Upload folder using huggingface_hub
Browse files- .gitattributes +2 -0
- README.md +119 -0
- added_tokens.json +30 -0
- assets/covoaudio-results-overview.png +3 -0
- assets/mel_filters.npz +3 -0
- chat_template.jinja +54 -0
- config.json +121 -0
- configuration_covo_audio.py +164 -0
- generation_config.json +4 -0
- merges.txt +0 -0
- model-00001-of-00004.safetensors +3 -0
- model-00002-of-00004.safetensors +3 -0
- model-00003-of-00004.safetensors +3 -0
- model-00004-of-00004.safetensors +3 -0
- model.safetensors.index.json +852 -0
- modeling_covo_audio.py +406 -0
- special_tokens_map.json +24 -0
- token2wav/global_mean_var.npy +3 -0
- token2wav/model.pt +3 -0
- tokenizer.json +3 -0
- tokenizer_config.json +248 -0
- vocab.json +0 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,5 @@ saved_model/**/* 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
|
|
|
|
|
|
|
|
|
| 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
|
| 36 |
+
assets/covoaudio-results-overview.png filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,119 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Covo-Audio
|
| 2 |
+
|
| 3 |
+
<div align="center">
|
| 4 |
+
|
| 5 |
+
<h1>
|
| 6 |
+
Covo-Audio Technical Report
|
| 7 |
+
</h1>
|
| 8 |
+
|
| 9 |
+
[](https://arxiv.org/abs/2602.09823)
|
| 10 |
+
[](https://github.com/Tencent/Covo-Audio)
|
| 11 |
+
[](https://huggingface.co/tencent/Covo-Audio-Chat)
|
| 12 |
+
|
| 13 |
+
</div>
|
| 14 |
+
|
| 15 |
+
## 📖 Overview
|
| 16 |
+
|
| 17 |
+
Covo-Audio is a 7B-parameter end-to-end large audio language model that directly processes continuous audio inputs and generates audio outputs within a single unified architecture, which is presented in the paper [Covo-Audio Technical Report](https://arxiv.org/abs/2602.09823). The report introduces Covo-Audio-Chat and its variant, Covo-Audio-Chat-FD, with the former being released in this repository.
|
| 18 |
+
|
| 19 |
+
<div align="center">
|
| 20 |
+
<figure>
|
| 21 |
+
<img src="assets/covoaudio-results-overview.png" alt="Covo-Audio-Chat Results" width="75%">
|
| 22 |
+
<br> <figcaption><em>An Overview of Comprehensive Performance Comparison.</em></figcaption>
|
| 23 |
+
</figure>
|
| 24 |
+
</div>
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
### Key Features
|
| 28 |
+
|
| 29 |
+
- **Hierarchical Tri-modal Speech-Text Interleaving**: We propose a framework designed to achieve deep alignment and fusion across modalities and scales. The Tri-modal aspect integrates continuous acoustic features, discrete speech tokens, and natural language text within a unified sequence, effectively bridging the gap between high-fidelity prosodic nuances and robust semantic structures.
|
| 30 |
+
|
| 31 |
+
- **Mitigating Intelligence-Speaker Coupling**: We propose a intelligence-speaker decoupling technique that decouples speaker from dialogue intelligence via multi-speaker training, then develop a contextual adaptation method to transfer and share high-quality TTS voice.
|
| 32 |
+
|
| 33 |
+
- **Native Full-Duplex Voice Interaction**: We evolve Covo-Audio into Covo-Audio-Chat-FD, a variant with native, low-latency full-duplex capability.
|
| 34 |
+
|
| 35 |
+
- **Comprehensive State-of-the-Art Performance**: Achieving state-of-the-art or competitive performance among models of comparable scale across a broad spectrum of tasks, including spoken dialogue, speech understanding, audio understanding, and full-duplex voice interaction.
|
| 36 |
+
|
| 37 |
+
|
| 38 |
+
## 🔧 Installation
|
| 39 |
+
### 1. Requirements
|
| 40 |
+
Recommends Python >= 3.11
|
| 41 |
+
|
| 42 |
+
```bash
|
| 43 |
+
conda create -n covoaudio python=3.11
|
| 44 |
+
conda activate covoaudio
|
| 45 |
+
pip install -r requirements.txt
|
| 46 |
+
```
|
| 47 |
+
|
| 48 |
+
### 2. Clone Repository
|
| 49 |
+
```bash
|
| 50 |
+
git clone https://github.com/Tencent/Covo-Audio.git
|
| 51 |
+
cd Covo-Audio
|
| 52 |
+
```
|
| 53 |
+
|
| 54 |
+
### 3. Download Pretrained Models
|
| 55 |
+
|
| 56 |
+
**Using HuggingFace:**
|
| 57 |
+
```bash
|
| 58 |
+
pip install huggingface-hub
|
| 59 |
+
hf download tencent/Covo-Audio-Chat --local-dir ./covoaudio
|
| 60 |
+
```
|
| 61 |
+
By running the above script, you can use the model downloaded from huggingface to override the directory of the same name in this repository. Or you can specify your own directory to store the model by modifying the `local-dir` argument (In this case, you need to edit the arguments `model_dir` and `decode_load_path` in `example.sh` accordingly before running the inference script).
|
| 62 |
+
|
| 63 |
+
|
| 64 |
+
## 🚀 Usage
|
| 65 |
+
|
| 66 |
+
### Run Inference Scripts
|
| 67 |
+
After completeing the configuration and model downloading, you can perform one-click inference by running the script:
|
| 68 |
+
```bash
|
| 69 |
+
bash example.sh
|
| 70 |
+
```
|
| 71 |
+
To perform interaction with our model, just replace the paths in `example.py` with your own audio files.
|
| 72 |
+
|
| 73 |
+
|
| 74 |
+
---
|
| 75 |
+
|
| 76 |
+
## 🙏 Acknowledgments
|
| 77 |
+
|
| 78 |
+
Part of the code for this project is based on the following open-source projects:
|
| 79 |
+
- [**Transformers**](https://github.com/huggingface/transformers)
|
| 80 |
+
- [**BigVGAN**](https://github.com/NVIDIA/BigVGAN)
|
| 81 |
+
|
| 82 |
+
The llm backbone and audio encoder of Covo-Audio are initialized respectively with the weights from:
|
| 83 |
+
- [**Qwen2.5-7B**](https://huggingface.co/Qwen/Qwen2.5-7B)
|
| 84 |
+
- [**Whisper**](https://huggingface.co/openai/whisper-large-v3)
|
| 85 |
+
|
| 86 |
+
---
|
| 87 |
+
|
| 88 |
+
## 🔗 Citation
|
| 89 |
+
If you find this model useful, please cite our paper:
|
| 90 |
+
|
| 91 |
+
```bibtex
|
| 92 |
+
@misc{wang2026covoaudiotechnicalreport,
|
| 93 |
+
title={Covo-Audio Technical Report},
|
| 94 |
+
author={Wenfu Wang and Chenxing Li and Liqiang Zhang and Yiyang Zhao and Yuxiang Zou and Hanzhao Li and Mingyu Cui and Hao Zhang and Kun Wei and Le Xu and Zikang Huang and Jiajun Xu and Jiliang Hu and Xiang He and Zeyu Xie and Jiawen Kang and Youjun Chen and Meng Yu and Dong Yu and Rilin Chen and Linlin Di and Shulin Feng and Na Hu and Yang Liu and Bang Wang and Shan Yang},
|
| 95 |
+
year={2026},
|
| 96 |
+
eprint={2602.09823},
|
| 97 |
+
archivePrefix={arXiv},
|
| 98 |
+
primaryClass={cs.SD},
|
| 99 |
+
url={https://arxiv.org/abs/2602.09823},
|
| 100 |
+
}
|
| 101 |
+
```
|
| 102 |
+
|
| 103 |
+
## 📄 License
|
| 104 |
+
Our model and code are licensed under [Apache 2.0](LICENSE) License.
|
| 105 |
+
|
| 106 |
+
|
| 107 |
+
## ✉️ Contact
|
| 108 |
+
If you have any questions or suggestions, feel free to contact us:
|
| 109 |
+
|
| 110 |
+
[](mailto:wenfuwang@tencent.com)
|
| 111 |
+
[](mailto:chenxingli@tencent.com)
|
| 112 |
+
[](mailto:tatelqzhang@tencent.com)
|
| 113 |
+
[](mailto:yyangyzhao@tencent.com)
|
| 114 |
+
[](mailto:yuxiangzou@tencent.com)
|
| 115 |
+
[](mailto:ericmycui@tencent.com)
|
| 116 |
+
[](mailto:draymondxu@tencent.com)
|
| 117 |
+
|
| 118 |
+
## 📔 Disclaimer
|
| 119 |
+
Covo-Audio-Chat is for research and experimental purposes only. It may occasionally produce inaccurate, inappropriate, biased, outdated, or factually incorrect content. Users should independently verify critical information, and are solely responsible for their use of the model and any consequences thereof.
|
added_tokens.json
ADDED
|
@@ -0,0 +1,30 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"</tool_call>": 151658,
|
| 3 |
+
"<tool_call>": 151657,
|
| 4 |
+
"<|begofcAUDIO|>": 151665,
|
| 5 |
+
"<|begofdAUDIO|>": 151668,
|
| 6 |
+
"<|box_end|>": 151649,
|
| 7 |
+
"<|box_start|>": 151648,
|
| 8 |
+
"<|cAUDIO|>": 151666,
|
| 9 |
+
"<|dAUDIO|>": 151669,
|
| 10 |
+
"<|endofcAUDIO|>": 151667,
|
| 11 |
+
"<|endofdAUDIO|>": 151670,
|
| 12 |
+
"<|endoftext|>": 151643,
|
| 13 |
+
"<|file_sep|>": 151664,
|
| 14 |
+
"<|fim_middle|>": 151660,
|
| 15 |
+
"<|fim_pad|>": 151662,
|
| 16 |
+
"<|fim_prefix|>": 151659,
|
| 17 |
+
"<|fim_suffix|>": 151661,
|
| 18 |
+
"<|im_end|>": 151645,
|
| 19 |
+
"<|im_start|>": 151644,
|
| 20 |
+
"<|image_pad|>": 151655,
|
| 21 |
+
"<|object_ref_end|>": 151647,
|
| 22 |
+
"<|object_ref_start|>": 151646,
|
| 23 |
+
"<|quad_end|>": 151651,
|
| 24 |
+
"<|quad_start|>": 151650,
|
| 25 |
+
"<|repo_name|>": 151663,
|
| 26 |
+
"<|video_pad|>": 151656,
|
| 27 |
+
"<|vision_end|>": 151653,
|
| 28 |
+
"<|vision_pad|>": 151654,
|
| 29 |
+
"<|vision_start|>": 151652
|
| 30 |
+
}
|
assets/covoaudio-results-overview.png
ADDED
|
Git LFS Details
|
assets/mel_filters.npz
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:7450ae70723a5ef9d341e3cee628c7cb0177f36ce42c44b7ed2bf3325f0f6d4c
|
| 3 |
+
size 4271
|
chat_template.jinja
ADDED
|
@@ -0,0 +1,54 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{%- if tools %}
|
| 2 |
+
{{- '<|im_start|>system\n' }}
|
| 3 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 4 |
+
{{- messages[0]['content'] }}
|
| 5 |
+
{%- else %}
|
| 6 |
+
{{- 'You are a helpful assistant.' }}
|
| 7 |
+
{%- endif %}
|
| 8 |
+
{{- "\n\n# Tools\n\nYou may call one or more functions to assist with the user query.\n\nYou are provided with function signatures within <tools></tools> XML tags:\n<tools>" }}
|
| 9 |
+
{%- for tool in tools %}
|
| 10 |
+
{{- "\n" }}
|
| 11 |
+
{{- tool | tojson }}
|
| 12 |
+
{%- endfor %}
|
| 13 |
+
{{- "\n</tools>\n\nFor each function call, return a json object with function name and arguments within <tool_call></tool_call> XML tags:\n<tool_call>\n{\"name\": <function-name>, \"arguments\": <args-json-object>}\n</tool_call><|im_end|>\n" }}
|
| 14 |
+
{%- else %}
|
| 15 |
+
{%- if messages[0]['role'] == 'system' %}
|
| 16 |
+
{{- '<|im_start|>system\n' + messages[0]['content'] + '<|im_end|>\n' }}
|
| 17 |
+
{%- else %}
|
| 18 |
+
{{- '<|im_start|>system\nYou are a helpful assistant.<|im_end|>\n' }}
|
| 19 |
+
{%- endif %}
|
| 20 |
+
{%- endif %}
|
| 21 |
+
{%- for message in messages %}
|
| 22 |
+
{%- if (message.role == "user") or (message.role == "system" and not loop.first) or (message.role == "assistant" and not message.tool_calls) %}
|
| 23 |
+
{{- '<|im_start|>' + message.role + '\n' + message.content + '<|im_end|>' + '\n' }}
|
| 24 |
+
{%- elif message.role == "assistant" %}
|
| 25 |
+
{{- '<|im_start|>' + message.role }}
|
| 26 |
+
{%- if message.content %}
|
| 27 |
+
{{- '\n' + message.content }}
|
| 28 |
+
{%- endif %}
|
| 29 |
+
{%- for tool_call in message.tool_calls %}
|
| 30 |
+
{%- if tool_call.function is defined %}
|
| 31 |
+
{%- set tool_call = tool_call.function %}
|
| 32 |
+
{%- endif %}
|
| 33 |
+
{{- '\n<tool_call>\n{"name": "' }}
|
| 34 |
+
{{- tool_call.name }}
|
| 35 |
+
{{- '", "arguments": ' }}
|
| 36 |
+
{{- tool_call.arguments | tojson }}
|
| 37 |
+
{{- '}\n</tool_call>' }}
|
| 38 |
+
{%- endfor %}
|
| 39 |
+
{{- '<|im_end|>\n' }}
|
| 40 |
+
{%- elif message.role == "tool" %}
|
| 41 |
+
{%- if (loop.index0 == 0) or (messages[loop.index0 - 1].role != "tool") %}
|
| 42 |
+
{{- '<|im_start|>user' }}
|
| 43 |
+
{%- endif %}
|
| 44 |
+
{{- '\n<tool_response>\n' }}
|
| 45 |
+
{{- message.content }}
|
| 46 |
+
{{- '\n</tool_response>' }}
|
| 47 |
+
{%- if loop.last or (messages[loop.index0 + 1].role != "tool") %}
|
| 48 |
+
{{- '<|im_end|>\n' }}
|
| 49 |
+
{%- endif %}
|
| 50 |
+
{%- endif %}
|
| 51 |
+
{%- endfor %}
|
| 52 |
+
{%- if add_generation_prompt %}
|
| 53 |
+
{{- '<|im_start|>assistant\n' }}
|
| 54 |
+
{%- endif %}
|
config.json
ADDED
|
@@ -0,0 +1,121 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_encoder_config_type": "whisper",
|
| 3 |
+
"_llm_config_type": "qwen2",
|
| 4 |
+
"adapter_downsample": 8,
|
| 5 |
+
"architectures": [
|
| 6 |
+
"CovoAudioForCausalLM"
|
| 7 |
+
],
|
| 8 |
+
"audio_token_index": 151671,
|
| 9 |
+
"auto_map": {
|
| 10 |
+
"AutoConfig": "configuration_covo_audio.CovoAudioConfig",
|
| 11 |
+
"AutoModel": "modeling_covo_audio.CovoAudioForCausalLM"
|
| 12 |
+
},
|
| 13 |
+
"dtype": "bfloat16",
|
| 14 |
+
"encoder_config": {
|
| 15 |
+
"_name_or_path": "openai/whisper-large-v3",
|
| 16 |
+
"activation_dropout": 0.0,
|
| 17 |
+
"activation_function": "gelu",
|
| 18 |
+
"apply_spec_augment": false,
|
| 19 |
+
"architectures": [
|
| 20 |
+
"WhisperForConditionalGeneration"
|
| 21 |
+
],
|
| 22 |
+
"attention_dropout": 0.0,
|
| 23 |
+
"begin_suppress_tokens": [
|
| 24 |
+
220,
|
| 25 |
+
50257
|
| 26 |
+
],
|
| 27 |
+
"bos_token_id": 50257,
|
| 28 |
+
"classifier_proj_size": 256,
|
| 29 |
+
"d_model": 1280,
|
| 30 |
+
"decoder_attention_heads": 20,
|
| 31 |
+
"decoder_ffn_dim": 5120,
|
| 32 |
+
"decoder_layerdrop": 0.0,
|
| 33 |
+
"decoder_layers": 32,
|
| 34 |
+
"decoder_start_token_id": 50258,
|
| 35 |
+
"dropout": 0.0,
|
| 36 |
+
"dtype": "float16",
|
| 37 |
+
"encoder_attention_heads": 20,
|
| 38 |
+
"encoder_ffn_dim": 5120,
|
| 39 |
+
"encoder_layerdrop": 0.0,
|
| 40 |
+
"encoder_layers": 32,
|
| 41 |
+
"eos_token_id": 50257,
|
| 42 |
+
"init_std": 0.02,
|
| 43 |
+
"mask_feature_length": 10,
|
| 44 |
+
"mask_feature_min_masks": 0,
|
| 45 |
+
"mask_feature_prob": 0.0,
|
| 46 |
+
"mask_time_length": 10,
|
| 47 |
+
"mask_time_min_masks": 2,
|
| 48 |
+
"mask_time_prob": 0.05,
|
| 49 |
+
"max_length": 448,
|
| 50 |
+
"max_source_positions": 1500,
|
| 51 |
+
"max_target_positions": 448,
|
| 52 |
+
"median_filter_width": 7,
|
| 53 |
+
"model_type": "whisper",
|
| 54 |
+
"num_hidden_layers": 32,
|
| 55 |
+
"num_mel_bins": 128,
|
| 56 |
+
"scale_embedding": false,
|
| 57 |
+
"use_cache": true,
|
| 58 |
+
"use_weighted_layer_sum": false,
|
| 59 |
+
"vocab_size": 51866
|
| 60 |
+
},
|
| 61 |
+
"llm_config": {
|
| 62 |
+
"architectures": [
|
| 63 |
+
"Qwen2ForCausalLM"
|
| 64 |
+
],
|
| 65 |
+
"attention_dropout": 0.0,
|
| 66 |
+
"bos_token_id": 151643,
|
| 67 |
+
"dtype": "bfloat16",
|
| 68 |
+
"eos_token_id": 151643,
|
| 69 |
+
"hidden_act": "silu",
|
| 70 |
+
"hidden_size": 3584,
|
| 71 |
+
"initializer_range": 0.02,
|
| 72 |
+
"intermediate_size": 18944,
|
| 73 |
+
"layer_types": [
|
| 74 |
+
"full_attention",
|
| 75 |
+
"full_attention",
|
| 76 |
+
"full_attention",
|
| 77 |
+
"full_attention",
|
| 78 |
+
"full_attention",
|
| 79 |
+
"full_attention",
|
| 80 |
+
"full_attention",
|
| 81 |
+
"full_attention",
|
| 82 |
+
"full_attention",
|
| 83 |
+
"full_attention",
|
| 84 |
+
"full_attention",
|
| 85 |
+
"full_attention",
|
| 86 |
+
"full_attention",
|
| 87 |
+
"full_attention",
|
| 88 |
+
"full_attention",
|
| 89 |
+
"full_attention",
|
| 90 |
+
"full_attention",
|
| 91 |
+
"full_attention",
|
| 92 |
+
"full_attention",
|
| 93 |
+
"full_attention",
|
| 94 |
+
"full_attention",
|
| 95 |
+
"full_attention",
|
| 96 |
+
"full_attention",
|
| 97 |
+
"full_attention",
|
| 98 |
+
"full_attention",
|
| 99 |
+
"full_attention",
|
| 100 |
+
"full_attention",
|
| 101 |
+
"full_attention"
|
| 102 |
+
],
|
| 103 |
+
"max_position_embeddings": 131072,
|
| 104 |
+
"max_window_layers": 28,
|
| 105 |
+
"model_type": "qwen2",
|
| 106 |
+
"num_attention_heads": 28,
|
| 107 |
+
"num_hidden_layers": 28,
|
| 108 |
+
"num_key_value_heads": 4,
|
| 109 |
+
"rms_norm_eps": 1e-06,
|
| 110 |
+
"rope_scaling": null,
|
| 111 |
+
"rope_theta": 1000000.0,
|
| 112 |
+
"sliding_window": null,
|
| 113 |
+
"use_cache": true,
|
| 114 |
+
"use_mrope": false,
|
| 115 |
+
"use_sliding_window": false,
|
| 116 |
+
"vocab_size": 168055
|
| 117 |
+
},
|
| 118 |
+
"model_type": "covo_audio",
|
| 119 |
+
"transformers_version": "4.57.1",
|
| 120 |
+
"whisper_feats_dim": 1280
|
| 121 |
+
}
|
configuration_covo_audio.py
ADDED
|
@@ -0,0 +1,164 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from typing import Optional
|
| 2 |
+
|
| 3 |
+
from transformers import Qwen2Config, WhisperConfig
|
| 4 |
+
from transformers.configuration_utils import PretrainedConfig
|
| 5 |
+
|
| 6 |
+
|
| 7 |
+
class CovoAudioConfig(PretrainedConfig):
|
| 8 |
+
model_type = "covo_audio"
|
| 9 |
+
sub_configs = {"llm_config": Qwen2Config, "encoder_config": WhisperConfig} # type: ignore
|
| 10 |
+
has_no_defaults_at_init = True
|
| 11 |
+
def __init__(self,
|
| 12 |
+
llm_config:Optional[Qwen2Config]=None,
|
| 13 |
+
encoder_config:Optional[WhisperConfig]=None,
|
| 14 |
+
audio_token_index=151671,
|
| 15 |
+
adapter_downsample=8,
|
| 16 |
+
**kwargs):
|
| 17 |
+
|
| 18 |
+
if llm_config is None:
|
| 19 |
+
llm_config = Qwen2Config(
|
| 20 |
+
architectures=[
|
| 21 |
+
"Qwen2ForCausalLM"
|
| 22 |
+
],
|
| 23 |
+
attention_dropout=0.0,
|
| 24 |
+
bos_token_id=151643,
|
| 25 |
+
eos_token_id=151643,
|
| 26 |
+
hidden_act="silu",
|
| 27 |
+
hidden_size=3584,
|
| 28 |
+
initializer_range=0.02,
|
| 29 |
+
intermediate_size=18944,
|
| 30 |
+
max_position_embeddings=131072,
|
| 31 |
+
max_window_layers=28,
|
| 32 |
+
model_type="qwen2",
|
| 33 |
+
num_attention_heads=28,
|
| 34 |
+
num_hidden_layers=28,
|
| 35 |
+
num_key_value_heads=4,
|
| 36 |
+
rms_norm_eps=1e-06,
|
| 37 |
+
rope_scaling=None,
|
| 38 |
+
rope_theta=1000000.0,
|
| 39 |
+
sliding_window=131072,
|
| 40 |
+
torch_dtype="bfloat16",
|
| 41 |
+
use_cache=True,
|
| 42 |
+
use_mrope=False,
|
| 43 |
+
use_sliding_window=False,
|
| 44 |
+
vocab_size=168055
|
| 45 |
+
)
|
| 46 |
+
if encoder_config is None:
|
| 47 |
+
encoder_config = WhisperConfig(
|
| 48 |
+
_name_or_path="openai/whisper-large-v3",
|
| 49 |
+
activation_dropout=0.0,
|
| 50 |
+
activation_function="gelu",
|
| 51 |
+
apply_spec_augment=False,
|
| 52 |
+
architectures=[
|
| 53 |
+
"WhisperForConditionalGeneration"
|
| 54 |
+
],
|
| 55 |
+
attention_dropout=0.0,
|
| 56 |
+
begin_suppress_tokens=[
|
| 57 |
+
220,
|
| 58 |
+
50257
|
| 59 |
+
],
|
| 60 |
+
bos_token_id=50257,
|
| 61 |
+
classifier_proj_size=256,
|
| 62 |
+
d_model=1280,
|
| 63 |
+
decoder_attention_heads=20,
|
| 64 |
+
decoder_ffn_dim=5120,
|
| 65 |
+
decoder_layerdrop=0.0,
|
| 66 |
+
decoder_layers=32,
|
| 67 |
+
decoder_start_token_id=50258,
|
| 68 |
+
dropout=0.0,
|
| 69 |
+
encoder_attention_heads=20,
|
| 70 |
+
encoder_ffn_dim=5120,
|
| 71 |
+
encoder_layerdrop=0.0,
|
| 72 |
+
encoder_layers=32,
|
| 73 |
+
eos_token_id=50257,
|
| 74 |
+
init_std=0.02,
|
| 75 |
+
mask_feature_length=10,
|
| 76 |
+
mask_feature_min_masks=0,
|
| 77 |
+
mask_feature_prob=0.0,
|
| 78 |
+
mask_time_length=10,
|
| 79 |
+
mask_time_min_masks=2,
|
| 80 |
+
mask_time_prob=0.05,
|
| 81 |
+
max_length=448,
|
| 82 |
+
max_source_positions=1500,
|
| 83 |
+
max_target_positions=448,
|
| 84 |
+
median_filter_width=7,
|
| 85 |
+
model_type="whisper",
|
| 86 |
+
num_hidden_layers=32,
|
| 87 |
+
num_mel_bins=128,
|
| 88 |
+
scale_embedding=False,
|
| 89 |
+
torch_dtype="float16",
|
| 90 |
+
use_cache=True,
|
| 91 |
+
use_weighted_layer_sum=False,
|
| 92 |
+
vocab_size=51866
|
| 93 |
+
)
|
| 94 |
+
|
| 95 |
+
self.audio_token_index = audio_token_index
|
| 96 |
+
self.adapter_downsample = adapter_downsample
|
| 97 |
+
self.llm_config = llm_config
|
| 98 |
+
self.encoder_config = encoder_config
|
| 99 |
+
self.whisper_feats_dim = encoder_config.d_model
|
| 100 |
+
|
| 101 |
+
if "dtype" not in kwargs:
|
| 102 |
+
kwargs["dtype"] = "bfloat16"
|
| 103 |
+
self.dtype = kwargs["dtype"]
|
| 104 |
+
|
| 105 |
+
super().__init__(**kwargs)
|
| 106 |
+
|
| 107 |
+
@property
|
| 108 |
+
def num_hidden_layers(self):
|
| 109 |
+
return self.llm_config.num_hidden_layers
|
| 110 |
+
|
| 111 |
+
@property
|
| 112 |
+
def hidden_size(self):
|
| 113 |
+
return self.llm_config.hidden_size
|
| 114 |
+
|
| 115 |
+
def to_dict(self):
|
| 116 |
+
"""Serializes this instance to a Python dictionary, ensuring nested
|
| 117 |
+
PretrainedConfig objects are serialized via their own to_dict().
|
| 118 |
+
"""
|
| 119 |
+
output = super().to_dict()
|
| 120 |
+
# replace nested config objects with their dict representation
|
| 121 |
+
if hasattr(self, "llm_config") and isinstance(self.llm_config, PretrainedConfig):
|
| 122 |
+
output["llm_config"] = self.llm_config.to_dict()
|
| 123 |
+
output["_llm_config_type"] = getattr(self.llm_config, "model_type", None)
|
| 124 |
+
if hasattr(self, "encoder_config") and isinstance(self.encoder_config, PretrainedConfig):
|
| 125 |
+
output["encoder_config"] = self.encoder_config.to_dict()
|
| 126 |
+
output["_encoder_config_type"] = getattr(self.encoder_config, "model_type", None)
|
| 127 |
+
|
| 128 |
+
return output
|
| 129 |
+
|
| 130 |
+
@classmethod
|
| 131 |
+
def from_dict(cls, config_dict: dict, **kwargs):
|
| 132 |
+
"""Create an CovoAudioConfig from a dict, reconstructing nested config
|
| 133 |
+
objects (llm_config, encoder_config) using the classes declared in
|
| 134 |
+
`sub_configs` if available.
|
| 135 |
+
"""
|
| 136 |
+
# Make a shallow copy to avoid mutating input
|
| 137 |
+
data = dict(config_dict)
|
| 138 |
+
|
| 139 |
+
llm_conf = None
|
| 140 |
+
enc_conf = None
|
| 141 |
+
|
| 142 |
+
if "llm_config" in data and data["llm_config"] is not None:
|
| 143 |
+
llm_cls = cls.sub_configs.get("llm_config") if hasattr(cls, "sub_configs") else None
|
| 144 |
+
if llm_cls is not None:
|
| 145 |
+
# use the sub-config class to reconstruct
|
| 146 |
+
llm_conf = llm_cls.from_dict(data.pop("llm_config"))
|
| 147 |
+
else:
|
| 148 |
+
# fallback to raw dict
|
| 149 |
+
llm_conf = data.pop("llm_config")
|
| 150 |
+
|
| 151 |
+
if "encoder_config" in data and data["encoder_config"] is not None:
|
| 152 |
+
enc_cls = cls.sub_configs.get("encoder_config") if hasattr(cls, "sub_configs") else None
|
| 153 |
+
if enc_cls is not None:
|
| 154 |
+
enc_conf = enc_cls.from_dict(data.pop("encoder_config"))
|
| 155 |
+
else:
|
| 156 |
+
enc_conf = data.pop("encoder_config")
|
| 157 |
+
# ensure HF-compatible fields reflect the underlying decoder (LLM）
|
| 158 |
+
|
| 159 |
+
# remove internal helper keys if present
|
| 160 |
+
data.pop("_llm_config_type", None)
|
| 161 |
+
data.pop("_encoder_config_type", None)
|
| 162 |
+
|
| 163 |
+
# now construct instance using reconstructed nested configs
|
| 164 |
+
return cls(llm_config=llm_conf, encoder_config=enc_conf, **data)
|
generation_config.json
ADDED
|
@@ -0,0 +1,4 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"transformers_version": "4.57.1"
|
| 4 |
+
}
|
merges.txt
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|
model-00001-of-00004.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:93e9dbb69e5484a5568667adf73f62407bd2a22d27c0564942736aada7afeda0
|
| 3 |
+
size 4992284680
|
model-00002-of-00004.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:52f0e2ce2b1323bb26979de09d964747e6f1626d553c1f719802f605800b6731
|
| 3 |
+
size 4932751496
|
model-00003-of-00004.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:96ff2a30bf7b2247bc59fc43d1455aea7a5bb150211944efc7942ec873835e86
|
| 3 |
+
size 4330865648
|
model-00004-of-00004.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b3c843610053d528de9c129d20a6804c7c09396f91a98c1198e2b4dfed087e1e
|
| 3 |
+
size 2549030600
|
model.safetensors.index.json
ADDED
|
@@ -0,0 +1,852 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"metadata": {
|
| 3 |
+
"total_parameters": 8402417920,
|
| 4 |
+
"total_size": 16804835840
|
| 5 |
+
},
|
| 6 |
+
"weight_map": {
|
| 7 |
+
"audio_adapter.downsample_layers.0.conv1d.bias": "model-00004-of-00004.safetensors",
|
| 8 |
+
"audio_adapter.downsample_layers.0.conv1d.weight": "model-00004-of-00004.safetensors",
|
| 9 |
+
"audio_adapter.downsample_layers.0.linear1.bias": "model-00004-of-00004.safetensors",
|
| 10 |
+
"audio_adapter.downsample_layers.0.linear1.weight": "model-00004-of-00004.safetensors",
|
| 11 |
+
"audio_adapter.downsample_layers.0.linear2.bias": "model-00004-of-00004.safetensors",
|
| 12 |
+
"audio_adapter.downsample_layers.0.linear2.weight": "model-00004-of-00004.safetensors",
|
| 13 |
+
"audio_adapter.downsample_layers.1.conv1d.bias": "model-00004-of-00004.safetensors",
|
| 14 |
+
"audio_adapter.downsample_layers.1.conv1d.weight": "model-00004-of-00004.safetensors",
|
| 15 |
+
"audio_adapter.downsample_layers.1.linear1.bias": "model-00004-of-00004.safetensors",
|
| 16 |
+
"audio_adapter.downsample_layers.1.linear1.weight": "model-00004-of-00004.safetensors",
|
| 17 |
+
"audio_adapter.downsample_layers.1.linear2.bias": "model-00004-of-00004.safetensors",
|
| 18 |
+
"audio_adapter.downsample_layers.1.linear2.weight": "model-00004-of-00004.safetensors",
|
| 19 |
+
"audio_adapter.downsample_layers.2.conv1d.bias": "model-00004-of-00004.safetensors",
|
| 20 |
+
"audio_adapter.downsample_layers.2.conv1d.weight": "model-00004-of-00004.safetensors",
|
| 21 |
+
"audio_adapter.downsample_layers.2.linear1.bias": "model-00004-of-00004.safetensors",
|
| 22 |
+
"audio_adapter.downsample_layers.2.linear1.weight": "model-00004-of-00004.safetensors",
|
| 23 |
+
"audio_adapter.downsample_layers.2.linear2.bias": "model-00004-of-00004.safetensors",
|
| 24 |
+
"audio_adapter.downsample_layers.2.linear2.weight": "model-00004-of-00004.safetensors",
|
| 25 |
+
"encoder.conv1.bias": "model-00004-of-00004.safetensors",
|
| 26 |
+
"encoder.conv1.weight": "model-00004-of-00004.safetensors",
|
| 27 |
+
"encoder.conv2.bias": "model-00004-of-00004.safetensors",
|
| 28 |
+
"encoder.conv2.weight": "model-00004-of-00004.safetensors",
|
| 29 |
+
"encoder.embed_positions.weight": "model-00004-of-00004.safetensors",
|
| 30 |
+
"encoder.layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 31 |
+
"encoder.layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 32 |
+
"encoder.layers.0.fc1.bias": "model-00004-of-00004.safetensors",
|
| 33 |
+
"encoder.layers.0.fc1.weight": "model-00004-of-00004.safetensors",
|
| 34 |
+
"encoder.layers.0.fc2.bias": "model-00004-of-00004.safetensors",
|
| 35 |
+
"encoder.layers.0.fc2.weight": "model-00004-of-00004.safetensors",
|
| 36 |
+
"encoder.layers.0.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 37 |
+
"encoder.layers.0.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 38 |
+
"encoder.layers.0.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 39 |
+
"encoder.layers.0.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 40 |
+
"encoder.layers.0.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 41 |
+
"encoder.layers.0.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 42 |
+
"encoder.layers.0.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 43 |
+
"encoder.layers.0.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 44 |
+
"encoder.layers.0.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 45 |
+
"encoder.layers.0.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 46 |
+
"encoder.layers.0.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 47 |
+
"encoder.layers.1.fc1.bias": "model-00004-of-00004.safetensors",
|
| 48 |
+
"encoder.layers.1.fc1.weight": "model-00004-of-00004.safetensors",
|
| 49 |
+
"encoder.layers.1.fc2.bias": "model-00004-of-00004.safetensors",
|
| 50 |
+
"encoder.layers.1.fc2.weight": "model-00004-of-00004.safetensors",
|
| 51 |
+
"encoder.layers.1.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 52 |
+
"encoder.layers.1.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 53 |
+
"encoder.layers.1.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 54 |
+
"encoder.layers.1.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 55 |
+
"encoder.layers.1.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 56 |
+
"encoder.layers.1.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 57 |
+
"encoder.layers.1.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 58 |
+
"encoder.layers.1.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 59 |
+
"encoder.layers.1.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 60 |
+
"encoder.layers.1.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 61 |
+
"encoder.layers.1.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 62 |
+
"encoder.layers.10.fc1.bias": "model-00004-of-00004.safetensors",
|
| 63 |
+
"encoder.layers.10.fc1.weight": "model-00004-of-00004.safetensors",
|
| 64 |
+
"encoder.layers.10.fc2.bias": "model-00004-of-00004.safetensors",
|
| 65 |
+
"encoder.layers.10.fc2.weight": "model-00004-of-00004.safetensors",
|
| 66 |
+
"encoder.layers.10.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 67 |
+
"encoder.layers.10.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 68 |
+
"encoder.layers.10.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 69 |
+
"encoder.layers.10.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 70 |
+
"encoder.layers.10.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 71 |
+
"encoder.layers.10.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 72 |
+
"encoder.layers.10.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 73 |
+
"encoder.layers.10.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 74 |
+
"encoder.layers.10.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 75 |
+
"encoder.layers.10.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 76 |
+
"encoder.layers.10.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 77 |
+
"encoder.layers.11.fc1.bias": "model-00004-of-00004.safetensors",
|
| 78 |
+
"encoder.layers.11.fc1.weight": "model-00004-of-00004.safetensors",
|
| 79 |
+
"encoder.layers.11.fc2.bias": "model-00004-of-00004.safetensors",
|
| 80 |
+
"encoder.layers.11.fc2.weight": "model-00004-of-00004.safetensors",
|
| 81 |
+
"encoder.layers.11.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 82 |
+
"encoder.layers.11.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 83 |
+
"encoder.layers.11.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 84 |
+
"encoder.layers.11.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 85 |
+
"encoder.layers.11.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 86 |
+
"encoder.layers.11.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 87 |
+
"encoder.layers.11.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 88 |
+
"encoder.layers.11.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 89 |
+
"encoder.layers.11.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 90 |
+
"encoder.layers.11.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 91 |
+
"encoder.layers.11.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 92 |
+
"encoder.layers.12.fc1.bias": "model-00004-of-00004.safetensors",
|
| 93 |
+
"encoder.layers.12.fc1.weight": "model-00004-of-00004.safetensors",
|
| 94 |
+
"encoder.layers.12.fc2.bias": "model-00004-of-00004.safetensors",
|
| 95 |
+
"encoder.layers.12.fc2.weight": "model-00004-of-00004.safetensors",
|
| 96 |
+
"encoder.layers.12.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 97 |
+
"encoder.layers.12.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 98 |
+
"encoder.layers.12.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 99 |
+
"encoder.layers.12.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 100 |
+
"encoder.layers.12.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 101 |
+
"encoder.layers.12.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 102 |
+
"encoder.layers.12.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 103 |
+
"encoder.layers.12.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 104 |
+
"encoder.layers.12.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 105 |
+
"encoder.layers.12.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 106 |
+
"encoder.layers.12.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 107 |
+
"encoder.layers.13.fc1.bias": "model-00004-of-00004.safetensors",
|
| 108 |
+
"encoder.layers.13.fc1.weight": "model-00004-of-00004.safetensors",
|
| 109 |
+
"encoder.layers.13.fc2.bias": "model-00004-of-00004.safetensors",
|
| 110 |
+
"encoder.layers.13.fc2.weight": "model-00004-of-00004.safetensors",
|
| 111 |
+
"encoder.layers.13.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 112 |
+
"encoder.layers.13.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 113 |
+
"encoder.layers.13.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 114 |
+
"encoder.layers.13.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 115 |
+
"encoder.layers.13.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 116 |
+
"encoder.layers.13.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 117 |
+
"encoder.layers.13.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 118 |
+
"encoder.layers.13.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 119 |
+
"encoder.layers.13.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 120 |
+
"encoder.layers.13.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 121 |
+
"encoder.layers.13.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 122 |
+
"encoder.layers.14.fc1.bias": "model-00004-of-00004.safetensors",
|
| 123 |
+
"encoder.layers.14.fc1.weight": "model-00004-of-00004.safetensors",
|
| 124 |
+
"encoder.layers.14.fc2.bias": "model-00004-of-00004.safetensors",
|
| 125 |
+
"encoder.layers.14.fc2.weight": "model-00004-of-00004.safetensors",
|
| 126 |
+
"encoder.layers.14.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 127 |
+
"encoder.layers.14.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 128 |
+
"encoder.layers.14.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 129 |
+
"encoder.layers.14.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 130 |
+
"encoder.layers.14.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 131 |
+
"encoder.layers.14.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 132 |
+
"encoder.layers.14.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 133 |
+
"encoder.layers.14.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 134 |
+
"encoder.layers.14.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 135 |
+
"encoder.layers.14.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 136 |
+
"encoder.layers.14.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 137 |
+
"encoder.layers.15.fc1.bias": "model-00004-of-00004.safetensors",
|
| 138 |
+
"encoder.layers.15.fc1.weight": "model-00004-of-00004.safetensors",
|
| 139 |
+
"encoder.layers.15.fc2.bias": "model-00004-of-00004.safetensors",
|
| 140 |
+
"encoder.layers.15.fc2.weight": "model-00004-of-00004.safetensors",
|
| 141 |
+
"encoder.layers.15.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 142 |
+
"encoder.layers.15.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 143 |
+
"encoder.layers.15.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 144 |
+
"encoder.layers.15.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 145 |
+
"encoder.layers.15.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 146 |
+
"encoder.layers.15.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 147 |
+
"encoder.layers.15.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 148 |
+
"encoder.layers.15.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 149 |
+
"encoder.layers.15.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 150 |
+
"encoder.layers.15.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 151 |
+
"encoder.layers.15.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 152 |
+
"encoder.layers.16.fc1.bias": "model-00004-of-00004.safetensors",
|
| 153 |
+
"encoder.layers.16.fc1.weight": "model-00004-of-00004.safetensors",
|
| 154 |
+
"encoder.layers.16.fc2.bias": "model-00004-of-00004.safetensors",
|
| 155 |
+
"encoder.layers.16.fc2.weight": "model-00004-of-00004.safetensors",
|
| 156 |
+
"encoder.layers.16.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 157 |
+
"encoder.layers.16.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 158 |
+
"encoder.layers.16.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 159 |
+
"encoder.layers.16.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 160 |
+
"encoder.layers.16.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 161 |
+
"encoder.layers.16.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 162 |
+
"encoder.layers.16.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 163 |
+
"encoder.layers.16.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 164 |
+
"encoder.layers.16.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 165 |
+
"encoder.layers.16.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 166 |
+
"encoder.layers.16.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 167 |
+
"encoder.layers.17.fc1.bias": "model-00004-of-00004.safetensors",
|
| 168 |
+
"encoder.layers.17.fc1.weight": "model-00004-of-00004.safetensors",
|
| 169 |
+
"encoder.layers.17.fc2.bias": "model-00004-of-00004.safetensors",
|
| 170 |
+
"encoder.layers.17.fc2.weight": "model-00004-of-00004.safetensors",
|
| 171 |
+
"encoder.layers.17.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 172 |
+
"encoder.layers.17.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 173 |
+
"encoder.layers.17.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 174 |
+
"encoder.layers.17.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 175 |
+
"encoder.layers.17.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 176 |
+
"encoder.layers.17.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 177 |
+
"encoder.layers.17.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 178 |
+
"encoder.layers.17.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 179 |
+
"encoder.layers.17.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 180 |
+
"encoder.layers.17.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 181 |
+
"encoder.layers.17.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 182 |
+
"encoder.layers.18.fc1.bias": "model-00004-of-00004.safetensors",
|
| 183 |
+
"encoder.layers.18.fc1.weight": "model-00004-of-00004.safetensors",
|
| 184 |
+
"encoder.layers.18.fc2.bias": "model-00004-of-00004.safetensors",
|
| 185 |
+
"encoder.layers.18.fc2.weight": "model-00004-of-00004.safetensors",
|
| 186 |
+
"encoder.layers.18.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 187 |
+
"encoder.layers.18.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 188 |
+
"encoder.layers.18.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 189 |
+
"encoder.layers.18.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 190 |
+
"encoder.layers.18.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 191 |
+
"encoder.layers.18.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 192 |
+
"encoder.layers.18.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 193 |
+
"encoder.layers.18.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 194 |
+
"encoder.layers.18.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 195 |
+
"encoder.layers.18.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 196 |
+
"encoder.layers.18.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 197 |
+
"encoder.layers.19.fc1.bias": "model-00004-of-00004.safetensors",
|
| 198 |
+
"encoder.layers.19.fc1.weight": "model-00004-of-00004.safetensors",
|
| 199 |
+
"encoder.layers.19.fc2.bias": "model-00004-of-00004.safetensors",
|
| 200 |
+
"encoder.layers.19.fc2.weight": "model-00004-of-00004.safetensors",
|
| 201 |
+
"encoder.layers.19.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 202 |
+
"encoder.layers.19.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 203 |
+
"encoder.layers.19.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 204 |
+
"encoder.layers.19.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 205 |
+
"encoder.layers.19.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 206 |
+
"encoder.layers.19.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 207 |
+
"encoder.layers.19.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 208 |
+
"encoder.layers.19.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 209 |
+
"encoder.layers.19.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 210 |
+
"encoder.layers.19.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 211 |
+
"encoder.layers.19.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 212 |
+
"encoder.layers.2.fc1.bias": "model-00004-of-00004.safetensors",
|
| 213 |
+
"encoder.layers.2.fc1.weight": "model-00004-of-00004.safetensors",
|
| 214 |
+
"encoder.layers.2.fc2.bias": "model-00004-of-00004.safetensors",
|
| 215 |
+
"encoder.layers.2.fc2.weight": "model-00004-of-00004.safetensors",
|
| 216 |
+
"encoder.layers.2.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 217 |
+
"encoder.layers.2.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 218 |
+
"encoder.layers.2.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 219 |
+
"encoder.layers.2.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 220 |
+
"encoder.layers.2.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 221 |
+
"encoder.layers.2.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 222 |
+
"encoder.layers.2.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 223 |
+
"encoder.layers.2.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 224 |
+
"encoder.layers.2.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 225 |
+
"encoder.layers.2.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 226 |
+
"encoder.layers.2.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 227 |
+
"encoder.layers.20.fc1.bias": "model-00004-of-00004.safetensors",
|
| 228 |
+
"encoder.layers.20.fc1.weight": "model-00004-of-00004.safetensors",
|
| 229 |
+
"encoder.layers.20.fc2.bias": "model-00004-of-00004.safetensors",
|
| 230 |
+
"encoder.layers.20.fc2.weight": "model-00004-of-00004.safetensors",
|
| 231 |
+
"encoder.layers.20.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 232 |
+
"encoder.layers.20.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 233 |
+
"encoder.layers.20.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 234 |
+
"encoder.layers.20.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 235 |
+
"encoder.layers.20.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 236 |
+
"encoder.layers.20.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 237 |
+
"encoder.layers.20.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 238 |
+
"encoder.layers.20.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 239 |
+
"encoder.layers.20.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 240 |
+
"encoder.layers.20.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 241 |
+
"encoder.layers.20.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 242 |
+
"encoder.layers.21.fc1.bias": "model-00004-of-00004.safetensors",
|
| 243 |
+
"encoder.layers.21.fc1.weight": "model-00004-of-00004.safetensors",
|
| 244 |
+
"encoder.layers.21.fc2.bias": "model-00004-of-00004.safetensors",
|
| 245 |
+
"encoder.layers.21.fc2.weight": "model-00004-of-00004.safetensors",
|
| 246 |
+
"encoder.layers.21.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 247 |
+
"encoder.layers.21.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 248 |
+
"encoder.layers.21.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 249 |
+
"encoder.layers.21.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 250 |
+
"encoder.layers.21.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 251 |
+
"encoder.layers.21.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 252 |
+
"encoder.layers.21.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 253 |
+
"encoder.layers.21.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 254 |
+
"encoder.layers.21.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 255 |
+
"encoder.layers.21.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 256 |
+
"encoder.layers.21.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 257 |
+
"encoder.layers.22.fc1.bias": "model-00004-of-00004.safetensors",
|
| 258 |
+
"encoder.layers.22.fc1.weight": "model-00004-of-00004.safetensors",
|
| 259 |
+
"encoder.layers.22.fc2.bias": "model-00004-of-00004.safetensors",
|
| 260 |
+
"encoder.layers.22.fc2.weight": "model-00004-of-00004.safetensors",
|
| 261 |
+
"encoder.layers.22.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 262 |
+
"encoder.layers.22.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 263 |
+
"encoder.layers.22.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 264 |
+
"encoder.layers.22.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 265 |
+
"encoder.layers.22.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 266 |
+
"encoder.layers.22.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 267 |
+
"encoder.layers.22.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 268 |
+
"encoder.layers.22.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 269 |
+
"encoder.layers.22.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 270 |
+
"encoder.layers.22.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 271 |
+
"encoder.layers.22.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 272 |
+
"encoder.layers.23.fc1.bias": "model-00004-of-00004.safetensors",
|
| 273 |
+
"encoder.layers.23.fc1.weight": "model-00004-of-00004.safetensors",
|
| 274 |
+
"encoder.layers.23.fc2.bias": "model-00004-of-00004.safetensors",
|
| 275 |
+
"encoder.layers.23.fc2.weight": "model-00004-of-00004.safetensors",
|
| 276 |
+
"encoder.layers.23.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 277 |
+
"encoder.layers.23.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 278 |
+
"encoder.layers.23.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 279 |
+
"encoder.layers.23.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 280 |
+
"encoder.layers.23.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 281 |
+
"encoder.layers.23.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 282 |
+
"encoder.layers.23.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 283 |
+
"encoder.layers.23.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 284 |
+
"encoder.layers.23.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 285 |
+
"encoder.layers.23.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 286 |
+
"encoder.layers.23.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 287 |
+
"encoder.layers.24.fc1.bias": "model-00004-of-00004.safetensors",
|
| 288 |
+
"encoder.layers.24.fc1.weight": "model-00004-of-00004.safetensors",
|
| 289 |
+
"encoder.layers.24.fc2.bias": "model-00004-of-00004.safetensors",
|
| 290 |
+
"encoder.layers.24.fc2.weight": "model-00004-of-00004.safetensors",
|
| 291 |
+
"encoder.layers.24.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 292 |
+
"encoder.layers.24.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 293 |
+
"encoder.layers.24.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 294 |
+
"encoder.layers.24.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 295 |
+
"encoder.layers.24.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 296 |
+
"encoder.layers.24.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 297 |
+
"encoder.layers.24.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 298 |
+
"encoder.layers.24.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 299 |
+
"encoder.layers.24.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 300 |
+
"encoder.layers.24.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 301 |
+
"encoder.layers.24.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 302 |
+
"encoder.layers.25.fc1.bias": "model-00004-of-00004.safetensors",
|
| 303 |
+
"encoder.layers.25.fc1.weight": "model-00004-of-00004.safetensors",
|
| 304 |
+
"encoder.layers.25.fc2.bias": "model-00004-of-00004.safetensors",
|
| 305 |
+
"encoder.layers.25.fc2.weight": "model-00004-of-00004.safetensors",
|
| 306 |
+
"encoder.layers.25.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 307 |
+
"encoder.layers.25.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 308 |
+
"encoder.layers.25.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 309 |
+
"encoder.layers.25.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 310 |
+
"encoder.layers.25.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 311 |
+
"encoder.layers.25.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 312 |
+
"encoder.layers.25.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 313 |
+
"encoder.layers.25.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 314 |
+
"encoder.layers.25.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 315 |
+
"encoder.layers.25.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 316 |
+
"encoder.layers.25.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 317 |
+
"encoder.layers.26.fc1.bias": "model-00004-of-00004.safetensors",
|
| 318 |
+
"encoder.layers.26.fc1.weight": "model-00004-of-00004.safetensors",
|
| 319 |
+
"encoder.layers.26.fc2.bias": "model-00004-of-00004.safetensors",
|
| 320 |
+
"encoder.layers.26.fc2.weight": "model-00004-of-00004.safetensors",
|
| 321 |
+
"encoder.layers.26.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 322 |
+
"encoder.layers.26.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 323 |
+
"encoder.layers.26.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 324 |
+
"encoder.layers.26.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 325 |
+
"encoder.layers.26.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 326 |
+
"encoder.layers.26.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 327 |
+
"encoder.layers.26.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 328 |
+
"encoder.layers.26.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 329 |
+
"encoder.layers.26.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 330 |
+
"encoder.layers.26.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 331 |
+
"encoder.layers.26.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 332 |
+
"encoder.layers.27.fc1.bias": "model-00004-of-00004.safetensors",
|
| 333 |
+
"encoder.layers.27.fc1.weight": "model-00004-of-00004.safetensors",
|
| 334 |
+
"encoder.layers.27.fc2.bias": "model-00004-of-00004.safetensors",
|
| 335 |
+
"encoder.layers.27.fc2.weight": "model-00004-of-00004.safetensors",
|
| 336 |
+
"encoder.layers.27.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 337 |
+
"encoder.layers.27.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 338 |
+
"encoder.layers.27.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 339 |
+
"encoder.layers.27.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 340 |
+
"encoder.layers.27.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 341 |
+
"encoder.layers.27.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 342 |
+
"encoder.layers.27.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 343 |
+
"encoder.layers.27.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 344 |
+
"encoder.layers.27.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 345 |
+
"encoder.layers.27.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 346 |
+
"encoder.layers.27.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 347 |
+
"encoder.layers.28.fc1.bias": "model-00004-of-00004.safetensors",
|
| 348 |
+
"encoder.layers.28.fc1.weight": "model-00004-of-00004.safetensors",
|
| 349 |
+
"encoder.layers.28.fc2.bias": "model-00004-of-00004.safetensors",
|
| 350 |
+
"encoder.layers.28.fc2.weight": "model-00004-of-00004.safetensors",
|
| 351 |
+
"encoder.layers.28.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 352 |
+
"encoder.layers.28.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 353 |
+
"encoder.layers.28.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 354 |
+
"encoder.layers.28.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 355 |
+
"encoder.layers.28.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 356 |
+
"encoder.layers.28.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 357 |
+
"encoder.layers.28.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 358 |
+
"encoder.layers.28.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 359 |
+
"encoder.layers.28.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 360 |
+
"encoder.layers.28.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 361 |
+
"encoder.layers.28.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 362 |
+
"encoder.layers.29.fc1.bias": "model-00004-of-00004.safetensors",
|
| 363 |
+
"encoder.layers.29.fc1.weight": "model-00004-of-00004.safetensors",
|
| 364 |
+
"encoder.layers.29.fc2.bias": "model-00004-of-00004.safetensors",
|
| 365 |
+
"encoder.layers.29.fc2.weight": "model-00004-of-00004.safetensors",
|
| 366 |
+
"encoder.layers.29.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 367 |
+
"encoder.layers.29.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 368 |
+
"encoder.layers.29.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 369 |
+
"encoder.layers.29.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 370 |
+
"encoder.layers.29.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 371 |
+
"encoder.layers.29.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 372 |
+
"encoder.layers.29.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 373 |
+
"encoder.layers.29.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 374 |
+
"encoder.layers.29.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 375 |
+
"encoder.layers.29.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 376 |
+
"encoder.layers.29.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 377 |
+
"encoder.layers.3.fc1.bias": "model-00004-of-00004.safetensors",
|
| 378 |
+
"encoder.layers.3.fc1.weight": "model-00004-of-00004.safetensors",
|
| 379 |
+
"encoder.layers.3.fc2.bias": "model-00004-of-00004.safetensors",
|
| 380 |
+
"encoder.layers.3.fc2.weight": "model-00004-of-00004.safetensors",
|
| 381 |
+
"encoder.layers.3.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 382 |
+
"encoder.layers.3.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 383 |
+
"encoder.layers.3.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 384 |
+
"encoder.layers.3.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 385 |
+
"encoder.layers.3.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 386 |
+
"encoder.layers.3.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 387 |
+
"encoder.layers.3.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 388 |
+
"encoder.layers.3.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 389 |
+
"encoder.layers.3.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 390 |
+
"encoder.layers.3.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 391 |
+
"encoder.layers.3.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 392 |
+
"encoder.layers.30.fc1.bias": "model-00004-of-00004.safetensors",
|
| 393 |
+
"encoder.layers.30.fc1.weight": "model-00004-of-00004.safetensors",
|
| 394 |
+
"encoder.layers.30.fc2.bias": "model-00004-of-00004.safetensors",
|
| 395 |
+
"encoder.layers.30.fc2.weight": "model-00004-of-00004.safetensors",
|
| 396 |
+
"encoder.layers.30.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 397 |
+
"encoder.layers.30.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 398 |
+
"encoder.layers.30.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 399 |
+
"encoder.layers.30.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 400 |
+
"encoder.layers.30.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 401 |
+
"encoder.layers.30.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 402 |
+
"encoder.layers.30.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 403 |
+
"encoder.layers.30.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 404 |
+
"encoder.layers.30.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 405 |
+
"encoder.layers.30.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 406 |
+
"encoder.layers.30.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 407 |
+
"encoder.layers.31.fc1.bias": "model-00004-of-00004.safetensors",
|
| 408 |
+
"encoder.layers.31.fc1.weight": "model-00004-of-00004.safetensors",
|
| 409 |
+
"encoder.layers.31.fc2.bias": "model-00004-of-00004.safetensors",
|
| 410 |
+
"encoder.layers.31.fc2.weight": "model-00004-of-00004.safetensors",
|
| 411 |
+
"encoder.layers.31.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 412 |
+
"encoder.layers.31.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 413 |
+
"encoder.layers.31.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 414 |
+
"encoder.layers.31.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 415 |
+
"encoder.layers.31.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 416 |
+
"encoder.layers.31.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 417 |
+
"encoder.layers.31.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 418 |
+
"encoder.layers.31.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 419 |
+
"encoder.layers.31.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 420 |
+
"encoder.layers.31.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 421 |
+
"encoder.layers.31.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 422 |
+
"encoder.layers.4.fc1.bias": "model-00004-of-00004.safetensors",
|
| 423 |
+
"encoder.layers.4.fc1.weight": "model-00004-of-00004.safetensors",
|
| 424 |
+
"encoder.layers.4.fc2.bias": "model-00004-of-00004.safetensors",
|
| 425 |
+
"encoder.layers.4.fc2.weight": "model-00004-of-00004.safetensors",
|
| 426 |
+
"encoder.layers.4.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 427 |
+
"encoder.layers.4.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 428 |
+
"encoder.layers.4.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 429 |
+
"encoder.layers.4.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 430 |
+
"encoder.layers.4.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 431 |
+
"encoder.layers.4.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 432 |
+
"encoder.layers.4.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 433 |
+
"encoder.layers.4.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 434 |
+
"encoder.layers.4.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 435 |
+
"encoder.layers.4.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 436 |
+
"encoder.layers.4.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 437 |
+
"encoder.layers.5.fc1.bias": "model-00004-of-00004.safetensors",
|
| 438 |
+
"encoder.layers.5.fc1.weight": "model-00004-of-00004.safetensors",
|
| 439 |
+
"encoder.layers.5.fc2.bias": "model-00004-of-00004.safetensors",
|
| 440 |
+
"encoder.layers.5.fc2.weight": "model-00004-of-00004.safetensors",
|
| 441 |
+
"encoder.layers.5.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 442 |
+
"encoder.layers.5.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 443 |
+
"encoder.layers.5.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 444 |
+
"encoder.layers.5.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 445 |
+
"encoder.layers.5.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 446 |
+
"encoder.layers.5.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 447 |
+
"encoder.layers.5.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 448 |
+
"encoder.layers.5.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 449 |
+
"encoder.layers.5.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 450 |
+
"encoder.layers.5.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 451 |
+
"encoder.layers.5.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 452 |
+
"encoder.layers.6.fc1.bias": "model-00004-of-00004.safetensors",
|
| 453 |
+
"encoder.layers.6.fc1.weight": "model-00004-of-00004.safetensors",
|
| 454 |
+
"encoder.layers.6.fc2.bias": "model-00004-of-00004.safetensors",
|
| 455 |
+
"encoder.layers.6.fc2.weight": "model-00004-of-00004.safetensors",
|
| 456 |
+
"encoder.layers.6.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 457 |
+
"encoder.layers.6.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 458 |
+
"encoder.layers.6.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 459 |
+
"encoder.layers.6.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 460 |
+
"encoder.layers.6.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 461 |
+
"encoder.layers.6.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 462 |
+
"encoder.layers.6.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 463 |
+
"encoder.layers.6.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 464 |
+
"encoder.layers.6.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 465 |
+
"encoder.layers.6.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 466 |
+
"encoder.layers.6.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 467 |
+
"encoder.layers.7.fc1.bias": "model-00004-of-00004.safetensors",
|
| 468 |
+
"encoder.layers.7.fc1.weight": "model-00004-of-00004.safetensors",
|
| 469 |
+
"encoder.layers.7.fc2.bias": "model-00004-of-00004.safetensors",
|
| 470 |
+
"encoder.layers.7.fc2.weight": "model-00004-of-00004.safetensors",
|
| 471 |
+
"encoder.layers.7.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 472 |
+
"encoder.layers.7.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 473 |
+
"encoder.layers.7.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 474 |
+
"encoder.layers.7.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 475 |
+
"encoder.layers.7.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 476 |
+
"encoder.layers.7.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 477 |
+
"encoder.layers.7.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 478 |
+
"encoder.layers.7.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 479 |
+
"encoder.layers.7.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 480 |
+
"encoder.layers.7.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 481 |
+
"encoder.layers.7.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 482 |
+
"encoder.layers.8.fc1.bias": "model-00004-of-00004.safetensors",
|
| 483 |
+
"encoder.layers.8.fc1.weight": "model-00004-of-00004.safetensors",
|
| 484 |
+
"encoder.layers.8.fc2.bias": "model-00004-of-00004.safetensors",
|
| 485 |
+
"encoder.layers.8.fc2.weight": "model-00004-of-00004.safetensors",
|
| 486 |
+
"encoder.layers.8.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 487 |
+
"encoder.layers.8.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 488 |
+
"encoder.layers.8.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 489 |
+
"encoder.layers.8.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 490 |
+
"encoder.layers.8.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 491 |
+
"encoder.layers.8.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 492 |
+
"encoder.layers.8.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 493 |
+
"encoder.layers.8.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 494 |
+
"encoder.layers.8.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 495 |
+
"encoder.layers.8.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 496 |
+
"encoder.layers.8.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 497 |
+
"encoder.layers.9.fc1.bias": "model-00004-of-00004.safetensors",
|
| 498 |
+
"encoder.layers.9.fc1.weight": "model-00004-of-00004.safetensors",
|
| 499 |
+
"encoder.layers.9.fc2.bias": "model-00004-of-00004.safetensors",
|
| 500 |
+
"encoder.layers.9.fc2.weight": "model-00004-of-00004.safetensors",
|
| 501 |
+
"encoder.layers.9.final_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 502 |
+
"encoder.layers.9.final_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 503 |
+
"encoder.layers.9.self_attn.k_proj.weight": "model-00004-of-00004.safetensors",
|
| 504 |
+
"encoder.layers.9.self_attn.out_proj.bias": "model-00004-of-00004.safetensors",
|
| 505 |
+
"encoder.layers.9.self_attn.out_proj.weight": "model-00004-of-00004.safetensors",
|
| 506 |
+
"encoder.layers.9.self_attn.q_proj.bias": "model-00004-of-00004.safetensors",
|
| 507 |
+
"encoder.layers.9.self_attn.q_proj.weight": "model-00004-of-00004.safetensors",
|
| 508 |
+
"encoder.layers.9.self_attn.v_proj.bias": "model-00004-of-00004.safetensors",
|
| 509 |
+
"encoder.layers.9.self_attn.v_proj.weight": "model-00004-of-00004.safetensors",
|
| 510 |
+
"encoder.layers.9.self_attn_layer_norm.bias": "model-00004-of-00004.safetensors",
|
| 511 |
+
"encoder.layers.9.self_attn_layer_norm.weight": "model-00004-of-00004.safetensors",
|
| 512 |
+
"llm.lm_head.weight": "model-00004-of-00004.safetensors",
|
| 513 |
+
"llm.model.embed_tokens.weight": "model-00001-of-00004.safetensors",
|
| 514 |
+
"llm.model.layers.0.input_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 515 |
+
"llm.model.layers.0.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
|
| 516 |
+
"llm.model.layers.0.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
|
| 517 |
+
"llm.model.layers.0.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
|
| 518 |
+
"llm.model.layers.0.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 519 |
+
"llm.model.layers.0.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
|
| 520 |
+
"llm.model.layers.0.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
|
| 521 |
+
"llm.model.layers.0.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
|
| 522 |
+
"llm.model.layers.0.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
|
| 523 |
+
"llm.model.layers.0.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
|
| 524 |
+
"llm.model.layers.0.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
|
| 525 |
+
"llm.model.layers.0.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
|
| 526 |
+
"llm.model.layers.1.input_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 527 |
+
"llm.model.layers.1.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
|
| 528 |
+
"llm.model.layers.1.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
|
| 529 |
+
"llm.model.layers.1.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
|
| 530 |
+
"llm.model.layers.1.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 531 |
+
"llm.model.layers.1.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
|
| 532 |
+
"llm.model.layers.1.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
|
| 533 |
+
"llm.model.layers.1.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
|
| 534 |
+
"llm.model.layers.1.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
|
| 535 |
+
"llm.model.layers.1.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
|
| 536 |
+
"llm.model.layers.1.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
|
| 537 |
+
"llm.model.layers.1.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
|
| 538 |
+
"llm.model.layers.10.input_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 539 |
+
"llm.model.layers.10.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
|
| 540 |
+
"llm.model.layers.10.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
|
| 541 |
+
"llm.model.layers.10.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
|
| 542 |
+
"llm.model.layers.10.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 543 |
+
"llm.model.layers.10.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
|
| 544 |
+
"llm.model.layers.10.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
|
| 545 |
+
"llm.model.layers.10.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
|
| 546 |
+
"llm.model.layers.10.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
|
| 547 |
+
"llm.model.layers.10.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
|
| 548 |
+
"llm.model.layers.10.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
|
| 549 |
+
"llm.model.layers.10.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
|
| 550 |
+
"llm.model.layers.11.input_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 551 |
+
"llm.model.layers.11.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
|
| 552 |
+
"llm.model.layers.11.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
|
| 553 |
+
"llm.model.layers.11.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
|
| 554 |
+
"llm.model.layers.11.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 555 |
+
"llm.model.layers.11.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
|
| 556 |
+
"llm.model.layers.11.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
|
| 557 |
+
"llm.model.layers.11.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
|
| 558 |
+
"llm.model.layers.11.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
|
| 559 |
+
"llm.model.layers.11.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
|
| 560 |
+
"llm.model.layers.11.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
|
| 561 |
+
"llm.model.layers.11.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
|
| 562 |
+
"llm.model.layers.12.input_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 563 |
+
"llm.model.layers.12.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
|
| 564 |
+
"llm.model.layers.12.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
|
| 565 |
+
"llm.model.layers.12.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
|
| 566 |
+
"llm.model.layers.12.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 567 |
+
"llm.model.layers.12.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
|
| 568 |
+
"llm.model.layers.12.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
|
| 569 |
+
"llm.model.layers.12.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
|
| 570 |
+
"llm.model.layers.12.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
|
| 571 |
+
"llm.model.layers.12.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
|
| 572 |
+
"llm.model.layers.12.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
|
| 573 |
+
"llm.model.layers.12.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
|
| 574 |
+
"llm.model.layers.13.input_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 575 |
+
"llm.model.layers.13.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
|
| 576 |
+
"llm.model.layers.13.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
|
| 577 |
+
"llm.model.layers.13.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
|
| 578 |
+
"llm.model.layers.13.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 579 |
+
"llm.model.layers.13.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
|
| 580 |
+
"llm.model.layers.13.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
|
| 581 |
+
"llm.model.layers.13.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
|
| 582 |
+
"llm.model.layers.13.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
|
| 583 |
+
"llm.model.layers.13.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
|
| 584 |
+
"llm.model.layers.13.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
|
| 585 |
+
"llm.model.layers.13.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
|
| 586 |
+
"llm.model.layers.14.input_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 587 |
+
"llm.model.layers.14.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
|
| 588 |
+
"llm.model.layers.14.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
|
| 589 |
+
"llm.model.layers.14.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
|
| 590 |
+
"llm.model.layers.14.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 591 |
+
"llm.model.layers.14.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
|
| 592 |
+
"llm.model.layers.14.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
|
| 593 |
+
"llm.model.layers.14.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
|
| 594 |
+
"llm.model.layers.14.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
|
| 595 |
+
"llm.model.layers.14.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
|
| 596 |
+
"llm.model.layers.14.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
|
| 597 |
+
"llm.model.layers.14.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
|
| 598 |
+
"llm.model.layers.15.input_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 599 |
+
"llm.model.layers.15.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
|
| 600 |
+
"llm.model.layers.15.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
|
| 601 |
+
"llm.model.layers.15.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
|
| 602 |
+
"llm.model.layers.15.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 603 |
+
"llm.model.layers.15.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
|
| 604 |
+
"llm.model.layers.15.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
|
| 605 |
+
"llm.model.layers.15.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
|
| 606 |
+
"llm.model.layers.15.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
|
| 607 |
+
"llm.model.layers.15.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
|
| 608 |
+
"llm.model.layers.15.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
|
| 609 |
+
"llm.model.layers.15.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
|
| 610 |
+
"llm.model.layers.16.input_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 611 |
+
"llm.model.layers.16.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
|
| 612 |
+
"llm.model.layers.16.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
|
| 613 |
+
"llm.model.layers.16.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
|
| 614 |
+
"llm.model.layers.16.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 615 |
+
"llm.model.layers.16.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
|
| 616 |
+
"llm.model.layers.16.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
|
| 617 |
+
"llm.model.layers.16.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
|
| 618 |
+
"llm.model.layers.16.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
|
| 619 |
+
"llm.model.layers.16.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
|
| 620 |
+
"llm.model.layers.16.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
|
| 621 |
+
"llm.model.layers.16.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
|
| 622 |
+
"llm.model.layers.17.input_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 623 |
+
"llm.model.layers.17.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
|
| 624 |
+
"llm.model.layers.17.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
|
| 625 |
+
"llm.model.layers.17.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
|
| 626 |
+
"llm.model.layers.17.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 627 |
+
"llm.model.layers.17.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
|
| 628 |
+
"llm.model.layers.17.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
|
| 629 |
+
"llm.model.layers.17.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
|
| 630 |
+
"llm.model.layers.17.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
|
| 631 |
+
"llm.model.layers.17.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
|
| 632 |
+
"llm.model.layers.17.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
|
| 633 |
+
"llm.model.layers.17.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
|
| 634 |
+
"llm.model.layers.18.input_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 635 |
+
"llm.model.layers.18.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
|
| 636 |
+
"llm.model.layers.18.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
|
| 637 |
+
"llm.model.layers.18.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
|
| 638 |
+
"llm.model.layers.18.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 639 |
+
"llm.model.layers.18.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
|
| 640 |
+
"llm.model.layers.18.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
|
| 641 |
+
"llm.model.layers.18.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
|
| 642 |
+
"llm.model.layers.18.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
|
| 643 |
+
"llm.model.layers.18.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
|
| 644 |
+
"llm.model.layers.18.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
|
| 645 |
+
"llm.model.layers.18.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
|
| 646 |
+
"llm.model.layers.19.input_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 647 |
+
"llm.model.layers.19.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
|
| 648 |
+
"llm.model.layers.19.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
|
| 649 |
+
"llm.model.layers.19.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
|
| 650 |
+
"llm.model.layers.19.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 651 |
+
"llm.model.layers.19.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
|
| 652 |
+
"llm.model.layers.19.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
|
| 653 |
+
"llm.model.layers.19.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
|
| 654 |
+
"llm.model.layers.19.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
|
| 655 |
+
"llm.model.layers.19.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
|
| 656 |
+
"llm.model.layers.19.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
|
| 657 |
+
"llm.model.layers.19.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
|
| 658 |
+
"llm.model.layers.2.input_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 659 |
+
"llm.model.layers.2.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
|
| 660 |
+
"llm.model.layers.2.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
|
| 661 |
+
"llm.model.layers.2.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
|
| 662 |
+
"llm.model.layers.2.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 663 |
+
"llm.model.layers.2.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
|
| 664 |
+
"llm.model.layers.2.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
|
| 665 |
+
"llm.model.layers.2.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
|
| 666 |
+
"llm.model.layers.2.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
|
| 667 |
+
"llm.model.layers.2.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
|
| 668 |
+
"llm.model.layers.2.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
|
| 669 |
+
"llm.model.layers.2.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
|
| 670 |
+
"llm.model.layers.20.input_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 671 |
+
"llm.model.layers.20.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
|
| 672 |
+
"llm.model.layers.20.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
|
| 673 |
+
"llm.model.layers.20.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
|
| 674 |
+
"llm.model.layers.20.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 675 |
+
"llm.model.layers.20.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
|
| 676 |
+
"llm.model.layers.20.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
|
| 677 |
+
"llm.model.layers.20.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
|
| 678 |
+
"llm.model.layers.20.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
|
| 679 |
+
"llm.model.layers.20.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
|
| 680 |
+
"llm.model.layers.20.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
|
| 681 |
+
"llm.model.layers.20.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
|
| 682 |
+
"llm.model.layers.21.input_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 683 |
+
"llm.model.layers.21.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
|
| 684 |
+
"llm.model.layers.21.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
|
| 685 |
+
"llm.model.layers.21.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
|
| 686 |
+
"llm.model.layers.21.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 687 |
+
"llm.model.layers.21.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
|
| 688 |
+
"llm.model.layers.21.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
|
| 689 |
+
"llm.model.layers.21.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
|
| 690 |
+
"llm.model.layers.21.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
|
| 691 |
+
"llm.model.layers.21.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
|
| 692 |
+
"llm.model.layers.21.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
|
| 693 |
+
"llm.model.layers.21.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
|
| 694 |
+
"llm.model.layers.22.input_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 695 |
+
"llm.model.layers.22.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
|
| 696 |
+
"llm.model.layers.22.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
|
| 697 |
+
"llm.model.layers.22.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
|
| 698 |
+
"llm.model.layers.22.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 699 |
+
"llm.model.layers.22.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
|
| 700 |
+
"llm.model.layers.22.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
|
| 701 |
+
"llm.model.layers.22.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
|
| 702 |
+
"llm.model.layers.22.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
|
| 703 |
+
"llm.model.layers.22.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
|
| 704 |
+
"llm.model.layers.22.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
|
| 705 |
+
"llm.model.layers.22.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
|
| 706 |
+
"llm.model.layers.23.input_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 707 |
+
"llm.model.layers.23.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
|
| 708 |
+
"llm.model.layers.23.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
|
| 709 |
+
"llm.model.layers.23.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
|
| 710 |
+
"llm.model.layers.23.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 711 |
+
"llm.model.layers.23.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
|
| 712 |
+
"llm.model.layers.23.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
|
| 713 |
+
"llm.model.layers.23.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
|
| 714 |
+
"llm.model.layers.23.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
|
| 715 |
+
"llm.model.layers.23.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
|
| 716 |
+
"llm.model.layers.23.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
|
| 717 |
+
"llm.model.layers.23.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
|
| 718 |
+
"llm.model.layers.24.input_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 719 |
+
"llm.model.layers.24.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
|
| 720 |
+
"llm.model.layers.24.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
|
| 721 |
+
"llm.model.layers.24.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
|
| 722 |
+
"llm.model.layers.24.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 723 |
+
"llm.model.layers.24.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
|
| 724 |
+
"llm.model.layers.24.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
|
| 725 |
+
"llm.model.layers.24.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
|
| 726 |
+
"llm.model.layers.24.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
|
| 727 |
+
"llm.model.layers.24.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
|
| 728 |
+
"llm.model.layers.24.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
|
| 729 |
+
"llm.model.layers.24.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
|
| 730 |
+
"llm.model.layers.25.input_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 731 |
+
"llm.model.layers.25.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
|
| 732 |
+
"llm.model.layers.25.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
|
| 733 |
+
"llm.model.layers.25.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
|
| 734 |
+
"llm.model.layers.25.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 735 |
+
"llm.model.layers.25.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
|
| 736 |
+
"llm.model.layers.25.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
|
| 737 |
+
"llm.model.layers.25.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
|
| 738 |
+
"llm.model.layers.25.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
|
| 739 |
+
"llm.model.layers.25.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
|
| 740 |
+
"llm.model.layers.25.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
|
| 741 |
+
"llm.model.layers.25.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
|
| 742 |
+
"llm.model.layers.26.input_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 743 |
+
"llm.model.layers.26.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
|
| 744 |
+
"llm.model.layers.26.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
|
| 745 |
+
"llm.model.layers.26.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
|
| 746 |
+
"llm.model.layers.26.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 747 |
+
"llm.model.layers.26.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
|
| 748 |
+
"llm.model.layers.26.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
|
| 749 |
+
"llm.model.layers.26.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
|
| 750 |
+
"llm.model.layers.26.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
|
| 751 |
+
"llm.model.layers.26.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
|
| 752 |
+
"llm.model.layers.26.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
|
| 753 |
+
"llm.model.layers.26.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
|
| 754 |
+
"llm.model.layers.27.input_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 755 |
+
"llm.model.layers.27.mlp.down_proj.weight": "model-00003-of-00004.safetensors",
|
| 756 |
+
"llm.model.layers.27.mlp.gate_proj.weight": "model-00003-of-00004.safetensors",
|
| 757 |
+
"llm.model.layers.27.mlp.up_proj.weight": "model-00003-of-00004.safetensors",
|
| 758 |
+
"llm.model.layers.27.post_attention_layernorm.weight": "model-00003-of-00004.safetensors",
|
| 759 |
+
"llm.model.layers.27.self_attn.k_proj.bias": "model-00003-of-00004.safetensors",
|
| 760 |
+
"llm.model.layers.27.self_attn.k_proj.weight": "model-00003-of-00004.safetensors",
|
| 761 |
+
"llm.model.layers.27.self_attn.o_proj.weight": "model-00003-of-00004.safetensors",
|
| 762 |
+
"llm.model.layers.27.self_attn.q_proj.bias": "model-00003-of-00004.safetensors",
|
| 763 |
+
"llm.model.layers.27.self_attn.q_proj.weight": "model-00003-of-00004.safetensors",
|
| 764 |
+
"llm.model.layers.27.self_attn.v_proj.bias": "model-00003-of-00004.safetensors",
|
| 765 |
+
"llm.model.layers.27.self_attn.v_proj.weight": "model-00003-of-00004.safetensors",
|
| 766 |
+
"llm.model.layers.3.input_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 767 |
+
"llm.model.layers.3.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
|
| 768 |
+
"llm.model.layers.3.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
|
| 769 |
+
"llm.model.layers.3.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
|
| 770 |
+
"llm.model.layers.3.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 771 |
+
"llm.model.layers.3.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
|
| 772 |
+
"llm.model.layers.3.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
|
| 773 |
+
"llm.model.layers.3.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
|
| 774 |
+
"llm.model.layers.3.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
|
| 775 |
+
"llm.model.layers.3.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
|
| 776 |
+
"llm.model.layers.3.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
|
| 777 |
+
"llm.model.layers.3.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
|
| 778 |
+
"llm.model.layers.4.input_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 779 |
+
"llm.model.layers.4.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
|
| 780 |
+
"llm.model.layers.4.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
|
| 781 |
+
"llm.model.layers.4.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
|
| 782 |
+
"llm.model.layers.4.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 783 |
+
"llm.model.layers.4.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
|
| 784 |
+
"llm.model.layers.4.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
|
| 785 |
+
"llm.model.layers.4.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
|
| 786 |
+
"llm.model.layers.4.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
|
| 787 |
+
"llm.model.layers.4.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
|
| 788 |
+
"llm.model.layers.4.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
|
| 789 |
+
"llm.model.layers.4.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
|
| 790 |
+
"llm.model.layers.5.input_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 791 |
+
"llm.model.layers.5.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
|
| 792 |
+
"llm.model.layers.5.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
|
| 793 |
+
"llm.model.layers.5.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
|
| 794 |
+
"llm.model.layers.5.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 795 |
+
"llm.model.layers.5.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
|
| 796 |
+
"llm.model.layers.5.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
|
| 797 |
+
"llm.model.layers.5.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
|
| 798 |
+
"llm.model.layers.5.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
|
| 799 |
+
"llm.model.layers.5.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
|
| 800 |
+
"llm.model.layers.5.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
|
| 801 |
+
"llm.model.layers.5.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
|
| 802 |
+
"llm.model.layers.6.input_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 803 |
+
"llm.model.layers.6.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
|
| 804 |
+
"llm.model.layers.6.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
|
| 805 |
+
"llm.model.layers.6.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
|
| 806 |
+
"llm.model.layers.6.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 807 |
+
"llm.model.layers.6.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
|
| 808 |
+
"llm.model.layers.6.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
|
| 809 |
+
"llm.model.layers.6.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
|
| 810 |
+
"llm.model.layers.6.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
|
| 811 |
+
"llm.model.layers.6.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
|
| 812 |
+
"llm.model.layers.6.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
|
| 813 |
+
"llm.model.layers.6.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
|
| 814 |
+
"llm.model.layers.7.input_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 815 |
+
"llm.model.layers.7.mlp.down_proj.weight": "model-00001-of-00004.safetensors",
|
| 816 |
+
"llm.model.layers.7.mlp.gate_proj.weight": "model-00001-of-00004.safetensors",
|
| 817 |
+
"llm.model.layers.7.mlp.up_proj.weight": "model-00001-of-00004.safetensors",
|
| 818 |
+
"llm.model.layers.7.post_attention_layernorm.weight": "model-00001-of-00004.safetensors",
|
| 819 |
+
"llm.model.layers.7.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
|
| 820 |
+
"llm.model.layers.7.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
|
| 821 |
+
"llm.model.layers.7.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
|
| 822 |
+
"llm.model.layers.7.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
|
| 823 |
+
"llm.model.layers.7.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
|
| 824 |
+
"llm.model.layers.7.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
|
| 825 |
+
"llm.model.layers.7.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
|
| 826 |
+
"llm.model.layers.8.input_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 827 |
+
"llm.model.layers.8.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
|
| 828 |
+
"llm.model.layers.8.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
|
| 829 |
+
"llm.model.layers.8.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
|
| 830 |
+
"llm.model.layers.8.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 831 |
+
"llm.model.layers.8.self_attn.k_proj.bias": "model-00001-of-00004.safetensors",
|
| 832 |
+
"llm.model.layers.8.self_attn.k_proj.weight": "model-00001-of-00004.safetensors",
|
| 833 |
+
"llm.model.layers.8.self_attn.o_proj.weight": "model-00001-of-00004.safetensors",
|
| 834 |
+
"llm.model.layers.8.self_attn.q_proj.bias": "model-00001-of-00004.safetensors",
|
| 835 |
+
"llm.model.layers.8.self_attn.q_proj.weight": "model-00001-of-00004.safetensors",
|
| 836 |
+
"llm.model.layers.8.self_attn.v_proj.bias": "model-00001-of-00004.safetensors",
|
| 837 |
+
"llm.model.layers.8.self_attn.v_proj.weight": "model-00001-of-00004.safetensors",
|
| 838 |
+
"llm.model.layers.9.input_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 839 |
+
"llm.model.layers.9.mlp.down_proj.weight": "model-00002-of-00004.safetensors",
|
| 840 |
+
"llm.model.layers.9.mlp.gate_proj.weight": "model-00002-of-00004.safetensors",
|
| 841 |
+
"llm.model.layers.9.mlp.up_proj.weight": "model-00002-of-00004.safetensors",
|
| 842 |
+
"llm.model.layers.9.post_attention_layernorm.weight": "model-00002-of-00004.safetensors",
|
| 843 |
+
"llm.model.layers.9.self_attn.k_proj.bias": "model-00002-of-00004.safetensors",
|
| 844 |
+
"llm.model.layers.9.self_attn.k_proj.weight": "model-00002-of-00004.safetensors",
|
| 845 |
+
"llm.model.layers.9.self_attn.o_proj.weight": "model-00002-of-00004.safetensors",
|
| 846 |
+
"llm.model.layers.9.self_attn.q_proj.bias": "model-00002-of-00004.safetensors",
|
| 847 |
+
"llm.model.layers.9.self_attn.q_proj.weight": "model-00002-of-00004.safetensors",
|
| 848 |
+
"llm.model.layers.9.self_attn.v_proj.bias": "model-00002-of-00004.safetensors",
|
| 849 |
+
"llm.model.layers.9.self_attn.v_proj.weight": "model-00002-of-00004.safetensors",
|
| 850 |
+
"llm.model.norm.weight": "model-00003-of-00004.safetensors"
|
| 851 |
+
}
|
| 852 |
+
}
|
modeling_covo_audio.py
ADDED
|
@@ -0,0 +1,406 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import torch
|
| 2 |
+
import torchaudio
|
| 3 |
+
from transformers.modeling_utils import PreTrainedModel
|
| 4 |
+
from transformers.generation.utils import GenerationMixin
|
| 5 |
+
from transformers.models.qwen2 import Qwen2ForCausalLM
|
| 6 |
+
from transformers.models.whisper.modeling_whisper import WhisperEncoder
|
| 7 |
+
from transformers.generation.logits_process import LogitsProcessor, LogitsProcessorList
|
| 8 |
+
|
| 9 |
+
from .configuration_covo_audio import CovoAudioConfig
|
| 10 |
+
|
| 11 |
+
from torch import nn
|
| 12 |
+
import numpy as np
|
| 13 |
+
import torch.nn.functional as F
|
| 14 |
+
import os
|
| 15 |
+
from functools import lru_cache
|
| 16 |
+
from typing import Optional, Union
|
| 17 |
+
|
| 18 |
+
|
| 19 |
+
@torch.no_grad()
|
| 20 |
+
def get_dialog_prompt(audio, tokenizer, device, first_round=True):
|
| 21 |
+
begofcAUDIO_id, cAUDIO_id, endofcAUDIO_id = tokenizer.convert_tokens_to_ids(["<|begofcAUDIO|>", "<|cAUDIO|>", "<|endofcAUDIO|>"])
|
| 22 |
+
|
| 23 |
+
wav, sr = torchaudio.load(audio)
|
| 24 |
+
if wav.shape[0] == 2: # stereo to mono
|
| 25 |
+
wav = wav.mean(dim=0, keepdim=True)
|
| 26 |
+
wav = wav.squeeze(0)
|
| 27 |
+
|
| 28 |
+
# hyperparameters
|
| 29 |
+
sample_rate = 24000
|
| 30 |
+
pad_multiple = True
|
| 31 |
+
multiple_of = 480
|
| 32 |
+
if sr != sample_rate:
|
| 33 |
+
wav = torchaudio.functional.resample(wav, orig_freq=sr, new_freq=sample_rate)
|
| 34 |
+
|
| 35 |
+
hop_size = sample_rate // 100
|
| 36 |
+
wav = wav[: len(wav) // hop_size * hop_size]
|
| 37 |
+
|
| 38 |
+
# pad wav
|
| 39 |
+
if pad_multiple and multiple_of is not None:
|
| 40 |
+
d = (wav.shape[0] + multiple_of - 1) // multiple_of * multiple_of - wav.shape[0]
|
| 41 |
+
if d > 0:
|
| 42 |
+
wav = F.pad(wav, (0, d), value=0)
|
| 43 |
+
num_token = calc_seq_len(len(wav) * 100 // sample_rate)
|
| 44 |
+
|
| 45 |
+
|
| 46 |
+
# first round dialog
|
| 47 |
+
if first_round:
|
| 48 |
+
sys_prompt = """你是"小腾"，英文名是"Covo"，由腾讯开发的AI助手。
|
| 49 |
+
1、请使用简洁、口语化的语言和用户聊天，你的态度积极、耐心，像一位值得信赖的朋友。
|
| 50 |
+
2、不要使用列表或编号，避免输出网址、表情符号和复杂的公式。
|
| 51 |
+
3、不评价竞争对手，不发表主观政治观点，针对色情类、政治类、恐怖类、歧视类、暴力类的用户问题，你要妥善应对潜在的安全风险，并给出幽默，情绪安抚以及安全的劝导。
|
| 52 |
+
请用文本和音频进行对话，交替生成5个文本token和15个音频token，音频部分使用发音人：default_female"""
|
| 53 |
+
interleave_text = "<|begofcAUDIO|>" + "<|cAUDIO|>" * num_token + "<|endofcAUDIO|>"
|
| 54 |
+
|
| 55 |
+
sys_prompt = "<|im_start|>system\n" + sys_prompt + "<|im_end|>\n"
|
| 56 |
+
prompt = sys_prompt + "<|im_start|>user\n" + interleave_text + "<|im_end|>\n<|im_start|>assistant\n"
|
| 57 |
+
# multi-round dialog
|
| 58 |
+
else:
|
| 59 |
+
interleave_text = "<|begofcAUDIO|>" + "<|cAUDIO|>" * num_token + "<|endofcAUDIO|>"
|
| 60 |
+
prompt = "\n<|im_start|>user\n" + interleave_text + "<|im_end|>\n<|im_start|>assistant\n"
|
| 61 |
+
|
| 62 |
+
|
| 63 |
+
text_inputs = tokenizer(prompt, padding=True, return_tensors="pt").to(device)
|
| 64 |
+
input_ids = text_inputs.input_ids
|
| 65 |
+
attention_mask = text_inputs.attention_mask
|
| 66 |
+
|
| 67 |
+
wav = wav.to(device)
|
| 68 |
+
# long audio (>30s) processing support
|
| 69 |
+
segment_length = 720000 # 30s * 24000Hz
|
| 70 |
+
# calculate total number of segments
|
| 71 |
+
total_length = wav.shape[0]
|
| 72 |
+
num_segments = (total_length + segment_length - 1) // segment_length
|
| 73 |
+
wav_segments = []
|
| 74 |
+
# split into 30s segments and collect
|
| 75 |
+
for i in range(num_segments):
|
| 76 |
+
start_idx = i * segment_length
|
| 77 |
+
end_idx = min((i + 1) * segment_length, total_length)
|
| 78 |
+
# extract current segment
|
| 79 |
+
segment = wav[start_idx:end_idx]
|
| 80 |
+
wav_segments.append(segment)
|
| 81 |
+
return wav_segments, input_ids, attention_mask
|
| 82 |
+
|
| 83 |
+
|
| 84 |
+
def sequence_mask(lengths, max_len=None, dtype=torch.bool):
|
| 85 |
+
if max_len is None:
|
| 86 |
+
max_len = lengths.max()
|
| 87 |
+
#mask = ~(torch.ones((len(lengths), max_len)).to(lengths.device).cumsum(dim=1).t() > lengths).t()
|
| 88 |
+
mask = ~(torch.ones((len(lengths), max_len)).to(lengths.device).cumsum(dim=1) > lengths.unsqueeze(1))
|
| 89 |
+
mask = mask.to(dtype)
|
| 90 |
+
return mask
|
| 91 |
+
|
| 92 |
+
|
| 93 |
+
def calc_seq_len(seq_len):
|
| 94 |
+
strides = [2, 2, 2, 2]
|
| 95 |
+
for s in strides:
|
| 96 |
+
seq_len = (seq_len + s - 1) // s
|
| 97 |
+
return seq_len
|
| 98 |
+
|
| 99 |
+
|
| 100 |
+
class DownsampleLayer(nn.Module):
|
| 101 |
+
"""
|
| 102 |
+
Downsample layer with 1D convolution and linear layers.
|
| 103 |
+
"""
|
| 104 |
+
def __init__(self, input_dim, output_dim, hidden_dim=2048):
|
| 105 |
+
super().__init__()
|
| 106 |
+
self.conv1d = nn.Conv1d(in_channels=input_dim, out_channels=input_dim, kernel_size=3, stride=2, padding=1)
|
| 107 |
+
self.linear1 = nn.Linear(input_dim, hidden_dim)
|
| 108 |
+
self.relu1 = nn.ReLU()
|
| 109 |
+
self.linear2 = nn.Linear(hidden_dim, output_dim)
|
| 110 |
+
self.relu2 = nn.ReLU()
|
| 111 |
+
|
| 112 |
+
def forward(self, x):
|
| 113 |
+
# x: (B, T, C)
|
| 114 |
+
x = x.transpose(1, 2) # -> (B, C, T)
|
| 115 |
+
x = self.conv1d(x) # -> (B, C, T // 2)
|
| 116 |
+
x = x.transpose(1, 2) # -> (B, T // 2, C)
|
| 117 |
+
x = self.relu1(x)
|
| 118 |
+
x = self.linear1(x) # -> (B, T // 2, hidden_dim)
|
| 119 |
+
x = self.relu2(x)
|
| 120 |
+
x = self.linear2(x) # -> (B, T // 2, output_dim)
|
| 121 |
+
return x
|
| 122 |
+
|
| 123 |
+
|
| 124 |
+
class AudioAdapter(nn.Module):
|
| 125 |
+
"""
|
| 126 |
+
Audio adapter with downsample layers.
|
| 127 |
+
"""
|
| 128 |
+
def __init__(self, input_dim, output_dim, downsample=8):
|
| 129 |
+
"""
|
| 130 |
+
Args:
|
| 131 |
+
input_dim (int): input feature dimension (number of channels)
|
| 132 |
+
output_dim (int): output feature dimension
|
| 133 |
+
downsample (int): total downsampling factor, must be a power of 2
|
| 134 |
+
"""
|
| 135 |
+
super(AudioAdapter, self).__init__()
|
| 136 |
+
assert downsample % 2 == 0 and downsample >= 2, "downsample must be even"
|
| 137 |
+
|
| 138 |
+
num_layers = downsample.bit_length() - 1 # calculate how many downsampling steps are needed to reach the target factor
|
| 139 |
+
|
| 140 |
+
layers = []
|
| 141 |
+
in_dim = input_dim
|
| 142 |
+
for i in range(num_layers):
|
| 143 |
+
is_last = (i == num_layers - 1)
|
| 144 |
+
out_dim = output_dim if is_last else input_dim
|
| 145 |
+
layers.append(DownsampleLayer(in_dim, out_dim))
|
| 146 |
+
in_dim = out_dim
|
| 147 |
+
|
| 148 |
+
self.downsample_layers = nn.ModuleList(layers)
|
| 149 |
+
|
| 150 |
+
def forward(self, x):
|
| 151 |
+
"""
|
| 152 |
+
Args:
|
| 153 |
+
x (Tensor): shape (B, T, C),C=input_dim
|
| 154 |
+
Returns:
|
| 155 |
+
Tensor: shape (B, T // downsample, output_dim)
|
| 156 |
+
"""
|
| 157 |
+
for layer in self.downsample_layers:
|
| 158 |
+
x = layer(x)
|
| 159 |
+
return x
|
| 160 |
+
|
| 161 |
+
|
| 162 |
+
# from openai-whisper
|
| 163 |
+
# hard-coded audio hyperparameters
|
| 164 |
+
SAMPLE_RATE = 16000
|
| 165 |
+
N_FFT = 400
|
| 166 |
+
HOP_LENGTH = 160
|
| 167 |
+
CHUNK_LENGTH = 30
|
| 168 |
+
N_SAMPLES = CHUNK_LENGTH * SAMPLE_RATE # 480000 samples in a 30-second chunk
|
| 169 |
+
|
| 170 |
+
|
| 171 |
+
def pad_or_trim(array, length: int = N_SAMPLES, *, axis: int = -1):
|
| 172 |
+
"""
|
| 173 |
+
Pad or trim the audio array to N_SAMPLES, as expected by the encoder.
|
| 174 |
+
"""
|
| 175 |
+
if torch.is_tensor(array):
|
| 176 |
+
if array.shape[axis] > length:
|
| 177 |
+
array = array.index_select(
|
| 178 |
+
dim=axis, index=torch.arange(length, device=array.device)
|
| 179 |
+
)
|
| 180 |
+
|
| 181 |
+
if array.shape[axis] < length:
|
| 182 |
+
pad_widths = [(0, 0)] * array.ndim
|
| 183 |
+
pad_widths[axis] = (0, length - array.shape[axis])
|
| 184 |
+
array = F.pad(array, [pad for sizes in pad_widths[::-1] for pad in sizes])
|
| 185 |
+
else:
|
| 186 |
+
if array.shape[axis] > length:
|
| 187 |
+
array = array.take(indices=range(length), axis=axis)
|
| 188 |
+
|
| 189 |
+
if array.shape[axis] < length:
|
| 190 |
+
pad_widths = [(0, 0)] * array.ndim
|
| 191 |
+
pad_widths[axis] = (0, length - array.shape[axis])
|
| 192 |
+
array = np.pad(array, pad_widths)
|
| 193 |
+
|
| 194 |
+
return array
|
| 195 |
+
|
| 196 |
+
|
| 197 |
+
@lru_cache(maxsize=None)
|
| 198 |
+
def mel_filters(device, n_mels: int) -> torch.Tensor:
|
| 199 |
+
"""
|
| 200 |
+
load the mel filterbank matrix for projecting STFT into a Mel spectrogram.
|
| 201 |
+
Allows decoupling librosa dependency; saved using:
|
| 202 |
+
|
| 203 |
+
np.savez_compressed(
|
| 204 |
+
"mel_filters.npz",
|
| 205 |
+
mel_80=librosa.filters.mel(sr=16000, n_fft=400, n_mels=80),
|
| 206 |
+
mel_128=librosa.filters.mel(sr=16000, n_fft=400, n_mels=128),
|
| 207 |
+
)
|
| 208 |
+
"""
|
| 209 |
+
assert n_mels in {80, 128}, f"Unsupported n_mels: {n_mels}"
|
| 210 |
+
|
| 211 |
+
filters_path = os.path.join(os.path.dirname(__file__), "assets", "mel_filters.npz")
|
| 212 |
+
with np.load(filters_path, allow_pickle=False) as f:
|
| 213 |
+
return torch.from_numpy(f[f"mel_{n_mels}"]).to(device)
|
| 214 |
+
|
| 215 |
+
|
| 216 |
+
def log_mel_spectrogram(
|
| 217 |
+
audio: torch.Tensor,
|
| 218 |
+
n_mels: int = 80,
|
| 219 |
+
padding: int = 0,
|
| 220 |
+
device: Optional[Union[str, torch.device]] = None,
|
| 221 |
+
):
|
| 222 |
+
"""
|
| 223 |
+
Compute the log-Mel spectrogram
|
| 224 |
+
|
| 225 |
+
Parameters
|
| 226 |
+
----------
|
| 227 |
+
audio: Union[str, np.ndarray, torch.Tensor], shape = (*)
|
| 228 |
+
The path to audio or either a NumPy array or Tensor containing the audio waveform in 16 kHz
|
| 229 |
+
|
| 230 |
+
n_mels: int
|
| 231 |
+
The number of Mel-frequency filters, only 80 is supported
|
| 232 |
+
|
| 233 |
+
padding: int
|
| 234 |
+
Number of zero samples to pad to the right
|
| 235 |
+
|
| 236 |
+
device: Optional[Union[str, torch.device]]
|
| 237 |
+
If given, the audio tensor is moved to this device before STFT
|
| 238 |
+
|
| 239 |
+
Returns
|
| 240 |
+
-------
|
| 241 |
+
torch.Tensor, shape = (80, n_frames)
|
| 242 |
+
A Tensor that contains the Mel spectrogram
|
| 243 |
+
"""
|
| 244 |
+
dtype = audio.dtype
|
| 245 |
+
if device is not None:
|
| 246 |
+
audio = audio.to(device)
|
| 247 |
+
if padding > 0:
|
| 248 |
+
audio = F.pad(audio, (0, padding))
|
| 249 |
+
window = torch.hann_window(N_FFT).to(audio.device)
|
| 250 |
+
stft = torch.stft(audio, N_FFT, HOP_LENGTH, window=window, return_complex=True)
|
| 251 |
+
magnitudes = stft[..., :-1].abs() ** 2
|
| 252 |
+
|
| 253 |
+
filters = mel_filters(audio.device, n_mels)
|
| 254 |
+
mel_spec = filters @ magnitudes
|
| 255 |
+
|
| 256 |
+
log_spec = torch.clamp(mel_spec, min=1e-10).log10()
|
| 257 |
+
log_spec = torch.maximum(log_spec, log_spec.max() - 8.0)
|
| 258 |
+
log_spec = (log_spec + 4.0) / 4.0
|
| 259 |
+
log_spec = log_spec.to(dtype)
|
| 260 |
+
return log_spec
|
| 261 |
+
|
| 262 |
+
|
| 263 |
+
class WindowedRepetitionPenaltyLogitsProcessor(LogitsProcessor):
|
| 264 |
+
def __init__(self, penalty: float, window_size: int):
|
| 265 |
+
self.penalty = penalty
|
| 266 |
+
self.window_size = window_size
|
| 267 |
+
|
| 268 |
+
def __call__(self, input_ids, scores):
|
| 269 |
+
for batch_idx, input_seq in enumerate(input_ids):
|
| 270 |
+
window = input_seq[-self.window_size:] if self.window_size > 0 else input_seq # get last 'window_size' tokens
|
| 271 |
+
for token_id in set(window.tolist()):
|
| 272 |
+
if scores[batch_idx, token_id] < 0:
|
| 273 |
+
scores[batch_idx, token_id] *= self.penalty
|
| 274 |
+
else:
|
| 275 |
+
scores[batch_idx, token_id] /= self.penalty
|
| 276 |
+
return scores
|
| 277 |
+
|
| 278 |
+
|
| 279 |
+
class CovoAudioForCausalLM(PreTrainedModel, GenerationMixin):
|
| 280 |
+
config_class = CovoAudioConfig
|
| 281 |
+
|
| 282 |
+
def __init__(self, config: CovoAudioConfig, **kwargs):
|
| 283 |
+
super().__init__(config, **kwargs)
|
| 284 |
+
self.llm = Qwen2ForCausalLM(config.llm_config)
|
| 285 |
+
self.encoder = WhisperEncoder(config.encoder_config)
|
| 286 |
+
self.audio_adapter = AudioAdapter(config.whisper_feats_dim,
|
| 287 |
+
config.llm_config.hidden_size,
|
| 288 |
+
config.adapter_downsample)
|
| 289 |
+
|
| 290 |
+
self.post_init()
|
| 291 |
+
|
| 292 |
+
#NOTE Force 'tie_weights' function to do nothing to
|
| 293 |
+
# avoid the memory sharing between input and output embeddings of llm
|
| 294 |
+
def tie_weights(self, **kwargs):
|
| 295 |
+
pass
|
| 296 |
+
|
| 297 |
+
def audio_encoder(self, wavs, device):
|
| 298 |
+
"""
|
| 299 |
+
Extract features from input waveform
|
| 300 |
+
"""
|
| 301 |
+
# move resampler to the correct device
|
| 302 |
+
resampler16k = torchaudio.transforms.Resample(24000, 16000).to(device)
|
| 303 |
+
|
| 304 |
+
mel_features_list = []
|
| 305 |
+
for wav in wavs:
|
| 306 |
+
wav = resampler16k(wav)
|
| 307 |
+
audio = pad_or_trim(wav)
|
| 308 |
+
# [B, 80, 3000] 30s 100hz
|
| 309 |
+
mel_features = log_mel_spectrogram(audio, n_mels=128).to(torch.bfloat16)
|
| 310 |
+
mel_features_list.append(mel_features)
|
| 311 |
+
mel_features = torch.stack(mel_features_list)
|
| 312 |
+
|
| 313 |
+
feats = self.encoder(mel_features).last_hidden_state
|
| 314 |
+
features = self.audio_adapter(feats)
|
| 315 |
+
features = features.view(1, -1, features.shape[2])
|
| 316 |
+
|
| 317 |
+
return features
|
| 318 |
+
|
| 319 |
+
def forward(
|
| 320 |
+
self,
|
| 321 |
+
input_ids=None,
|
| 322 |
+
inputs_embeds=None,
|
| 323 |
+
wavs=None,
|
| 324 |
+
attention_mask=None,
|
| 325 |
+
past_key_values=None,
|
| 326 |
+
labels=None,
|
| 327 |
+
position_ids=None,
|
| 328 |
+
**kwargs
|
| 329 |
+
):
|
| 330 |
+
outputs = self.llm(
|
| 331 |
+
inputs_embeds=inputs_embeds,
|
| 332 |
+
attention_mask=attention_mask,
|
| 333 |
+
past_key_values=past_key_values,
|
| 334 |
+
labels=labels,
|
| 335 |
+
position_ids=position_ids,
|
| 336 |
+
**kwargs
|
| 337 |
+
)
|
| 338 |
+
return outputs
|
| 339 |
+
|
| 340 |
+
def get_input_embeddings(self):
|
| 341 |
+
"""
|
| 342 |
+
Return the model's input embeddings - required for GenerationMixin
|
| 343 |
+
"""
|
| 344 |
+
return self.llm.get_input_embeddings()
|
| 345 |
+
|
| 346 |
+
def get_output_embeddings(self):
|
| 347 |
+
"""
|
| 348 |
+
Return the model's output embeddings - required for GenerationMixin
|
| 349 |
+
"""
|
| 350 |
+
# return self.llm.get_output_embeddings()
|
| 351 |
+
return self.llm.lm_head
|
| 352 |
+
|
| 353 |
+
def prepare_inputs_for_generation(
|
| 354 |
+
self,
|
| 355 |
+
input_ids,
|
| 356 |
+
attention_mask=None,
|
| 357 |
+
**kwargs
|
| 358 |
+
):
|
| 359 |
+
wavs = kwargs.get("wavs", None)
|
| 360 |
+
is_first_iteration = kwargs.get("is_first_iteration", False)
|
| 361 |
+
past_key_values = kwargs.get("past_key_values", None)
|
| 362 |
+
|
| 363 |
+
if is_first_iteration: # First generation step, include audio processing
|
| 364 |
+
inputs_embeds = self.llm.get_input_embeddings()(input_ids)
|
| 365 |
+
cAUDIO_id = 151666 # tokenizer.convert_tokens_to_ids("<|cAUDIO|>")
|
| 366 |
+
audio_features = self.audio_encoder(wavs, inputs_embeds.device)
|
| 367 |
+
feature_lengths = (input_ids == cAUDIO_id).sum(1)
|
| 368 |
+
feature_seq_mask = sequence_mask(feature_lengths, max_len=audio_features.size(1), dtype=torch.bool)
|
| 369 |
+
audio_features = audio_features.to(inputs_embeds.device, inputs_embeds.dtype)
|
| 370 |
+
audio_features = audio_features[feature_seq_mask]
|
| 371 |
+
|
| 372 |
+
audio_mask = input_ids == cAUDIO_id
|
| 373 |
+
audio_mask = audio_mask.unsqueeze(-1)
|
| 374 |
+
inputs_embeds = inputs_embeds.masked_scatter(audio_mask, audio_features)
|
| 375 |
+
|
| 376 |
+
return {
|
| 377 |
+
"inputs_embeds": inputs_embeds,
|
| 378 |
+
"attention_mask": attention_mask,
|
| 379 |
+
"past_key_values": past_key_values,
|
| 380 |
+
}
|
| 381 |
+
else: # We're in a generation step, no need to process audio again
|
| 382 |
+
input_ids = input_ids[:, -1:]
|
| 383 |
+
inputs_embeds = self.llm.get_input_embeddings()(input_ids)
|
| 384 |
+
return {
|
| 385 |
+
"inputs_embeds": inputs_embeds,
|
| 386 |
+
"attention_mask": attention_mask,
|
| 387 |
+
"past_key_values": past_key_values,
|
| 388 |
+
}
|
| 389 |
+
|
| 390 |
+
def _set_gradient_checkpointing(self, module, value=False):
|
| 391 |
+
# For Qwen2
|
| 392 |
+
if hasattr(self.llm, 'gradient_checkpointing'):
|
| 393 |
+
self.llm.gradient_checkpointing = value
|
| 394 |
+
|
| 395 |
+
# Add the missing _gradient_checkpointing_func method to Qwen2Model
|
| 396 |
+
if value and not hasattr(self.llm, '_gradient_checkpointing_func'):
|
| 397 |
+
def _gradient_checkpointing_func(module_to_run, *args, **kwargs):
|
| 398 |
+
return torch.utils.checkpoint.checkpoint(module_to_run, *args, **kwargs)
|
| 399 |
+
|
| 400 |
+
self.llm._gradient_checkpointing_func = _gradient_checkpointing_func
|
| 401 |
+
|
| 402 |
+
# For custom encoder and adapter
|
| 403 |
+
if hasattr(self.encoder, 'gradient_checkpointing'):
|
| 404 |
+
self.encoder.gradient_checkpointing = value
|
| 405 |
+
if hasattr(self.audio_adapter, 'gradient_checkpointing'):
|
| 406 |
+
self.audio_adapter.gradient_checkpointing = value
|
special_tokens_map.json
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"additional_special_tokens": [
|
| 3 |
+
"<|begofcAUDIO|>",
|
| 4 |
+
"<|cAUDIO|>",
|
| 5 |
+
"<|endofcAUDIO|>",
|
| 6 |
+
"<|begofdAUDIO|>",
|
| 7 |
+
"<|dAUDIO|>",
|
| 8 |
+
"<|endofdAUDIO|>"
|
| 9 |
+
],
|
| 10 |
+
"eos_token": {
|
| 11 |
+
"content": "<|endoftext|>",
|
| 12 |
+
"lstrip": false,
|
| 13 |
+
"normalized": false,
|
| 14 |
+
"rstrip": false,
|
| 15 |
+
"single_word": false
|
| 16 |
+
},
|
| 17 |
+
"pad_token": {
|
| 18 |
+
"content": "<|endoftext|>",
|
| 19 |
+
"lstrip": false,
|
| 20 |
+
"normalized": false,
|
| 21 |
+
"rstrip": false,
|
| 22 |
+
"single_word": false
|
| 23 |
+
}
|
| 24 |
+
}
|
token2wav/global_mean_var.npy
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:91968d1064eff334fd218e1ba6764e66c760f2af4ac28ea453265d740f7bcae5
|
| 3 |
+
size 640
|
token2wav/model.pt
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:92f98f7be0a24913b48d58ef7dbd85c3a306727a929a7e9c912140433f5f7fcc
|
| 3 |
+
size 1861521759
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:25daf65a8df9d4f2c3d73bf6a6930a5cacd45ea66572fbf8a7fd5b46927fce72
|
| 3 |
+
size 11423038
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,248 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 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": "<|begofcAUDIO|>",
|
| 183 |
+
"lstrip": false,
|
| 184 |
+
"normalized": false,
|
| 185 |
+
"rstrip": false,
|
| 186 |
+
"single_word": false,
|
| 187 |
+
"special": true
|
| 188 |
+
},
|
| 189 |
+
"151666": {
|
| 190 |
+
"content": "<|cAUDIO|>",
|
| 191 |
+
"lstrip": false,
|
| 192 |
+
"normalized": false,
|
| 193 |
+
"rstrip": false,
|
| 194 |
+
"single_word": false,
|
| 195 |
+
"special": true
|
| 196 |
+
},
|
| 197 |
+
"151667": {
|
| 198 |
+
"content": "<|endofcAUDIO|>",
|
| 199 |
+
"lstrip": false,
|
| 200 |
+
"normalized": false,
|
| 201 |
+
"rstrip": false,
|
| 202 |
+
"single_word": false,
|
| 203 |
+
"special": true
|
| 204 |
+
},
|
| 205 |
+
"151668": {
|
| 206 |
+
"content": "<|begofdAUDIO|>",
|
| 207 |
+
"lstrip": false,
|
| 208 |
+
"normalized": false,
|
| 209 |
+
"rstrip": false,
|
| 210 |
+
"single_word": false,
|
| 211 |
+
"special": true
|
| 212 |
+
},
|
| 213 |
+
"151669": {
|
| 214 |
+
"content": "<|dAUDIO|>",
|
| 215 |
+
"lstrip": false,
|
| 216 |
+
"normalized": false,
|
| 217 |
+
"rstrip": false,
|
| 218 |
+
"single_word": false,
|
| 219 |
+
"special": true
|
| 220 |
+
},
|
| 221 |
+
"151670": {
|
| 222 |
+
"content": "<|endofdAUDIO|>",
|
| 223 |
+
"lstrip": false,
|
| 224 |
+
"normalized": false,
|
| 225 |
+
"rstrip": false,
|
| 226 |
+
"single_word": false,
|
| 227 |
+
"special": true
|
| 228 |
+
}
|
| 229 |
+
},
|
| 230 |
+
"additional_special_tokens": [
|
| 231 |
+
"<|begofcAUDIO|>",
|
| 232 |
+
"<|cAUDIO|>",
|
| 233 |
+
"<|endofcAUDIO|>",
|
| 234 |
+
"<|begofdAUDIO|>",
|
| 235 |
+
"<|dAUDIO|>",
|
| 236 |
+
"<|endofdAUDIO|>"
|
| 237 |
+
],
|
| 238 |
+
"bos_token": null,
|
| 239 |
+
"clean_up_tokenization_spaces": false,
|
| 240 |
+
"eos_token": "<|endoftext|>",
|
| 241 |
+
"errors": "replace",
|
| 242 |
+
"extra_special_tokens": {},
|
| 243 |
+
"model_max_length": 131072,
|
| 244 |
+
"pad_token": "<|endoftext|>",
|
| 245 |
+
"split_special_tokens": false,
|
| 246 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 247 |
+
"unk_token": null
|
| 248 |
+
}
|
vocab.json
ADDED
|
The diff for this file is too large to render.
See raw diff
|
|
|