": false
|
| 140 |
+
},
|
| 141 |
+
"151660": {
|
| 142 |
+
"content": "<|fim_middle|>",
|
| 143 |
+
"lstrip": false,
|
| 144 |
+
"normalized": false,
|
| 145 |
+
"rstrip": false,
|
| 146 |
+
"single_word": false,
|
| 147 |
+
"special": false
|
| 148 |
+
},
|
| 149 |
+
"151661": {
|
| 150 |
+
"content": "<|fim_suffix|>",
|
| 151 |
+
"lstrip": false,
|
| 152 |
+
"normalized": false,
|
| 153 |
+
"rstrip": false,
|
| 154 |
+
"single_word": false,
|
| 155 |
+
"special": false
|
| 156 |
+
},
|
| 157 |
+
"151662": {
|
| 158 |
+
"content": "<|fim_pad|>",
|
| 159 |
+
"lstrip": false,
|
| 160 |
+
"normalized": false,
|
| 161 |
+
"rstrip": false,
|
| 162 |
+
"single_word": false,
|
| 163 |
+
"special": false
|
| 164 |
+
},
|
| 165 |
+
"151663": {
|
| 166 |
+
"content": "<|repo_name|>",
|
| 167 |
+
"lstrip": false,
|
| 168 |
+
"normalized": false,
|
| 169 |
+
"rstrip": false,
|
| 170 |
+
"single_word": false,
|
| 171 |
+
"special": false
|
| 172 |
+
},
|
| 173 |
+
"151664": {
|
| 174 |
+
"content": "<|file_sep|>",
|
| 175 |
+
"lstrip": false,
|
| 176 |
+
"normalized": false,
|
| 177 |
+
"rstrip": false,
|
| 178 |
+
"single_word": false,
|
| 179 |
+
"special": false
|
| 180 |
+
},
|
| 181 |
+
"151665": {
|
| 182 |
+
"content": "<tool_response>",
|
| 183 |
+
"lstrip": false,
|
| 184 |
+
"normalized": false,
|
| 185 |
+
"rstrip": false,
|
| 186 |
+
"single_word": false,
|
| 187 |
+
"special": false
|
| 188 |
+
},
|
| 189 |
+
"151666": {
|
| 190 |
+
"content": "</tool_response>",
|
| 191 |
+
"lstrip": false,
|
| 192 |
+
"normalized": false,
|
| 193 |
+
"rstrip": false,
|
| 194 |
+
"single_word": false,
|
| 195 |
+
"special": false
|
| 196 |
+
},
|
| 197 |
+
"151667": {
|
| 198 |
+
"content": "<think>",
|
| 199 |
+
"lstrip": false,
|
| 200 |
+
"normalized": false,
|
| 201 |
+
"rstrip": false,
|
| 202 |
+
"single_word": false,
|
| 203 |
+
"special": false
|
| 204 |
+
},
|
| 205 |
+
"151668": {
|
| 206 |
+
"content": "</think>",
|
| 207 |
+
"lstrip": false,
|
| 208 |
+
"normalized": false,
|
| 209 |
+
"rstrip": false,
|
| 210 |
+
"single_word": false,
|
| 211 |
+
"special": false
|
| 212 |
+
},
|
| 213 |
+
"151669": {
|
| 214 |
+
"content": "<|audio_start|>",
|
| 215 |
+
"lstrip": false,
|
| 216 |
+
"normalized": false,
|
| 217 |
+
"rstrip": false,
|
| 218 |
+
"single_word": false,
|
| 219 |
+
"special": true
|
| 220 |
+
},
|
| 221 |
+
"151670": {
|
| 222 |
+
"content": "<|audio_end|>",
|
| 223 |
+
"lstrip": false,
|
| 224 |
+
"normalized": false,
|
| 225 |
+
"rstrip": false,
|
| 226 |
+
"single_word": false,
|
| 227 |
+
"special": true
|
| 228 |
+
},
|
| 229 |
+
"151671": {
|
| 230 |
+
"content": "<tts_pad>",
|
| 231 |
+
"lstrip": false,
|
| 232 |
+
"normalized": false,
|
| 233 |
+
"rstrip": false,
|
| 234 |
+
"single_word": false,
|
| 235 |
+
"special": true
|
| 236 |
+
},
|
| 237 |
+
"151672": {
|
| 238 |
+
"content": "<tts_text_bos>",
|
| 239 |
+
"lstrip": false,
|
| 240 |
+
"normalized": false,
|
| 241 |
+
"rstrip": false,
|
| 242 |
+
"single_word": false,
|
| 243 |
+
"special": true
|
| 244 |
+
},
|
| 245 |
+
"151673": {
|
| 246 |
+
"content": "<tts_text_eod>",
|
| 247 |
+
"lstrip": false,
|
| 248 |
+
"normalized": false,
|
| 249 |
+
"rstrip": false,
|
| 250 |
+
"single_word": false,
|
| 251 |
+
"special": true
|
| 252 |
+
},
|
| 253 |
+
"151674": {
|
| 254 |
+
"content": "<tts_text_bos_single>",
|
| 255 |
+
"lstrip": false,
|
| 256 |
+
"normalized": false,
|
| 257 |
+
"rstrip": false,
|
| 258 |
+
"single_word": false,
|
| 259 |
+
"special": true
|
| 260 |
+
},
|
| 261 |
+
"151675": {
|
| 262 |
+
"content": "<|audio_pad|>",
|
| 263 |
+
"lstrip": false,
|
| 264 |
+
"normalized": false,
|
| 265 |
+
"rstrip": false,
|
| 266 |
+
"single_word": false,
|
| 267 |
+
"special": true
|
| 268 |
+
}
|
| 269 |
+
},
|
| 270 |
+
"additional_special_tokens": [
|
| 271 |
+
"<|im_start|>",
|
| 272 |
+
"<|im_end|>",
|
| 273 |
+
"<|object_ref_start|>",
|
| 274 |
+
"<|object_ref_end|>",
|
| 275 |
+
"<|box_start|>",
|
| 276 |
+
"<|box_end|>",
|
| 277 |
+
"<|quad_start|>",
|
| 278 |
+
"<|quad_end|>",
|
| 279 |
+
"<|vision_start|>",
|
| 280 |
+
"<|vision_end|>",
|
| 281 |
+
"<|vision_pad|>",
|
| 282 |
+
"<|image_pad|>",
|
| 283 |
+
"<|video_pad|>",
|
| 284 |
+
"<|audio_start|>",
|
| 285 |
+
"<|audio_end|>",
|
| 286 |
+
"<tts_pad>",
|
| 287 |
+
"<tts_text_bos>",
|
| 288 |
+
"<tts_text_bos_single>",
|
| 289 |
+
"<|audio_pad|>"
|
| 290 |
+
],
|
| 291 |
+
"extra_special_tokens": {
|
| 292 |
+
"image_token": "<|image_pad|>",
|
| 293 |
+
"audio_token": "<|audio_pad|>",
|
| 294 |
+
"video_token": "<|video_pad|>",
|
| 295 |
+
"vision_bos_token": "<|vision_start|>",
|
| 296 |
+
"vision_eos_token": "<|vision_end|>",
|
| 297 |
+
"audio_bos_token": "<|audio_start|>",
|
| 298 |
+
"audio_eos_token": "<|audio_end|>"
|
| 299 |
+
},
|
| 300 |
+
"bos_token": null,
|
| 301 |
+
"clean_up_tokenization_spaces": false,
|
| 302 |
+
"eos_token": "<|im_end|>",
|
| 303 |
+
"errors": "replace",
|
| 304 |
+
"model_max_length": 131072,
|
| 305 |
+
"pad_token": "<|endoftext|>",
|
| 306 |
+
"split_special_tokens": false,
|
| 307 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 308 |
+
"unk_token": null,
|
| 309 |
+
"image_token": "<|image_pad|>",
|
| 310 |
+
"audio_token": "<|audio_pad|>",
|
| 311 |
+
"video_token": "<|video_pad|>",
|
| 312 |
+
"vision_bos_token": "<|vision_start|>",
|
| 313 |
+
"vision_eos_token": "<|vision_end|>",
|
| 314 |
+
"audio_bos_token": "<|audio_start|>",
|
| 315 |
+
"audio_eos_token": "<|audio_end|>"
|
| 316 |
+
}
|
qwen3-tts-12hz-1.7b-voicedesign/tokenizer/vocab.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ca10d7e9fb3ed18575dd1e277a2579c16d108e32f27439684afa0e10b1440910
|
| 3 |
+
size 2776833
|
qwen3-tts-12hz-1.7b-voicedesign/weights/model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:c0240a5d663f431683c63606770665321ef21594dbd7559fd4bd2154b26091ea
|
| 3 |
+
size 3833402584
|
qwen3-tts-12hz-1.7b-voicedesign/weights/speech_tokenizer.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:ee778743f0734e5e35f751f771ffe42a8943ca3ce4c9cc2e39b58719989c20c6
|
| 3 |
+
size 682293164
|