Commit ·
5f986ca
0
Parent(s):
DirectS2ST checkpoints
Browse files- .gitattributes +35 -0
- README.md +55 -0
- en-de/model_config.json +221 -0
- en-de/pytorch_model.bin +3 -0
- en-de/text_tokenizer/added_tokens.json +6 -0
- en-de/text_tokenizer/special_tokens_map.json +38 -0
- en-de/text_tokenizer/tokenizer.json +0 -0
- en-de/text_tokenizer/tokenizer.model +3 -0
- en-de/text_tokenizer/tokenizer_config.json +107 -0
- en-es/model_config.json +221 -0
- en-es/pytorch_model.bin +3 -0
- en-es/text_tokenizer/added_tokens.json +6 -0
- en-es/text_tokenizer/special_tokens_map.json +38 -0
- en-es/text_tokenizer/tokenizer.json +0 -0
- en-es/text_tokenizer/tokenizer.model +3 -0
- en-es/text_tokenizer/tokenizer_config.json +107 -0
.gitattributes
ADDED
|
@@ -0,0 +1,35 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
*.7z filter=lfs diff=lfs merge=lfs -text
|
| 2 |
+
*.arrow filter=lfs diff=lfs merge=lfs -text
|
| 3 |
+
*.bin filter=lfs diff=lfs merge=lfs -text
|
| 4 |
+
*.bz2 filter=lfs diff=lfs merge=lfs -text
|
| 5 |
+
*.ckpt filter=lfs diff=lfs merge=lfs -text
|
| 6 |
+
*.ftz filter=lfs diff=lfs merge=lfs -text
|
| 7 |
+
*.gz filter=lfs diff=lfs merge=lfs -text
|
| 8 |
+
*.h5 filter=lfs diff=lfs merge=lfs -text
|
| 9 |
+
*.joblib filter=lfs diff=lfs merge=lfs -text
|
| 10 |
+
*.lfs.* filter=lfs diff=lfs merge=lfs -text
|
| 11 |
+
*.mlmodel filter=lfs diff=lfs merge=lfs -text
|
| 12 |
+
*.model filter=lfs diff=lfs merge=lfs -text
|
| 13 |
+
*.msgpack filter=lfs diff=lfs merge=lfs -text
|
| 14 |
+
*.npy filter=lfs diff=lfs merge=lfs -text
|
| 15 |
+
*.npz filter=lfs diff=lfs merge=lfs -text
|
| 16 |
+
*.onnx filter=lfs diff=lfs merge=lfs -text
|
| 17 |
+
*.ot filter=lfs diff=lfs merge=lfs -text
|
| 18 |
+
*.parquet filter=lfs diff=lfs merge=lfs -text
|
| 19 |
+
*.pb filter=lfs diff=lfs merge=lfs -text
|
| 20 |
+
*.pickle filter=lfs diff=lfs merge=lfs -text
|
| 21 |
+
*.pkl filter=lfs diff=lfs merge=lfs -text
|
| 22 |
+
*.pt filter=lfs diff=lfs merge=lfs -text
|
| 23 |
+
*.pth filter=lfs diff=lfs merge=lfs -text
|
| 24 |
+
*.rar filter=lfs diff=lfs merge=lfs -text
|
| 25 |
+
*.safetensors filter=lfs diff=lfs merge=lfs -text
|
| 26 |
+
saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
| 27 |
+
*.tar.* filter=lfs diff=lfs merge=lfs -text
|
| 28 |
+
*.tar filter=lfs diff=lfs merge=lfs -text
|
| 29 |
+
*.tflite filter=lfs diff=lfs merge=lfs -text
|
| 30 |
+
*.tgz filter=lfs diff=lfs merge=lfs -text
|
| 31 |
+
*.wasm filter=lfs diff=lfs merge=lfs -text
|
| 32 |
+
*.xz filter=lfs diff=lfs merge=lfs -text
|
| 33 |
+
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
+
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
+
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,55 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
tags:
|
| 3 |
+
- speech-to-speech-translation
|
| 4 |
+
- expressive-speech-translation
|
| 5 |
+
- mimi
|
| 6 |
+
language:
|
| 7 |
+
- en
|
| 8 |
+
- de
|
| 9 |
+
- es
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# DirectS2ST
|
| 13 |
+
|
| 14 |
+
Checkpoints of DirectS2ST, the end-to-end expressive speech-to-speech translation
|
| 15 |
+
model from *Holistic Parallel Supervision for Expressive Speech-to-Speech
|
| 16 |
+
Translation*, trained on VC-Dub (professional dubbing + multilingual voice
|
| 17 |
+
conversion).
|
| 18 |
+
|
| 19 |
+
| Folder | Direction |
|
| 20 |
+
| --- | --- |
|
| 21 |
+
| `en-de/` | English → German |
|
| 22 |
+
| `en-es/` | English → Spanish |
|
| 23 |
+
|
| 24 |
+
Each folder holds `pytorch_model.bin`, `model_config.json` and the
|
| 25 |
+
SentencePiece-8k text tokenizer (`text_tokenizer/`).
|
| 26 |
+
|
| 27 |
+
Given English source speech, the model predicts all 16 Mimi codec streams of
|
| 28 |
+
the target speech directly: a frozen w2v-BERT 2.0 encoder, an AR temporal
|
| 29 |
+
decoder (source text, target text and codec stream c0 with one shared head), and
|
| 30 |
+
an NAR depth decoder (c1-c15) conditioned on a CAMPPlus speaker prefix and a
|
| 31 |
+
source codec prompt.
|
| 32 |
+
|
| 33 |
+
## Usage
|
| 34 |
+
|
| 35 |
+
Code, setup and training: <https://github.com/hzted/truly-e2e-expressive-s2st-demo/tree/main/DirectS2ST>
|
| 36 |
+
|
| 37 |
+
```
|
| 38 |
+
hf download TedZhangHao/DirectS2ST --include "en-es/*" --local-dir checkpoints
|
| 39 |
+
python inference/infer.py \
|
| 40 |
+
--checkpoint-dir checkpoints/en-es \
|
| 41 |
+
--campplus-model-root /path/to/seed-vc \
|
| 42 |
+
--campplus-checkpoint-path /path/to/campplus_cn_en_common.pt \
|
| 43 |
+
--source-wav input_en.wav --output-wav output_es.wav
|
| 44 |
+
```
|
| 45 |
+
|
| 46 |
+
Demo: <https://hzted.github.io/truly-e2e-expressive-s2st-demo/>
|
| 47 |
+
|
| 48 |
+
## Third-party components
|
| 49 |
+
|
| 50 |
+
The checkpoints contain the frozen weights of
|
| 51 |
+
[w2v-BERT 2.0](https://huggingface.co/facebook/w2v-bert-2.0) and
|
| 52 |
+
[Mimi](https://huggingface.co/kyutai/mimi); their licenses apply to those
|
| 53 |
+
parts. The CAMPPlus speaker encoder
|
| 54 |
+
([iic/speech_campplus_sv_zh_en_16k-common_advanced](https://modelscope.cn/models/iic/speech_campplus_sv_zh_en_16k-common_advanced))
|
| 55 |
+
is not included and is loaded at run time.
|
en-de/model_config.json
ADDED
|
@@ -0,0 +1,221 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"c0_auxiliary_loss_weight": 0.0,
|
| 3 |
+
"c0_gold_anchor_mass": 0.75,
|
| 4 |
+
"c0_gold_anchored_distill_temperature": 1.0,
|
| 5 |
+
"c0_gold_anchored_distill_weight": 0.0,
|
| 6 |
+
"c0_teacher_distill_temperature": 1.0,
|
| 7 |
+
"c0_teacher_distill_weight": 0.0,
|
| 8 |
+
"c0_teacher_hidden_dim": 1024,
|
| 9 |
+
"c0_teacher_hidden_distill_weight": 0.0,
|
| 10 |
+
"c0_teacher_hidden_loss_type": "cosine",
|
| 11 |
+
"c0_teacher_hidden_use_projection": true,
|
| 12 |
+
"campplus_depth_conditioning_mode": "prefix_token",
|
| 13 |
+
"campplus_depth_identity_loss_weight": 0.1,
|
| 14 |
+
"campplus_depth_start_level": 1,
|
| 15 |
+
"campplus_embedding_size": 192,
|
| 16 |
+
"campplus_film_scale": 1.0,
|
| 17 |
+
"campplus_identity_level_aggregation": "conditioned_mean",
|
| 18 |
+
"campplus_identity_supervision": "target",
|
| 19 |
+
"codebook_loss_weights": null,
|
| 20 |
+
"codebook_model_order": null,
|
| 21 |
+
"codebook_size": 2048,
|
| 22 |
+
"codec_mass_loss_weight": 0.0,
|
| 23 |
+
"depth_auto_adjust_heads": true,
|
| 24 |
+
"depth_chain_scheduled_sampling_mode": "argmax",
|
| 25 |
+
"depth_chain_scheduled_sampling_prob": 0.0,
|
| 26 |
+
"depth_chain_scheduled_sampling_start_step": 0,
|
| 27 |
+
"depth_chain_scheduled_sampling_temperature": 1.0,
|
| 28 |
+
"depth_chain_scheduled_sampling_topk": 0,
|
| 29 |
+
"depth_chain_scheduled_sampling_warmup_steps": 0,
|
| 30 |
+
"depth_dropout": 0.1,
|
| 31 |
+
"depth_ffn_dim": 4096,
|
| 32 |
+
"depth_hidden_size": 1024,
|
| 33 |
+
"depth_loss_weight": 1.0,
|
| 34 |
+
"depth_num_heads": 16,
|
| 35 |
+
"depth_num_layers": 8,
|
| 36 |
+
"depth_objective_start_level": 1,
|
| 37 |
+
"depth_scheduled_sampling_mode": "argmax",
|
| 38 |
+
"depth_scheduled_sampling_prob": 0.0,
|
| 39 |
+
"depth_scheduled_sampling_start_step": 0,
|
| 40 |
+
"depth_scheduled_sampling_temperature": 1.0,
|
| 41 |
+
"depth_scheduled_sampling_topk": 0,
|
| 42 |
+
"depth_scheduled_sampling_warmup_steps": 0,
|
| 43 |
+
"depth_source_conditioning_detach": true,
|
| 44 |
+
"depth_source_conditioning_dropout": 0.1,
|
| 45 |
+
"depth_source_conditioning_num_heads": 16,
|
| 46 |
+
"depth_source_conditioning_residual_scale": 0.2,
|
| 47 |
+
"detach_temporal_for_depth_loss": true,
|
| 48 |
+
"direct_text_to_c0_input": "target",
|
| 49 |
+
"direct_text_to_c0_kd_temperature": 1.0,
|
| 50 |
+
"direct_text_to_c0_kd_weight": 0.0,
|
| 51 |
+
"direct_text_to_c0_loss_weight": 0.0,
|
| 52 |
+
"direct_text_to_c0_zero_speech_memory": false,
|
| 53 |
+
"enable_c0_auxiliary_loss": false,
|
| 54 |
+
"enable_c0_gold_anchored_distill": false,
|
| 55 |
+
"enable_c0_teacher_distill": false,
|
| 56 |
+
"enable_c0_teacher_hidden_distill": false,
|
| 57 |
+
"enable_campplus_depth_conditioning": true,
|
| 58 |
+
"enable_codec_mass_loss": false,
|
| 59 |
+
"enable_depth_source_conditioning": false,
|
| 60 |
+
"enable_direct_text_to_c0_auxiliary": false,
|
| 61 |
+
"enable_mimi_nar_depth_prompt": true,
|
| 62 |
+
"enable_mimi_source_acoustic_conditioning": false,
|
| 63 |
+
"enable_mimi_source_acoustic_temporal_conditioning": false,
|
| 64 |
+
"enable_mimi_source_speaker_prompt": true,
|
| 65 |
+
"enable_mimi_source_style_token_bank": false,
|
| 66 |
+
"enable_mixed_source_target_auxiliary": false,
|
| 67 |
+
"enable_predicted_source_target_auxiliary": false,
|
| 68 |
+
"enable_quality_cot": true,
|
| 69 |
+
"enable_semantic_prefix_ar": false,
|
| 70 |
+
"enable_source_unit_auxiliary": false,
|
| 71 |
+
"enable_source_unit_transformer_auxiliary": false,
|
| 72 |
+
"enable_target_only_speech_auxiliary": false,
|
| 73 |
+
"enable_text_codec_ar": true,
|
| 74 |
+
"enable_text_codec_kd_loss": false,
|
| 75 |
+
"enable_text_codec_text_path_loss": false,
|
| 76 |
+
"enable_text_prefix_ar": true,
|
| 77 |
+
"enable_transvip_style_loss_aggregation": false,
|
| 78 |
+
"enable_uniss_content_controls": true,
|
| 79 |
+
"enable_unity_source_memory_fusion": false,
|
| 80 |
+
"input_num_quantizers": 16,
|
| 81 |
+
"level0_depth_objective": true,
|
| 82 |
+
"level0_eos_loss_weight": 1.0,
|
| 83 |
+
"level0_label_smoothing": 0.0,
|
| 84 |
+
"level0_label_smoothing_train_only": false,
|
| 85 |
+
"level0_loss_weight": 1.0,
|
| 86 |
+
"level0_tail_loss_last_n": 0,
|
| 87 |
+
"level0_tail_loss_weight": 1.0,
|
| 88 |
+
"load_text_path_encoder": false,
|
| 89 |
+
"mimi_nar_depth_attention_mode": "full",
|
| 90 |
+
"mimi_nar_depth_max_positions": 4096,
|
| 91 |
+
"mimi_nar_depth_teacher_force_level_batch_size": 1,
|
| 92 |
+
"mimi_source_acoustic_depth_start_level": 99,
|
| 93 |
+
"mimi_source_acoustic_detach": true,
|
| 94 |
+
"mimi_source_acoustic_input_sample_rate": 16000,
|
| 95 |
+
"mimi_source_acoustic_model_name": "kyutai/mimi",
|
| 96 |
+
"mimi_source_acoustic_pooling": "mean_std",
|
| 97 |
+
"mimi_source_acoustic_residual_scale": 0.0,
|
| 98 |
+
"mimi_source_acoustic_temporal_detach": true,
|
| 99 |
+
"mimi_source_acoustic_temporal_residual_scale": 0.0,
|
| 100 |
+
"mimi_source_speaker_prompt_detach": true,
|
| 101 |
+
"mimi_source_style_token_c1_gate_init": 0.03,
|
| 102 |
+
"mimi_source_style_token_count": 8,
|
| 103 |
+
"mimi_source_style_token_depth_start_level": 1,
|
| 104 |
+
"mimi_source_style_token_dropout": 0.1,
|
| 105 |
+
"mimi_source_style_token_gate_init": 0.1,
|
| 106 |
+
"mimi_source_style_token_num_heads": 8,
|
| 107 |
+
"mimi_source_style_token_residual_scale": 0.7,
|
| 108 |
+
"mixed_source_auxiliary_seed": 59,
|
| 109 |
+
"mixed_source_prediction_prob": 0.0,
|
| 110 |
+
"mixed_source_start_step": 0,
|
| 111 |
+
"mixed_source_target_loss_weight": 0.0,
|
| 112 |
+
"mixed_source_warmup_steps": 0,
|
| 113 |
+
"num_codebook_levels": 16,
|
| 114 |
+
"num_temporal_codebook_levels": 1,
|
| 115 |
+
"predicted_source_target_batch_size": 2,
|
| 116 |
+
"predicted_source_target_loss_weight": 0.0,
|
| 117 |
+
"predicted_source_target_max_source_tokens": 64,
|
| 118 |
+
"predicted_source_target_start_step": 0,
|
| 119 |
+
"predicted_source_target_warmup_steps": 0,
|
| 120 |
+
"quality_cot_code0_loss_weight": 0.0,
|
| 121 |
+
"quality_cot_code1_loss_weight": 0.0,
|
| 122 |
+
"quality_cot_depth_loss_weight": 3.0,
|
| 123 |
+
"quality_cot_prompt_all_seps": false,
|
| 124 |
+
"quality_cot_source_text_loss_weight": 0.0,
|
| 125 |
+
"quality_cot_source_text_max_tokens": 128,
|
| 126 |
+
"quality_cot_source_text_prefix_extra_token_ids_after_bos": [
|
| 127 |
+
4
|
| 128 |
+
],
|
| 129 |
+
"quality_cot_target_text_loss_weight": 0.0,
|
| 130 |
+
"quality_cot_target_text_max_tokens": 128,
|
| 131 |
+
"skip_mimi_source_acoustic_encoder_when_precomputed": false,
|
| 132 |
+
"source_adapter_dropout": 0.1,
|
| 133 |
+
"source_adapter_ffn_dim": 4096,
|
| 134 |
+
"source_adapter_num_heads": 16,
|
| 135 |
+
"source_adapter_num_layers": 0,
|
| 136 |
+
"source_adapter_residual": true,
|
| 137 |
+
"source_adapter_residual_scale": 0.1,
|
| 138 |
+
"source_speaker_prompt_backend": "mimi_mean_std",
|
| 139 |
+
"source_speaker_prompt_dropout": 0.1,
|
| 140 |
+
"source_speaker_prompt_ffn_dim": 4096,
|
| 141 |
+
"source_speaker_prompt_num_heads": 16,
|
| 142 |
+
"source_speaker_prompt_num_layers": 2,
|
| 143 |
+
"source_speaker_prompt_speech_layer": 12,
|
| 144 |
+
"source_unit_blank_id": 1000,
|
| 145 |
+
"source_unit_loss_weight": 0.2,
|
| 146 |
+
"source_unit_transformer_dropout": 0.1,
|
| 147 |
+
"source_unit_transformer_ffn_dim": 4096,
|
| 148 |
+
"source_unit_transformer_loss_weight": 0.5,
|
| 149 |
+
"source_unit_transformer_max_positions": 513,
|
| 150 |
+
"source_unit_transformer_num_heads": 8,
|
| 151 |
+
"source_unit_transformer_num_layers": 2,
|
| 152 |
+
"source_unit_vocab_size": 1000,
|
| 153 |
+
"speech_encoder_name": "facebook/w2v-bert-2.0",
|
| 154 |
+
"speech_encoder_output_layer": 12,
|
| 155 |
+
"speech_encoder_type": "w2vbert2",
|
| 156 |
+
"target_only_speech_code0_loss_weight": 0.0,
|
| 157 |
+
"target_only_speech_text_loss_weight": 0.0,
|
| 158 |
+
"temporal_activation": "gelu",
|
| 159 |
+
"temporal_auto_adjust_heads": true,
|
| 160 |
+
"temporal_context_levels": 1,
|
| 161 |
+
"temporal_decoder_backend": "direct",
|
| 162 |
+
"temporal_dropout": 0.1,
|
| 163 |
+
"temporal_ffn_dim": 4096,
|
| 164 |
+
"temporal_hidden_size": 1024,
|
| 165 |
+
"temporal_max_positions": 2048,
|
| 166 |
+
"temporal_num_heads": 16,
|
| 167 |
+
"temporal_num_layers": 12,
|
| 168 |
+
"temporal_scheduled_sampling_decay_start_step": -1,
|
| 169 |
+
"temporal_scheduled_sampling_decay_steps": 0,
|
| 170 |
+
"temporal_scheduled_sampling_final_prob": 0.0,
|
| 171 |
+
"temporal_scheduled_sampling_mode": "argmax",
|
| 172 |
+
"temporal_scheduled_sampling_preserve_last_n": 0,
|
| 173 |
+
"temporal_scheduled_sampling_prob": 0.0,
|
| 174 |
+
"temporal_scheduled_sampling_start_step": 0,
|
| 175 |
+
"temporal_scheduled_sampling_temperature": 1.0,
|
| 176 |
+
"temporal_scheduled_sampling_topk": 0,
|
| 177 |
+
"temporal_scheduled_sampling_warmup_steps": 0,
|
| 178 |
+
"text_bos_token_id": 8000,
|
| 179 |
+
"text_codec_kd_loss_weight": 0.0,
|
| 180 |
+
"text_codec_t2c_loss_weight": 0.0,
|
| 181 |
+
"text_codec_t2t_loss_weight": 0.0,
|
| 182 |
+
"text_codec_text_path_input": "source",
|
| 183 |
+
"text_label_smoothing": 0.0,
|
| 184 |
+
"text_loss_weight": 1.0,
|
| 185 |
+
"text_max_positions": 255,
|
| 186 |
+
"text_pad_token_id": 0,
|
| 187 |
+
"text_path_source_token_dropout": 0.0,
|
| 188 |
+
"text_prefix_beam_size": 5,
|
| 189 |
+
"text_prefix_extra_token_ids_after_bos": [
|
| 190 |
+
5
|
| 191 |
+
],
|
| 192 |
+
"text_prefix_length_penalty": 1.0,
|
| 193 |
+
"text_prefix_min_tokens": 4,
|
| 194 |
+
"text_prefix_no_repeat_ngram_size": 3,
|
| 195 |
+
"text_prefix_sep_penalty": 0.0,
|
| 196 |
+
"text_sep_token_id": 8001,
|
| 197 |
+
"text_tokenizer_name": "./text_tokenizer",
|
| 198 |
+
"text_vocab_size": 8004,
|
| 199 |
+
"tie_text_embeddings": true,
|
| 200 |
+
"train_only_depth_decoder": true,
|
| 201 |
+
"transvip_load_source_encoder_from_checkpoint": false,
|
| 202 |
+
"transvip_load_text_decoder_from_checkpoint": false,
|
| 203 |
+
"transvip_model_name": "",
|
| 204 |
+
"transvip_num_new_tokens": null,
|
| 205 |
+
"transvip_style_depth_loss_weight": 0.0,
|
| 206 |
+
"transvip_style_kd_loss_weight": 0.0,
|
| 207 |
+
"transvip_style_speech_loss_weight": 0.0,
|
| 208 |
+
"transvip_style_text_loss_weight": 0.0,
|
| 209 |
+
"transvip_text_decoder_text_vocab_size": null,
|
| 210 |
+
"transvip_use_length_control": false,
|
| 211 |
+
"uniss_end_content_token_id": 8003,
|
| 212 |
+
"uniss_start_content_token_id": 8002,
|
| 213 |
+
"unity_source_memory_detach": true,
|
| 214 |
+
"unity_source_memory_fusion_mode": "concat",
|
| 215 |
+
"unity_source_memory_layer": 8,
|
| 216 |
+
"unity_source_memory_scale": 0.5,
|
| 217 |
+
"unity_source_memory_source": "inner",
|
| 218 |
+
"use_precomputed_source_acoustic_embeddings": true,
|
| 219 |
+
"use_unified_code0_loss_in_objective": true,
|
| 220 |
+
"vocab_size": 2050
|
| 221 |
+
}
|
en-de/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:e5f996500058680f7797a3a2b16b3fbc6f7ca0d27ed25aead5dc29e03a966cb7
|
| 3 |
+
size 4995031222
|
en-de/text_tokenizer/added_tokens.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<|end_content|>": 8003,
|
| 3 |
+
"<|start_content|>": 8002,
|
| 4 |
+
"<|text_bos|>": 8000,
|
| 5 |
+
"<|text_sep|>": 8001
|
| 6 |
+
}
|
en-de/text_tokenizer/special_tokens_map.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<en>",
|
| 4 |
+
"<de>",
|
| 5 |
+
"<|text_bos|>",
|
| 6 |
+
"<|text_sep|>",
|
| 7 |
+
"<|start_content|>",
|
| 8 |
+
"<|end_content|>"
|
| 9 |
+
],
|
| 10 |
+
"bos_token": {
|
| 11 |
+
"content": "<s>",
|
| 12 |
+
"lstrip": false,
|
| 13 |
+
"normalized": false,
|
| 14 |
+
"rstrip": false,
|
| 15 |
+
"single_word": false
|
| 16 |
+
},
|
| 17 |
+
"eos_token": {
|
| 18 |
+
"content": "</s>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
},
|
| 24 |
+
"pad_token": {
|
| 25 |
+
"content": "<pad>",
|
| 26 |
+
"lstrip": false,
|
| 27 |
+
"normalized": false,
|
| 28 |
+
"rstrip": false,
|
| 29 |
+
"single_word": false
|
| 30 |
+
},
|
| 31 |
+
"unk_token": {
|
| 32 |
+
"content": "<unk>",
|
| 33 |
+
"lstrip": false,
|
| 34 |
+
"normalized": false,
|
| 35 |
+
"rstrip": false,
|
| 36 |
+
"single_word": false
|
| 37 |
+
}
|
| 38 |
+
}
|
en-de/text_tokenizer/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
en-de/text_tokenizer/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:18627b96648618340a1ab873ac814639fe397fb87eac7ce0cdfa91a597769a44
|
| 3 |
+
size 364791
|
en-de/text_tokenizer/tokenizer_config.json
ADDED
|
@@ -0,0 +1,107 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": true,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"add_prefix_space": true,
|
| 5 |
+
"added_tokens_decoder": {
|
| 6 |
+
"0": {
|
| 7 |
+
"content": "<pad>",
|
| 8 |
+
"lstrip": false,
|
| 9 |
+
"normalized": false,
|
| 10 |
+
"rstrip": false,
|
| 11 |
+
"single_word": false,
|
| 12 |
+
"special": true
|
| 13 |
+
},
|
| 14 |
+
"1": {
|
| 15 |
+
"content": "<unk>",
|
| 16 |
+
"lstrip": false,
|
| 17 |
+
"normalized": false,
|
| 18 |
+
"rstrip": false,
|
| 19 |
+
"single_word": false,
|
| 20 |
+
"special": true
|
| 21 |
+
},
|
| 22 |
+
"2": {
|
| 23 |
+
"content": "<s>",
|
| 24 |
+
"lstrip": false,
|
| 25 |
+
"normalized": false,
|
| 26 |
+
"rstrip": false,
|
| 27 |
+
"single_word": false,
|
| 28 |
+
"special": true
|
| 29 |
+
},
|
| 30 |
+
"3": {
|
| 31 |
+
"content": "</s>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false,
|
| 36 |
+
"special": true
|
| 37 |
+
},
|
| 38 |
+
"4": {
|
| 39 |
+
"content": "<en>",
|
| 40 |
+
"lstrip": false,
|
| 41 |
+
"normalized": false,
|
| 42 |
+
"rstrip": false,
|
| 43 |
+
"single_word": false,
|
| 44 |
+
"special": true
|
| 45 |
+
},
|
| 46 |
+
"5": {
|
| 47 |
+
"content": "<de>",
|
| 48 |
+
"lstrip": false,
|
| 49 |
+
"normalized": false,
|
| 50 |
+
"rstrip": false,
|
| 51 |
+
"single_word": false,
|
| 52 |
+
"special": true
|
| 53 |
+
},
|
| 54 |
+
"8000": {
|
| 55 |
+
"content": "<|text_bos|>",
|
| 56 |
+
"lstrip": false,
|
| 57 |
+
"normalized": false,
|
| 58 |
+
"rstrip": false,
|
| 59 |
+
"single_word": false,
|
| 60 |
+
"special": true
|
| 61 |
+
},
|
| 62 |
+
"8001": {
|
| 63 |
+
"content": "<|text_sep|>",
|
| 64 |
+
"lstrip": false,
|
| 65 |
+
"normalized": false,
|
| 66 |
+
"rstrip": false,
|
| 67 |
+
"single_word": false,
|
| 68 |
+
"special": true
|
| 69 |
+
},
|
| 70 |
+
"8002": {
|
| 71 |
+
"content": "<|start_content|>",
|
| 72 |
+
"lstrip": false,
|
| 73 |
+
"normalized": false,
|
| 74 |
+
"rstrip": false,
|
| 75 |
+
"single_word": false,
|
| 76 |
+
"special": true
|
| 77 |
+
},
|
| 78 |
+
"8003": {
|
| 79 |
+
"content": "<|end_content|>",
|
| 80 |
+
"lstrip": false,
|
| 81 |
+
"normalized": false,
|
| 82 |
+
"rstrip": false,
|
| 83 |
+
"single_word": false,
|
| 84 |
+
"special": true
|
| 85 |
+
}
|
| 86 |
+
},
|
| 87 |
+
"additional_special_tokens": [
|
| 88 |
+
"<en>",
|
| 89 |
+
"<de>",
|
| 90 |
+
"<|text_bos|>",
|
| 91 |
+
"<|text_sep|>",
|
| 92 |
+
"<|start_content|>",
|
| 93 |
+
"<|end_content|>"
|
| 94 |
+
],
|
| 95 |
+
"bos_token": "<s>",
|
| 96 |
+
"clean_up_tokenization_spaces": false,
|
| 97 |
+
"eos_token": "</s>",
|
| 98 |
+
"extra_special_tokens": {},
|
| 99 |
+
"legacy": true,
|
| 100 |
+
"model_max_length": 2048,
|
| 101 |
+
"pad_token": "<pad>",
|
| 102 |
+
"sp_model_kwargs": {},
|
| 103 |
+
"spaces_between_special_tokens": false,
|
| 104 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 105 |
+
"unk_token": "<unk>",
|
| 106 |
+
"use_default_system_prompt": false
|
| 107 |
+
}
|
en-es/model_config.json
ADDED
|
@@ -0,0 +1,221 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"c0_auxiliary_loss_weight": 0.0,
|
| 3 |
+
"c0_gold_anchor_mass": 0.75,
|
| 4 |
+
"c0_gold_anchored_distill_temperature": 1.0,
|
| 5 |
+
"c0_gold_anchored_distill_weight": 0.0,
|
| 6 |
+
"c0_teacher_distill_temperature": 1.0,
|
| 7 |
+
"c0_teacher_distill_weight": 0.0,
|
| 8 |
+
"c0_teacher_hidden_dim": 1024,
|
| 9 |
+
"c0_teacher_hidden_distill_weight": 0.0,
|
| 10 |
+
"c0_teacher_hidden_loss_type": "cosine",
|
| 11 |
+
"c0_teacher_hidden_use_projection": true,
|
| 12 |
+
"campplus_depth_conditioning_mode": "prefix_token",
|
| 13 |
+
"campplus_depth_identity_loss_weight": 0.1,
|
| 14 |
+
"campplus_depth_start_level": 1,
|
| 15 |
+
"campplus_embedding_size": 192,
|
| 16 |
+
"campplus_film_scale": 1.0,
|
| 17 |
+
"campplus_identity_level_aggregation": "conditioned_mean",
|
| 18 |
+
"campplus_identity_supervision": "target",
|
| 19 |
+
"codebook_loss_weights": null,
|
| 20 |
+
"codebook_model_order": null,
|
| 21 |
+
"codebook_size": 2048,
|
| 22 |
+
"codec_mass_loss_weight": 0.0,
|
| 23 |
+
"depth_auto_adjust_heads": true,
|
| 24 |
+
"depth_chain_scheduled_sampling_mode": "argmax",
|
| 25 |
+
"depth_chain_scheduled_sampling_prob": 0.0,
|
| 26 |
+
"depth_chain_scheduled_sampling_start_step": 0,
|
| 27 |
+
"depth_chain_scheduled_sampling_temperature": 1.0,
|
| 28 |
+
"depth_chain_scheduled_sampling_topk": 0,
|
| 29 |
+
"depth_chain_scheduled_sampling_warmup_steps": 0,
|
| 30 |
+
"depth_dropout": 0.1,
|
| 31 |
+
"depth_ffn_dim": 4096,
|
| 32 |
+
"depth_hidden_size": 1024,
|
| 33 |
+
"depth_loss_weight": 1.0,
|
| 34 |
+
"depth_num_heads": 16,
|
| 35 |
+
"depth_num_layers": 8,
|
| 36 |
+
"depth_objective_start_level": 1,
|
| 37 |
+
"depth_scheduled_sampling_mode": "argmax",
|
| 38 |
+
"depth_scheduled_sampling_prob": 0.0,
|
| 39 |
+
"depth_scheduled_sampling_start_step": 0,
|
| 40 |
+
"depth_scheduled_sampling_temperature": 1.0,
|
| 41 |
+
"depth_scheduled_sampling_topk": 0,
|
| 42 |
+
"depth_scheduled_sampling_warmup_steps": 0,
|
| 43 |
+
"depth_source_conditioning_detach": true,
|
| 44 |
+
"depth_source_conditioning_dropout": 0.1,
|
| 45 |
+
"depth_source_conditioning_num_heads": 16,
|
| 46 |
+
"depth_source_conditioning_residual_scale": 0.2,
|
| 47 |
+
"detach_temporal_for_depth_loss": true,
|
| 48 |
+
"direct_text_to_c0_input": "target",
|
| 49 |
+
"direct_text_to_c0_kd_temperature": 1.0,
|
| 50 |
+
"direct_text_to_c0_kd_weight": 0.0,
|
| 51 |
+
"direct_text_to_c0_loss_weight": 0.0,
|
| 52 |
+
"direct_text_to_c0_zero_speech_memory": false,
|
| 53 |
+
"enable_c0_auxiliary_loss": false,
|
| 54 |
+
"enable_c0_gold_anchored_distill": false,
|
| 55 |
+
"enable_c0_teacher_distill": false,
|
| 56 |
+
"enable_c0_teacher_hidden_distill": false,
|
| 57 |
+
"enable_campplus_depth_conditioning": true,
|
| 58 |
+
"enable_codec_mass_loss": false,
|
| 59 |
+
"enable_depth_source_conditioning": false,
|
| 60 |
+
"enable_direct_text_to_c0_auxiliary": false,
|
| 61 |
+
"enable_mimi_nar_depth_prompt": true,
|
| 62 |
+
"enable_mimi_source_acoustic_conditioning": false,
|
| 63 |
+
"enable_mimi_source_acoustic_temporal_conditioning": false,
|
| 64 |
+
"enable_mimi_source_speaker_prompt": true,
|
| 65 |
+
"enable_mimi_source_style_token_bank": false,
|
| 66 |
+
"enable_mixed_source_target_auxiliary": false,
|
| 67 |
+
"enable_predicted_source_target_auxiliary": false,
|
| 68 |
+
"enable_quality_cot": true,
|
| 69 |
+
"enable_semantic_prefix_ar": false,
|
| 70 |
+
"enable_source_unit_auxiliary": false,
|
| 71 |
+
"enable_source_unit_transformer_auxiliary": false,
|
| 72 |
+
"enable_target_only_speech_auxiliary": false,
|
| 73 |
+
"enable_text_codec_ar": true,
|
| 74 |
+
"enable_text_codec_kd_loss": false,
|
| 75 |
+
"enable_text_codec_text_path_loss": true,
|
| 76 |
+
"enable_text_prefix_ar": true,
|
| 77 |
+
"enable_transvip_style_loss_aggregation": false,
|
| 78 |
+
"enable_uniss_content_controls": true,
|
| 79 |
+
"enable_unity_source_memory_fusion": false,
|
| 80 |
+
"input_num_quantizers": 16,
|
| 81 |
+
"level0_depth_objective": true,
|
| 82 |
+
"level0_eos_loss_weight": 1.0,
|
| 83 |
+
"level0_label_smoothing": 0.0,
|
| 84 |
+
"level0_label_smoothing_train_only": false,
|
| 85 |
+
"level0_loss_weight": 1.0,
|
| 86 |
+
"level0_tail_loss_last_n": 0,
|
| 87 |
+
"level0_tail_loss_weight": 1.0,
|
| 88 |
+
"load_text_path_encoder": false,
|
| 89 |
+
"mimi_nar_depth_attention_mode": "full",
|
| 90 |
+
"mimi_nar_depth_max_positions": 4096,
|
| 91 |
+
"mimi_nar_depth_teacher_force_level_batch_size": 1,
|
| 92 |
+
"mimi_source_acoustic_depth_start_level": 99,
|
| 93 |
+
"mimi_source_acoustic_detach": true,
|
| 94 |
+
"mimi_source_acoustic_input_sample_rate": 16000,
|
| 95 |
+
"mimi_source_acoustic_model_name": "kyutai/mimi",
|
| 96 |
+
"mimi_source_acoustic_pooling": "mean_std",
|
| 97 |
+
"mimi_source_acoustic_residual_scale": 0.0,
|
| 98 |
+
"mimi_source_acoustic_temporal_detach": true,
|
| 99 |
+
"mimi_source_acoustic_temporal_residual_scale": 0.0,
|
| 100 |
+
"mimi_source_speaker_prompt_detach": true,
|
| 101 |
+
"mimi_source_style_token_c1_gate_init": 0.03,
|
| 102 |
+
"mimi_source_style_token_count": 8,
|
| 103 |
+
"mimi_source_style_token_depth_start_level": 1,
|
| 104 |
+
"mimi_source_style_token_dropout": 0.1,
|
| 105 |
+
"mimi_source_style_token_gate_init": 0.1,
|
| 106 |
+
"mimi_source_style_token_num_heads": 8,
|
| 107 |
+
"mimi_source_style_token_residual_scale": 0.7,
|
| 108 |
+
"mixed_source_auxiliary_seed": 59,
|
| 109 |
+
"mixed_source_prediction_prob": 0.0,
|
| 110 |
+
"mixed_source_start_step": 0,
|
| 111 |
+
"mixed_source_target_loss_weight": 0.0,
|
| 112 |
+
"mixed_source_warmup_steps": 0,
|
| 113 |
+
"num_codebook_levels": 16,
|
| 114 |
+
"num_temporal_codebook_levels": 1,
|
| 115 |
+
"predicted_source_target_batch_size": 2,
|
| 116 |
+
"predicted_source_target_loss_weight": 0.0,
|
| 117 |
+
"predicted_source_target_max_source_tokens": 64,
|
| 118 |
+
"predicted_source_target_start_step": 0,
|
| 119 |
+
"predicted_source_target_warmup_steps": 0,
|
| 120 |
+
"quality_cot_code0_loss_weight": 1.0,
|
| 121 |
+
"quality_cot_code1_loss_weight": 0.0,
|
| 122 |
+
"quality_cot_depth_loss_weight": 3.0,
|
| 123 |
+
"quality_cot_prompt_all_seps": false,
|
| 124 |
+
"quality_cot_source_text_loss_weight": 0.4,
|
| 125 |
+
"quality_cot_source_text_max_tokens": 128,
|
| 126 |
+
"quality_cot_source_text_prefix_extra_token_ids_after_bos": [
|
| 127 |
+
4
|
| 128 |
+
],
|
| 129 |
+
"quality_cot_target_text_loss_weight": 0.6,
|
| 130 |
+
"quality_cot_target_text_max_tokens": 128,
|
| 131 |
+
"skip_mimi_source_acoustic_encoder_when_precomputed": false,
|
| 132 |
+
"source_adapter_dropout": 0.1,
|
| 133 |
+
"source_adapter_ffn_dim": 4096,
|
| 134 |
+
"source_adapter_num_heads": 16,
|
| 135 |
+
"source_adapter_num_layers": 0,
|
| 136 |
+
"source_adapter_residual": true,
|
| 137 |
+
"source_adapter_residual_scale": 0.1,
|
| 138 |
+
"source_speaker_prompt_backend": "mimi_mean_std",
|
| 139 |
+
"source_speaker_prompt_dropout": 0.1,
|
| 140 |
+
"source_speaker_prompt_ffn_dim": 4096,
|
| 141 |
+
"source_speaker_prompt_num_heads": 16,
|
| 142 |
+
"source_speaker_prompt_num_layers": 2,
|
| 143 |
+
"source_speaker_prompt_speech_layer": 12,
|
| 144 |
+
"source_unit_blank_id": 1000,
|
| 145 |
+
"source_unit_loss_weight": 0.2,
|
| 146 |
+
"source_unit_transformer_dropout": 0.1,
|
| 147 |
+
"source_unit_transformer_ffn_dim": 4096,
|
| 148 |
+
"source_unit_transformer_loss_weight": 0.5,
|
| 149 |
+
"source_unit_transformer_max_positions": 513,
|
| 150 |
+
"source_unit_transformer_num_heads": 8,
|
| 151 |
+
"source_unit_transformer_num_layers": 2,
|
| 152 |
+
"source_unit_vocab_size": 1000,
|
| 153 |
+
"speech_encoder_name": "facebook/w2v-bert-2.0",
|
| 154 |
+
"speech_encoder_output_layer": 12,
|
| 155 |
+
"speech_encoder_type": "w2vbert2",
|
| 156 |
+
"target_only_speech_code0_loss_weight": 0.0,
|
| 157 |
+
"target_only_speech_text_loss_weight": 0.0,
|
| 158 |
+
"temporal_activation": "gelu",
|
| 159 |
+
"temporal_auto_adjust_heads": true,
|
| 160 |
+
"temporal_context_levels": 1,
|
| 161 |
+
"temporal_decoder_backend": "direct",
|
| 162 |
+
"temporal_dropout": 0.1,
|
| 163 |
+
"temporal_ffn_dim": 4096,
|
| 164 |
+
"temporal_hidden_size": 1024,
|
| 165 |
+
"temporal_max_positions": 2048,
|
| 166 |
+
"temporal_num_heads": 16,
|
| 167 |
+
"temporal_num_layers": 12,
|
| 168 |
+
"temporal_scheduled_sampling_decay_start_step": -1,
|
| 169 |
+
"temporal_scheduled_sampling_decay_steps": 0,
|
| 170 |
+
"temporal_scheduled_sampling_final_prob": 0.0,
|
| 171 |
+
"temporal_scheduled_sampling_mode": "argmax",
|
| 172 |
+
"temporal_scheduled_sampling_preserve_last_n": 0,
|
| 173 |
+
"temporal_scheduled_sampling_prob": 0.0,
|
| 174 |
+
"temporal_scheduled_sampling_start_step": 0,
|
| 175 |
+
"temporal_scheduled_sampling_temperature": 1.0,
|
| 176 |
+
"temporal_scheduled_sampling_topk": 0,
|
| 177 |
+
"temporal_scheduled_sampling_warmup_steps": 0,
|
| 178 |
+
"text_bos_token_id": 8000,
|
| 179 |
+
"text_codec_kd_loss_weight": 0.0,
|
| 180 |
+
"text_codec_t2c_loss_weight": 0.5,
|
| 181 |
+
"text_codec_t2t_loss_weight": 0.3,
|
| 182 |
+
"text_codec_text_path_input": "source",
|
| 183 |
+
"text_label_smoothing": 0.0,
|
| 184 |
+
"text_loss_weight": 1.0,
|
| 185 |
+
"text_max_positions": 255,
|
| 186 |
+
"text_pad_token_id": 0,
|
| 187 |
+
"text_path_source_token_dropout": 0.0,
|
| 188 |
+
"text_prefix_beam_size": 5,
|
| 189 |
+
"text_prefix_extra_token_ids_after_bos": [
|
| 190 |
+
5
|
| 191 |
+
],
|
| 192 |
+
"text_prefix_length_penalty": 1.0,
|
| 193 |
+
"text_prefix_min_tokens": 4,
|
| 194 |
+
"text_prefix_no_repeat_ngram_size": 3,
|
| 195 |
+
"text_prefix_sep_penalty": 0.0,
|
| 196 |
+
"text_sep_token_id": 8001,
|
| 197 |
+
"text_tokenizer_name": "./text_tokenizer",
|
| 198 |
+
"text_vocab_size": 8004,
|
| 199 |
+
"tie_text_embeddings": true,
|
| 200 |
+
"train_only_depth_decoder": true,
|
| 201 |
+
"transvip_load_source_encoder_from_checkpoint": false,
|
| 202 |
+
"transvip_load_text_decoder_from_checkpoint": false,
|
| 203 |
+
"transvip_model_name": "",
|
| 204 |
+
"transvip_num_new_tokens": null,
|
| 205 |
+
"transvip_style_depth_loss_weight": 0.0,
|
| 206 |
+
"transvip_style_kd_loss_weight": 0.0,
|
| 207 |
+
"transvip_style_speech_loss_weight": 0.0,
|
| 208 |
+
"transvip_style_text_loss_weight": 0.0,
|
| 209 |
+
"transvip_text_decoder_text_vocab_size": null,
|
| 210 |
+
"transvip_use_length_control": false,
|
| 211 |
+
"uniss_end_content_token_id": 8003,
|
| 212 |
+
"uniss_start_content_token_id": 8002,
|
| 213 |
+
"unity_source_memory_detach": true,
|
| 214 |
+
"unity_source_memory_fusion_mode": "concat",
|
| 215 |
+
"unity_source_memory_layer": 8,
|
| 216 |
+
"unity_source_memory_scale": 0.5,
|
| 217 |
+
"unity_source_memory_source": "inner",
|
| 218 |
+
"use_precomputed_source_acoustic_embeddings": true,
|
| 219 |
+
"use_unified_code0_loss_in_objective": true,
|
| 220 |
+
"vocab_size": 2050
|
| 221 |
+
}
|
en-es/pytorch_model.bin
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:5bdb865b68a6c9cd8f442d9369bf14d5559fb7ec9e36f88f706f8f805373446c
|
| 3 |
+
size 4995031222
|
en-es/text_tokenizer/added_tokens.json
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"<|end_content|>": 8003,
|
| 3 |
+
"<|start_content|>": 8002,
|
| 4 |
+
"<|text_bos|>": 8000,
|
| 5 |
+
"<|text_sep|>": 8001
|
| 6 |
+
}
|
en-es/text_tokenizer/special_tokens_map.json
ADDED
|
@@ -0,0 +1,38 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<en>",
|
| 4 |
+
"<es>",
|
| 5 |
+
"<|text_bos|>",
|
| 6 |
+
"<|text_sep|>",
|
| 7 |
+
"<|start_content|>",
|
| 8 |
+
"<|end_content|>"
|
| 9 |
+
],
|
| 10 |
+
"bos_token": {
|
| 11 |
+
"content": "<s>",
|
| 12 |
+
"lstrip": false,
|
| 13 |
+
"normalized": false,
|
| 14 |
+
"rstrip": false,
|
| 15 |
+
"single_word": false
|
| 16 |
+
},
|
| 17 |
+
"eos_token": {
|
| 18 |
+
"content": "</s>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
},
|
| 24 |
+
"pad_token": {
|
| 25 |
+
"content": "<pad>",
|
| 26 |
+
"lstrip": false,
|
| 27 |
+
"normalized": false,
|
| 28 |
+
"rstrip": false,
|
| 29 |
+
"single_word": false
|
| 30 |
+
},
|
| 31 |
+
"unk_token": {
|
| 32 |
+
"content": "<unk>",
|
| 33 |
+
"lstrip": false,
|
| 34 |
+
"normalized": false,
|
| 35 |
+
"rstrip": false,
|
| 36 |
+
"single_word": false
|
| 37 |
+
}
|
| 38 |
+
}
|
en-es/text_tokenizer/tokenizer.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
en-es/text_tokenizer/tokenizer.model
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:0111d3c52fd33496b60a079ff0ced24af6ebf66e849c67c7d63d6c93de19f320
|
| 3 |
+
size 367186
|
en-es/text_tokenizer/tokenizer_config.json
ADDED
|
@@ -0,0 +1,107 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_bos_token": true,
|
| 3 |
+
"add_eos_token": false,
|
| 4 |
+
"add_prefix_space": true,
|
| 5 |
+
"added_tokens_decoder": {
|
| 6 |
+
"0": {
|
| 7 |
+
"content": "<pad>",
|
| 8 |
+
"lstrip": false,
|
| 9 |
+
"normalized": false,
|
| 10 |
+
"rstrip": false,
|
| 11 |
+
"single_word": false,
|
| 12 |
+
"special": true
|
| 13 |
+
},
|
| 14 |
+
"1": {
|
| 15 |
+
"content": "<unk>",
|
| 16 |
+
"lstrip": false,
|
| 17 |
+
"normalized": false,
|
| 18 |
+
"rstrip": false,
|
| 19 |
+
"single_word": false,
|
| 20 |
+
"special": true
|
| 21 |
+
},
|
| 22 |
+
"2": {
|
| 23 |
+
"content": "<s>",
|
| 24 |
+
"lstrip": false,
|
| 25 |
+
"normalized": false,
|
| 26 |
+
"rstrip": false,
|
| 27 |
+
"single_word": false,
|
| 28 |
+
"special": true
|
| 29 |
+
},
|
| 30 |
+
"3": {
|
| 31 |
+
"content": "</s>",
|
| 32 |
+
"lstrip": false,
|
| 33 |
+
"normalized": false,
|
| 34 |
+
"rstrip": false,
|
| 35 |
+
"single_word": false,
|
| 36 |
+
"special": true
|
| 37 |
+
},
|
| 38 |
+
"4": {
|
| 39 |
+
"content": "<en>",
|
| 40 |
+
"lstrip": false,
|
| 41 |
+
"normalized": false,
|
| 42 |
+
"rstrip": false,
|
| 43 |
+
"single_word": false,
|
| 44 |
+
"special": true
|
| 45 |
+
},
|
| 46 |
+
"5": {
|
| 47 |
+
"content": "<es>",
|
| 48 |
+
"lstrip": false,
|
| 49 |
+
"normalized": false,
|
| 50 |
+
"rstrip": false,
|
| 51 |
+
"single_word": false,
|
| 52 |
+
"special": true
|
| 53 |
+
},
|
| 54 |
+
"8000": {
|
| 55 |
+
"content": "<|text_bos|>",
|
| 56 |
+
"lstrip": false,
|
| 57 |
+
"normalized": false,
|
| 58 |
+
"rstrip": false,
|
| 59 |
+
"single_word": false,
|
| 60 |
+
"special": true
|
| 61 |
+
},
|
| 62 |
+
"8001": {
|
| 63 |
+
"content": "<|text_sep|>",
|
| 64 |
+
"lstrip": false,
|
| 65 |
+
"normalized": false,
|
| 66 |
+
"rstrip": false,
|
| 67 |
+
"single_word": false,
|
| 68 |
+
"special": true
|
| 69 |
+
},
|
| 70 |
+
"8002": {
|
| 71 |
+
"content": "<|start_content|>",
|
| 72 |
+
"lstrip": false,
|
| 73 |
+
"normalized": false,
|
| 74 |
+
"rstrip": false,
|
| 75 |
+
"single_word": false,
|
| 76 |
+
"special": true
|
| 77 |
+
},
|
| 78 |
+
"8003": {
|
| 79 |
+
"content": "<|end_content|>",
|
| 80 |
+
"lstrip": false,
|
| 81 |
+
"normalized": false,
|
| 82 |
+
"rstrip": false,
|
| 83 |
+
"single_word": false,
|
| 84 |
+
"special": true
|
| 85 |
+
}
|
| 86 |
+
},
|
| 87 |
+
"additional_special_tokens": [
|
| 88 |
+
"<en>",
|
| 89 |
+
"<es>",
|
| 90 |
+
"<|text_bos|>",
|
| 91 |
+
"<|text_sep|>",
|
| 92 |
+
"<|start_content|>",
|
| 93 |
+
"<|end_content|>"
|
| 94 |
+
],
|
| 95 |
+
"bos_token": "<s>",
|
| 96 |
+
"clean_up_tokenization_spaces": false,
|
| 97 |
+
"eos_token": "</s>",
|
| 98 |
+
"extra_special_tokens": {},
|
| 99 |
+
"legacy": true,
|
| 100 |
+
"model_max_length": 2048,
|
| 101 |
+
"pad_token": "<pad>",
|
| 102 |
+
"sp_model_kwargs": {},
|
| 103 |
+
"spaces_between_special_tokens": false,
|
| 104 |
+
"tokenizer_class": "LlamaTokenizer",
|
| 105 |
+
"unk_token": "<unk>",
|
| 106 |
+
"use_default_system_prompt": false
|
| 107 |
+
}
|