Upload quantized model DeepSeek-R1-Distill-Qwen-1.5B-AutoRound-NVFP4-Tuning
Browse files- .gitattributes +1 -0
- README.md +178 -0
- chat_template.jinja +1 -0
- config.json +758 -0
- generation_config.json +9 -0
- model.safetensors +3 -0
- quantization_config.json +696 -0
- tokenizer.json +3 -0
- tokenizer_config.json +15 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,4 @@ 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 |
+
tokenizer.json filter=lfs diff=lfs merge=lfs -text
|
README.md
ADDED
|
@@ -0,0 +1,178 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
base_model:
|
| 3 |
+
- deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B
|
| 4 |
+
pipeline_tag: text-generation
|
| 5 |
+
tags:
|
| 6 |
+
- quantized
|
| 7 |
+
- nvfp4
|
| 8 |
+
- tuning
|
| 9 |
+
- low-bit-open-llm-leaderboard
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
+
# DeepSeek-R1-Distill-Qwen-1.5B-AutoRound-NVFP4-Tuning
|
| 13 |
+
|
| 14 |
+
## Model Details
|
| 15 |
+
|
| 16 |
+
This model is a NVFP4 (NVIDIA FP4) quantization of [deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B) generated by TUNING. Please follow the license of the original model.
|
| 17 |
+
|
| 18 |
+
## Quantization Details
|
| 19 |
+
|
| 20 |
+
| Attribute | Value |
|
| 21 |
+
|-----------|-------|
|
| 22 |
+
| Base Model | [deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B](https://huggingface.co/deepseek-ai/DeepSeek-R1-Distill-Qwen-1.5B) |
|
| 23 |
+
| Quantization Tool | TUNING |
|
| 24 |
+
| Quantization Scheme | NVFP4 |
|
| 25 |
+
| Quantized Size | 1805 MB |
|
| 26 |
+
|
| 27 |
+
## Evaluation Results
|
| 28 |
+
|
| 29 |
+
| Task | Accuracy |
|
| 30 |
+
|------|----------|
|
| 31 |
+
| hellaswag | 0.3582 |
|
| 32 |
+
| mmlu | 0.3565 |
|
| 33 |
+
| mmlu_abstract_algebra | 0.3100 |
|
| 34 |
+
| mmlu_anatomy | 0.3481 |
|
| 35 |
+
| mmlu_astronomy | 0.3421 |
|
| 36 |
+
| mmlu_business_ethics | 0.4000 |
|
| 37 |
+
| mmlu_clinical_knowledge | 0.3358 |
|
| 38 |
+
| mmlu_college_biology | 0.3264 |
|
| 39 |
+
| mmlu_college_chemistry | 0.3500 |
|
| 40 |
+
| mmlu_college_computer_science | 0.3900 |
|
| 41 |
+
| mmlu_college_mathematics | 0.3800 |
|
| 42 |
+
| mmlu_college_medicine | 0.3526 |
|
| 43 |
+
| mmlu_college_physics | 0.2745 |
|
| 44 |
+
| mmlu_computer_security | 0.3400 |
|
| 45 |
+
| mmlu_conceptual_physics | 0.4298 |
|
| 46 |
+
| mmlu_econometrics | 0.2719 |
|
| 47 |
+
| mmlu_electrical_engineering | 0.4069 |
|
| 48 |
+
| mmlu_elementary_mathematics | 0.4392 |
|
| 49 |
+
| mmlu_formal_logic | 0.3968 |
|
| 50 |
+
| mmlu_global_facts | 0.3400 |
|
| 51 |
+
| mmlu_high_school_biology | 0.4032 |
|
| 52 |
+
| mmlu_high_school_chemistry | 0.3744 |
|
| 53 |
+
| mmlu_high_school_computer_science | 0.4400 |
|
| 54 |
+
| mmlu_high_school_european_history | 0.3576 |
|
| 55 |
+
| mmlu_high_school_geography | 0.3586 |
|
| 56 |
+
| mmlu_high_school_government_and_politics | 0.3316 |
|
| 57 |
+
| mmlu_high_school_macroeconomics | 0.3590 |
|
| 58 |
+
| mmlu_high_school_mathematics | 0.3000 |
|
| 59 |
+
| mmlu_high_school_microeconomics | 0.4412 |
|
| 60 |
+
| mmlu_high_school_physics | 0.2649 |
|
| 61 |
+
| mmlu_high_school_psychology | 0.4330 |
|
| 62 |
+
| mmlu_high_school_statistics | 0.3657 |
|
| 63 |
+
| mmlu_high_school_us_history | 0.3039 |
|
| 64 |
+
| mmlu_high_school_world_history | 0.3713 |
|
| 65 |
+
| mmlu_human_aging | 0.4126 |
|
| 66 |
+
| mmlu_human_sexuality | 0.4198 |
|
| 67 |
+
| mmlu_humanities | 0.3105 |
|
| 68 |
+
| mmlu_international_law | 0.4298 |
|
| 69 |
+
| mmlu_jurisprudence | 0.4630 |
|
| 70 |
+
| mmlu_logical_fallacies | 0.3926 |
|
| 71 |
+
| mmlu_machine_learning | 0.2143 |
|
| 72 |
+
| mmlu_management | 0.4563 |
|
| 73 |
+
| mmlu_marketing | 0.5897 |
|
| 74 |
+
| mmlu_medical_genetics | 0.4000 |
|
| 75 |
+
| mmlu_miscellaneous | 0.4125 |
|
| 76 |
+
| mmlu_moral_disputes | 0.3671 |
|
| 77 |
+
| mmlu_moral_scenarios | 0.2324 |
|
| 78 |
+
| mmlu_nutrition | 0.4150 |
|
| 79 |
+
| mmlu_other | 0.3911 |
|
| 80 |
+
| mmlu_philosophy | 0.3923 |
|
| 81 |
+
| mmlu_prehistory | 0.3488 |
|
| 82 |
+
| mmlu_professional_accounting | 0.2589 |
|
| 83 |
+
| mmlu_professional_law | 0.2725 |
|
| 84 |
+
| mmlu_professional_medicine | 0.2978 |
|
| 85 |
+
| mmlu_professional_psychology | 0.3203 |
|
| 86 |
+
| mmlu_public_relations | 0.4455 |
|
| 87 |
+
| mmlu_security_studies | 0.3878 |
|
| 88 |
+
| mmlu_social_sciences | 0.3848 |
|
| 89 |
+
| mmlu_sociology | 0.4677 |
|
| 90 |
+
| mmlu_stem | 0.3635 |
|
| 91 |
+
| mmlu_us_foreign_policy | 0.4800 |
|
| 92 |
+
| mmlu_virology | 0.4217 |
|
| 93 |
+
| mmlu_world_religions | 0.2807 |
|
| 94 |
+
| piqa | 0.6485 |
|
| 95 |
+
|
| 96 |
+
## How to Use
|
| 97 |
+
|
| 98 |
+
### HF Usage
|
| 99 |
+
|
| 100 |
+
**Step 1: Install [AutoRound](https://github.com/intel/auto-round)**
|
| 101 |
+
|
| 102 |
+
```bash
|
| 103 |
+
pip install auto-round
|
| 104 |
+
```
|
| 105 |
+
|
| 106 |
+
**Step 2: Load and run the quantized model**
|
| 107 |
+
|
| 108 |
+
```python
|
| 109 |
+
from transformers import AutoModelForCausalLM, AutoTokenizer
|
| 110 |
+
|
| 111 |
+
model_name = "DeepSeek-R1-Distill-Qwen-1.5B-AutoRound-NVFP4-Tuning"
|
| 112 |
+
|
| 113 |
+
# load the tokenizer and the model
|
| 114 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name)
|
| 115 |
+
model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype="auto", device_map="auto")
|
| 116 |
+
|
| 117 |
+
# prepare the model input
|
| 118 |
+
prompt = "Write a quick sort algorithm."
|
| 119 |
+
messages = [{"role": "user", "content": prompt}]
|
| 120 |
+
text = tokenizer.apply_chat_template(
|
| 121 |
+
messages,
|
| 122 |
+
tokenize=False,
|
| 123 |
+
add_generation_prompt=True,
|
| 124 |
+
)
|
| 125 |
+
model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
|
| 126 |
+
|
| 127 |
+
# conduct text completion
|
| 128 |
+
generated_ids = model.generate(**model_inputs, max_new_tokens=512)
|
| 129 |
+
output_ids = generated_ids[0][len(model_inputs.input_ids[0]) :].tolist()
|
| 130 |
+
|
| 131 |
+
content = tokenizer.decode(output_ids, skip_special_tokens=True)
|
| 132 |
+
print("content:", content)
|
| 133 |
+
```
|
| 134 |
+
|
| 135 |
+
### VLLM Usage
|
| 136 |
+
|
| 137 |
+
```bash
|
| 138 |
+
vllm serve DeepSeek-R1-Distill-Qwen-1.5B-AutoRound-NVFP4-Tuning \
|
| 139 |
+
--trust-remote-code \
|
| 140 |
+
--dtype bfloat16 \
|
| 141 |
+
--tensor_parallel_size 1
|
| 142 |
+
```
|
| 143 |
+
|
| 144 |
+
If you encounter any issues, feel free to open an issue on the [AutoRound GitHub repo](https://github.com/intel/auto-round/issues) or provide feedback on the [Low-Bit Open LLM Leaderboard](https://huggingface.co/spaces/Intel/low_bit_open_llm_leaderboard).
|
| 145 |
+
|
| 146 |
+
## Ethical Considerations and Limitations
|
| 147 |
+
|
| 148 |
+
The model can produce factually incorrect output, and should not be relied on to produce factually accurate information. Because of the limitations of the pretrained model and the finetuning datasets, it is possible that this model could generate lewd, biased or otherwise offensive outputs.
|
| 149 |
+
Therefore, before deploying any applications of the model, developers should perform safety testing.
|
| 150 |
+
|
| 151 |
+
## Caveats and Recommendations
|
| 152 |
+
|
| 153 |
+
Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model.
|
| 154 |
+
Here are a couple of useful links to learn more about Intel's AI software:
|
| 155 |
+
|
| 156 |
+
- [Intel Neural Compressor](https://github.com/intel/neural-compressor)
|
| 157 |
+
- [AutoRound](https://github.com/intel/auto-round)
|
| 158 |
+
|
| 159 |
+
## Disclaimer
|
| 160 |
+
|
| 161 |
+
The license on this model does not constitute legal advice. We are not responsible for the actions of third parties who use this model. Please consult an attorney before using this model for commercial purposes.
|
| 162 |
+
|
| 163 |
+
## Cite
|
| 164 |
+
|
| 165 |
+
```
|
| 166 |
+
@article{cheng2023optimize,
|
| 167 |
+
title={Optimize weight rounding via signed gradient descent for the quantization of llms},
|
| 168 |
+
author={Cheng, Wenhua and Zhang, Weiwei and Shen, Haihao and Cai, Yiyang and He, Xin and Lv, Kaokao and Liu, Yi},
|
| 169 |
+
journal={arXiv preprint arXiv:2309.05516},
|
| 170 |
+
year={2023}
|
| 171 |
+
}
|
| 172 |
+
```
|
| 173 |
+
|
| 174 |
+
[arxiv](https://arxiv.org/abs/2309.05516) [github](https://github.com/intel/auto-round)
|
| 175 |
+
|
| 176 |
+
---
|
| 177 |
+
|
| 178 |
+
*This model is part of the [Intel Low-Bit Open LLM Leaderboard](https://huggingface.co/spaces/Intel/low_bit_open_llm_leaderboard) initiative.*
|
chat_template.jinja
ADDED
|
@@ -0,0 +1 @@
|
|
|
|
|
|
|
| 1 |
+
{% if not add_generation_prompt is defined %}{% set add_generation_prompt = false %}{% endif %}{% set ns = namespace(is_first=false, is_tool=false, is_output_first=true, system_prompt='') %}{%- for message in messages %}{%- if message['role'] == 'system' %}{% set ns.system_prompt = message['content'] %}{%- endif %}{%- endfor %}{{bos_token}}{{ns.system_prompt}}{%- for message in messages %}{%- if message['role'] == 'user' %}{%- set ns.is_tool = false -%}{{'<｜User｜>' + message['content']}}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is none %}{%- set ns.is_tool = false -%}{%- for tool in message['tool_calls']%}{%- if not ns.is_first %}{{'<｜Assistant｜><｜tool▁calls▁begin｜><｜tool▁call▁begin｜>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\n' + '```json' + '\n' + tool['function']['arguments'] + '\n' + '```' + '<｜tool▁call▁end｜>'}}{%- set ns.is_first = true -%}{%- else %}{{'\n' + '<｜tool▁call▁begin｜>' + tool['type'] + '<｜tool▁sep｜>' + tool['function']['name'] + '\n' + '```json' + '\n' + tool['function']['arguments'] + '\n' + '```' + '<｜tool▁call▁end｜>'}}{{'<｜tool▁calls▁end｜><｜end▁of▁sentence｜>'}}{%- endif %}{%- endfor %}{%- endif %}{%- if message['role'] == 'assistant' and message['content'] is not none %}{%- if ns.is_tool %}{{'<｜tool▁outputs▁end｜>' + message['content'] + '<｜end▁of▁sentence｜>'}}{%- set ns.is_tool = false -%}{%- else %}{% set content = message['content'] %}{% if '</think>' in content %}{% set content = content.split('</think>')[-1] %}{% endif %}{{'<｜Assistant｜>' + content + '<｜end▁of▁sentence｜>'}}{%- endif %}{%- endif %}{%- if message['role'] == 'tool' %}{%- set ns.is_tool = true -%}{%- if ns.is_output_first %}{{'<｜tool▁outputs▁begin｜><｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- set ns.is_output_first = false %}{%- else %}{{'\n<｜tool▁output▁begin｜>' + message['content'] + '<｜tool▁output▁end｜>'}}{%- endif %}{%- endif %}{%- endfor -%}{% if ns.is_tool %}{{'<｜tool▁outputs▁end｜>'}}{% endif %}{% if add_generation_prompt and not ns.is_tool %}{{'<｜Assistant｜><think>\n'}}{% endif %}
|
config.json
ADDED
|
@@ -0,0 +1,758 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"architectures": [
|
| 3 |
+
"Qwen2ForCausalLM"
|
| 4 |
+
],
|
| 5 |
+
"attention_dropout": 0.0,
|
| 6 |
+
"bos_token_id": 151643,
|
| 7 |
+
"dtype": "bfloat16",
|
| 8 |
+
"eos_token_id": 151643,
|
| 9 |
+
"hidden_act": "silu",
|
| 10 |
+
"hidden_size": 1536,
|
| 11 |
+
"initializer_range": 0.02,
|
| 12 |
+
"intermediate_size": 8960,
|
| 13 |
+
"layer_types": [
|
| 14 |
+
"full_attention",
|
| 15 |
+
"full_attention",
|
| 16 |
+
"full_attention",
|
| 17 |
+
"full_attention",
|
| 18 |
+
"full_attention",
|
| 19 |
+
"full_attention",
|
| 20 |
+
"full_attention",
|
| 21 |
+
"full_attention",
|
| 22 |
+
"full_attention",
|
| 23 |
+
"full_attention",
|
| 24 |
+
"full_attention",
|
| 25 |
+
"full_attention",
|
| 26 |
+
"full_attention",
|
| 27 |
+
"full_attention",
|
| 28 |
+
"full_attention",
|
| 29 |
+
"full_attention",
|
| 30 |
+
"full_attention",
|
| 31 |
+
"full_attention",
|
| 32 |
+
"full_attention",
|
| 33 |
+
"full_attention",
|
| 34 |
+
"full_attention",
|
| 35 |
+
"full_attention",
|
| 36 |
+
"full_attention",
|
| 37 |
+
"full_attention",
|
| 38 |
+
"full_attention",
|
| 39 |
+
"full_attention",
|
| 40 |
+
"full_attention",
|
| 41 |
+
"full_attention"
|
| 42 |
+
],
|
| 43 |
+
"max_position_embeddings": 131072,
|
| 44 |
+
"max_window_layers": 21,
|
| 45 |
+
"model_type": "qwen2",
|
| 46 |
+
"num_attention_heads": 12,
|
| 47 |
+
"num_hidden_layers": 28,
|
| 48 |
+
"num_key_value_heads": 2,
|
| 49 |
+
"pad_token_id": null,
|
| 50 |
+
"quantization_config": {
|
| 51 |
+
"act_bits": 4,
|
| 52 |
+
"act_data_type": "nv_fp4_with_static_gs",
|
| 53 |
+
"act_dynamic": true,
|
| 54 |
+
"act_group_size": 16,
|
| 55 |
+
"act_sym": true,
|
| 56 |
+
"autoround_version": "0.13.1",
|
| 57 |
+
"bits": 4,
|
| 58 |
+
"block_name_to_quantize": "model.layers",
|
| 59 |
+
"data_type": "nv_fp",
|
| 60 |
+
"extra_config": {
|
| 61 |
+
".*self_attn.*": {
|
| 62 |
+
"act_bits": 16,
|
| 63 |
+
"act_data_type": "float",
|
| 64 |
+
"bits": 16,
|
| 65 |
+
"data_type": "float"
|
| 66 |
+
},
|
| 67 |
+
"model.layers.0.self_attn.k_proj": {
|
| 68 |
+
"act_bits": 16,
|
| 69 |
+
"act_data_type": "float",
|
| 70 |
+
"bits": 16,
|
| 71 |
+
"data_type": "float"
|
| 72 |
+
},
|
| 73 |
+
"model.layers.0.self_attn.o_proj": {
|
| 74 |
+
"act_bits": 16,
|
| 75 |
+
"act_data_type": "float",
|
| 76 |
+
"bits": 16,
|
| 77 |
+
"data_type": "float"
|
| 78 |
+
},
|
| 79 |
+
"model.layers.0.self_attn.q_proj": {
|
| 80 |
+
"act_bits": 16,
|
| 81 |
+
"act_data_type": "float",
|
| 82 |
+
"bits": 16,
|
| 83 |
+
"data_type": "float"
|
| 84 |
+
},
|
| 85 |
+
"model.layers.0.self_attn.v_proj": {
|
| 86 |
+
"act_bits": 16,
|
| 87 |
+
"act_data_type": "float",
|
| 88 |
+
"bits": 16,
|
| 89 |
+
"data_type": "float"
|
| 90 |
+
},
|
| 91 |
+
"model.layers.1.self_attn.k_proj": {
|
| 92 |
+
"act_bits": 16,
|
| 93 |
+
"act_data_type": "float",
|
| 94 |
+
"bits": 16,
|
| 95 |
+
"data_type": "float"
|
| 96 |
+
},
|
| 97 |
+
"model.layers.1.self_attn.o_proj": {
|
| 98 |
+
"act_bits": 16,
|
| 99 |
+
"act_data_type": "float",
|
| 100 |
+
"bits": 16,
|
| 101 |
+
"data_type": "float"
|
| 102 |
+
},
|
| 103 |
+
"model.layers.1.self_attn.q_proj": {
|
| 104 |
+
"act_bits": 16,
|
| 105 |
+
"act_data_type": "float",
|
| 106 |
+
"bits": 16,
|
| 107 |
+
"data_type": "float"
|
| 108 |
+
},
|
| 109 |
+
"model.layers.1.self_attn.v_proj": {
|
| 110 |
+
"act_bits": 16,
|
| 111 |
+
"act_data_type": "float",
|
| 112 |
+
"bits": 16,
|
| 113 |
+
"data_type": "float"
|
| 114 |
+
},
|
| 115 |
+
"model.layers.10.self_attn.k_proj": {
|
| 116 |
+
"act_bits": 16,
|
| 117 |
+
"act_data_type": "float",
|
| 118 |
+
"bits": 16,
|
| 119 |
+
"data_type": "float"
|
| 120 |
+
},
|
| 121 |
+
"model.layers.10.self_attn.o_proj": {
|
| 122 |
+
"act_bits": 16,
|
| 123 |
+
"act_data_type": "float",
|
| 124 |
+
"bits": 16,
|
| 125 |
+
"data_type": "float"
|
| 126 |
+
},
|
| 127 |
+
"model.layers.10.self_attn.q_proj": {
|
| 128 |
+
"act_bits": 16,
|
| 129 |
+
"act_data_type": "float",
|
| 130 |
+
"bits": 16,
|
| 131 |
+
"data_type": "float"
|
| 132 |
+
},
|
| 133 |
+
"model.layers.10.self_attn.v_proj": {
|
| 134 |
+
"act_bits": 16,
|
| 135 |
+
"act_data_type": "float",
|
| 136 |
+
"bits": 16,
|
| 137 |
+
"data_type": "float"
|
| 138 |
+
},
|
| 139 |
+
"model.layers.11.self_attn.k_proj": {
|
| 140 |
+
"act_bits": 16,
|
| 141 |
+
"act_data_type": "float",
|
| 142 |
+
"bits": 16,
|
| 143 |
+
"data_type": "float"
|
| 144 |
+
},
|
| 145 |
+
"model.layers.11.self_attn.o_proj": {
|
| 146 |
+
"act_bits": 16,
|
| 147 |
+
"act_data_type": "float",
|
| 148 |
+
"bits": 16,
|
| 149 |
+
"data_type": "float"
|
| 150 |
+
},
|
| 151 |
+
"model.layers.11.self_attn.q_proj": {
|
| 152 |
+
"act_bits": 16,
|
| 153 |
+
"act_data_type": "float",
|
| 154 |
+
"bits": 16,
|
| 155 |
+
"data_type": "float"
|
| 156 |
+
},
|
| 157 |
+
"model.layers.11.self_attn.v_proj": {
|
| 158 |
+
"act_bits": 16,
|
| 159 |
+
"act_data_type": "float",
|
| 160 |
+
"bits": 16,
|
| 161 |
+
"data_type": "float"
|
| 162 |
+
},
|
| 163 |
+
"model.layers.12.self_attn.k_proj": {
|
| 164 |
+
"act_bits": 16,
|
| 165 |
+
"act_data_type": "float",
|
| 166 |
+
"bits": 16,
|
| 167 |
+
"data_type": "float"
|
| 168 |
+
},
|
| 169 |
+
"model.layers.12.self_attn.o_proj": {
|
| 170 |
+
"act_bits": 16,
|
| 171 |
+
"act_data_type": "float",
|
| 172 |
+
"bits": 16,
|
| 173 |
+
"data_type": "float"
|
| 174 |
+
},
|
| 175 |
+
"model.layers.12.self_attn.q_proj": {
|
| 176 |
+
"act_bits": 16,
|
| 177 |
+
"act_data_type": "float",
|
| 178 |
+
"bits": 16,
|
| 179 |
+
"data_type": "float"
|
| 180 |
+
},
|
| 181 |
+
"model.layers.12.self_attn.v_proj": {
|
| 182 |
+
"act_bits": 16,
|
| 183 |
+
"act_data_type": "float",
|
| 184 |
+
"bits": 16,
|
| 185 |
+
"data_type": "float"
|
| 186 |
+
},
|
| 187 |
+
"model.layers.13.self_attn.k_proj": {
|
| 188 |
+
"act_bits": 16,
|
| 189 |
+
"act_data_type": "float",
|
| 190 |
+
"bits": 16,
|
| 191 |
+
"data_type": "float"
|
| 192 |
+
},
|
| 193 |
+
"model.layers.13.self_attn.o_proj": {
|
| 194 |
+
"act_bits": 16,
|
| 195 |
+
"act_data_type": "float",
|
| 196 |
+
"bits": 16,
|
| 197 |
+
"data_type": "float"
|
| 198 |
+
},
|
| 199 |
+
"model.layers.13.self_attn.q_proj": {
|
| 200 |
+
"act_bits": 16,
|
| 201 |
+
"act_data_type": "float",
|
| 202 |
+
"bits": 16,
|
| 203 |
+
"data_type": "float"
|
| 204 |
+
},
|
| 205 |
+
"model.layers.13.self_attn.v_proj": {
|
| 206 |
+
"act_bits": 16,
|
| 207 |
+
"act_data_type": "float",
|
| 208 |
+
"bits": 16,
|
| 209 |
+
"data_type": "float"
|
| 210 |
+
},
|
| 211 |
+
"model.layers.14.self_attn.k_proj": {
|
| 212 |
+
"act_bits": 16,
|
| 213 |
+
"act_data_type": "float",
|
| 214 |
+
"bits": 16,
|
| 215 |
+
"data_type": "float"
|
| 216 |
+
},
|
| 217 |
+
"model.layers.14.self_attn.o_proj": {
|
| 218 |
+
"act_bits": 16,
|
| 219 |
+
"act_data_type": "float",
|
| 220 |
+
"bits": 16,
|
| 221 |
+
"data_type": "float"
|
| 222 |
+
},
|
| 223 |
+
"model.layers.14.self_attn.q_proj": {
|
| 224 |
+
"act_bits": 16,
|
| 225 |
+
"act_data_type": "float",
|
| 226 |
+
"bits": 16,
|
| 227 |
+
"data_type": "float"
|
| 228 |
+
},
|
| 229 |
+
"model.layers.14.self_attn.v_proj": {
|
| 230 |
+
"act_bits": 16,
|
| 231 |
+
"act_data_type": "float",
|
| 232 |
+
"bits": 16,
|
| 233 |
+
"data_type": "float"
|
| 234 |
+
},
|
| 235 |
+
"model.layers.15.self_attn.k_proj": {
|
| 236 |
+
"act_bits": 16,
|
| 237 |
+
"act_data_type": "float",
|
| 238 |
+
"bits": 16,
|
| 239 |
+
"data_type": "float"
|
| 240 |
+
},
|
| 241 |
+
"model.layers.15.self_attn.o_proj": {
|
| 242 |
+
"act_bits": 16,
|
| 243 |
+
"act_data_type": "float",
|
| 244 |
+
"bits": 16,
|
| 245 |
+
"data_type": "float"
|
| 246 |
+
},
|
| 247 |
+
"model.layers.15.self_attn.q_proj": {
|
| 248 |
+
"act_bits": 16,
|
| 249 |
+
"act_data_type": "float",
|
| 250 |
+
"bits": 16,
|
| 251 |
+
"data_type": "float"
|
| 252 |
+
},
|
| 253 |
+
"model.layers.15.self_attn.v_proj": {
|
| 254 |
+
"act_bits": 16,
|
| 255 |
+
"act_data_type": "float",
|
| 256 |
+
"bits": 16,
|
| 257 |
+
"data_type": "float"
|
| 258 |
+
},
|
| 259 |
+
"model.layers.16.self_attn.k_proj": {
|
| 260 |
+
"act_bits": 16,
|
| 261 |
+
"act_data_type": "float",
|
| 262 |
+
"bits": 16,
|
| 263 |
+
"data_type": "float"
|
| 264 |
+
},
|
| 265 |
+
"model.layers.16.self_attn.o_proj": {
|
| 266 |
+
"act_bits": 16,
|
| 267 |
+
"act_data_type": "float",
|
| 268 |
+
"bits": 16,
|
| 269 |
+
"data_type": "float"
|
| 270 |
+
},
|
| 271 |
+
"model.layers.16.self_attn.q_proj": {
|
| 272 |
+
"act_bits": 16,
|
| 273 |
+
"act_data_type": "float",
|
| 274 |
+
"bits": 16,
|
| 275 |
+
"data_type": "float"
|
| 276 |
+
},
|
| 277 |
+
"model.layers.16.self_attn.v_proj": {
|
| 278 |
+
"act_bits": 16,
|
| 279 |
+
"act_data_type": "float",
|
| 280 |
+
"bits": 16,
|
| 281 |
+
"data_type": "float"
|
| 282 |
+
},
|
| 283 |
+
"model.layers.17.self_attn.k_proj": {
|
| 284 |
+
"act_bits": 16,
|
| 285 |
+
"act_data_type": "float",
|
| 286 |
+
"bits": 16,
|
| 287 |
+
"data_type": "float"
|
| 288 |
+
},
|
| 289 |
+
"model.layers.17.self_attn.o_proj": {
|
| 290 |
+
"act_bits": 16,
|
| 291 |
+
"act_data_type": "float",
|
| 292 |
+
"bits": 16,
|
| 293 |
+
"data_type": "float"
|
| 294 |
+
},
|
| 295 |
+
"model.layers.17.self_attn.q_proj": {
|
| 296 |
+
"act_bits": 16,
|
| 297 |
+
"act_data_type": "float",
|
| 298 |
+
"bits": 16,
|
| 299 |
+
"data_type": "float"
|
| 300 |
+
},
|
| 301 |
+
"model.layers.17.self_attn.v_proj": {
|
| 302 |
+
"act_bits": 16,
|
| 303 |
+
"act_data_type": "float",
|
| 304 |
+
"bits": 16,
|
| 305 |
+
"data_type": "float"
|
| 306 |
+
},
|
| 307 |
+
"model.layers.18.self_attn.k_proj": {
|
| 308 |
+
"act_bits": 16,
|
| 309 |
+
"act_data_type": "float",
|
| 310 |
+
"bits": 16,
|
| 311 |
+
"data_type": "float"
|
| 312 |
+
},
|
| 313 |
+
"model.layers.18.self_attn.o_proj": {
|
| 314 |
+
"act_bits": 16,
|
| 315 |
+
"act_data_type": "float",
|
| 316 |
+
"bits": 16,
|
| 317 |
+
"data_type": "float"
|
| 318 |
+
},
|
| 319 |
+
"model.layers.18.self_attn.q_proj": {
|
| 320 |
+
"act_bits": 16,
|
| 321 |
+
"act_data_type": "float",
|
| 322 |
+
"bits": 16,
|
| 323 |
+
"data_type": "float"
|
| 324 |
+
},
|
| 325 |
+
"model.layers.18.self_attn.v_proj": {
|
| 326 |
+
"act_bits": 16,
|
| 327 |
+
"act_data_type": "float",
|
| 328 |
+
"bits": 16,
|
| 329 |
+
"data_type": "float"
|
| 330 |
+
},
|
| 331 |
+
"model.layers.19.self_attn.k_proj": {
|
| 332 |
+
"act_bits": 16,
|
| 333 |
+
"act_data_type": "float",
|
| 334 |
+
"bits": 16,
|
| 335 |
+
"data_type": "float"
|
| 336 |
+
},
|
| 337 |
+
"model.layers.19.self_attn.o_proj": {
|
| 338 |
+
"act_bits": 16,
|
| 339 |
+
"act_data_type": "float",
|
| 340 |
+
"bits": 16,
|
| 341 |
+
"data_type": "float"
|
| 342 |
+
},
|
| 343 |
+
"model.layers.19.self_attn.q_proj": {
|
| 344 |
+
"act_bits": 16,
|
| 345 |
+
"act_data_type": "float",
|
| 346 |
+
"bits": 16,
|
| 347 |
+
"data_type": "float"
|
| 348 |
+
},
|
| 349 |
+
"model.layers.19.self_attn.v_proj": {
|
| 350 |
+
"act_bits": 16,
|
| 351 |
+
"act_data_type": "float",
|
| 352 |
+
"bits": 16,
|
| 353 |
+
"data_type": "float"
|
| 354 |
+
},
|
| 355 |
+
"model.layers.2.self_attn.k_proj": {
|
| 356 |
+
"act_bits": 16,
|
| 357 |
+
"act_data_type": "float",
|
| 358 |
+
"bits": 16,
|
| 359 |
+
"data_type": "float"
|
| 360 |
+
},
|
| 361 |
+
"model.layers.2.self_attn.o_proj": {
|
| 362 |
+
"act_bits": 16,
|
| 363 |
+
"act_data_type": "float",
|
| 364 |
+
"bits": 16,
|
| 365 |
+
"data_type": "float"
|
| 366 |
+
},
|
| 367 |
+
"model.layers.2.self_attn.q_proj": {
|
| 368 |
+
"act_bits": 16,
|
| 369 |
+
"act_data_type": "float",
|
| 370 |
+
"bits": 16,
|
| 371 |
+
"data_type": "float"
|
| 372 |
+
},
|
| 373 |
+
"model.layers.2.self_attn.v_proj": {
|
| 374 |
+
"act_bits": 16,
|
| 375 |
+
"act_data_type": "float",
|
| 376 |
+
"bits": 16,
|
| 377 |
+
"data_type": "float"
|
| 378 |
+
},
|
| 379 |
+
"model.layers.20.self_attn.k_proj": {
|
| 380 |
+
"act_bits": 16,
|
| 381 |
+
"act_data_type": "float",
|
| 382 |
+
"bits": 16,
|
| 383 |
+
"data_type": "float"
|
| 384 |
+
},
|
| 385 |
+
"model.layers.20.self_attn.o_proj": {
|
| 386 |
+
"act_bits": 16,
|
| 387 |
+
"act_data_type": "float",
|
| 388 |
+
"bits": 16,
|
| 389 |
+
"data_type": "float"
|
| 390 |
+
},
|
| 391 |
+
"model.layers.20.self_attn.q_proj": {
|
| 392 |
+
"act_bits": 16,
|
| 393 |
+
"act_data_type": "float",
|
| 394 |
+
"bits": 16,
|
| 395 |
+
"data_type": "float"
|
| 396 |
+
},
|
| 397 |
+
"model.layers.20.self_attn.v_proj": {
|
| 398 |
+
"act_bits": 16,
|
| 399 |
+
"act_data_type": "float",
|
| 400 |
+
"bits": 16,
|
| 401 |
+
"data_type": "float"
|
| 402 |
+
},
|
| 403 |
+
"model.layers.21.self_attn.k_proj": {
|
| 404 |
+
"act_bits": 16,
|
| 405 |
+
"act_data_type": "float",
|
| 406 |
+
"bits": 16,
|
| 407 |
+
"data_type": "float"
|
| 408 |
+
},
|
| 409 |
+
"model.layers.21.self_attn.o_proj": {
|
| 410 |
+
"act_bits": 16,
|
| 411 |
+
"act_data_type": "float",
|
| 412 |
+
"bits": 16,
|
| 413 |
+
"data_type": "float"
|
| 414 |
+
},
|
| 415 |
+
"model.layers.21.self_attn.q_proj": {
|
| 416 |
+
"act_bits": 16,
|
| 417 |
+
"act_data_type": "float",
|
| 418 |
+
"bits": 16,
|
| 419 |
+
"data_type": "float"
|
| 420 |
+
},
|
| 421 |
+
"model.layers.21.self_attn.v_proj": {
|
| 422 |
+
"act_bits": 16,
|
| 423 |
+
"act_data_type": "float",
|
| 424 |
+
"bits": 16,
|
| 425 |
+
"data_type": "float"
|
| 426 |
+
},
|
| 427 |
+
"model.layers.22.self_attn.k_proj": {
|
| 428 |
+
"act_bits": 16,
|
| 429 |
+
"act_data_type": "float",
|
| 430 |
+
"bits": 16,
|
| 431 |
+
"data_type": "float"
|
| 432 |
+
},
|
| 433 |
+
"model.layers.22.self_attn.o_proj": {
|
| 434 |
+
"act_bits": 16,
|
| 435 |
+
"act_data_type": "float",
|
| 436 |
+
"bits": 16,
|
| 437 |
+
"data_type": "float"
|
| 438 |
+
},
|
| 439 |
+
"model.layers.22.self_attn.q_proj": {
|
| 440 |
+
"act_bits": 16,
|
| 441 |
+
"act_data_type": "float",
|
| 442 |
+
"bits": 16,
|
| 443 |
+
"data_type": "float"
|
| 444 |
+
},
|
| 445 |
+
"model.layers.22.self_attn.v_proj": {
|
| 446 |
+
"act_bits": 16,
|
| 447 |
+
"act_data_type": "float",
|
| 448 |
+
"bits": 16,
|
| 449 |
+
"data_type": "float"
|
| 450 |
+
},
|
| 451 |
+
"model.layers.23.self_attn.k_proj": {
|
| 452 |
+
"act_bits": 16,
|
| 453 |
+
"act_data_type": "float",
|
| 454 |
+
"bits": 16,
|
| 455 |
+
"data_type": "float"
|
| 456 |
+
},
|
| 457 |
+
"model.layers.23.self_attn.o_proj": {
|
| 458 |
+
"act_bits": 16,
|
| 459 |
+
"act_data_type": "float",
|
| 460 |
+
"bits": 16,
|
| 461 |
+
"data_type": "float"
|
| 462 |
+
},
|
| 463 |
+
"model.layers.23.self_attn.q_proj": {
|
| 464 |
+
"act_bits": 16,
|
| 465 |
+
"act_data_type": "float",
|
| 466 |
+
"bits": 16,
|
| 467 |
+
"data_type": "float"
|
| 468 |
+
},
|
| 469 |
+
"model.layers.23.self_attn.v_proj": {
|
| 470 |
+
"act_bits": 16,
|
| 471 |
+
"act_data_type": "float",
|
| 472 |
+
"bits": 16,
|
| 473 |
+
"data_type": "float"
|
| 474 |
+
},
|
| 475 |
+
"model.layers.24.self_attn.k_proj": {
|
| 476 |
+
"act_bits": 16,
|
| 477 |
+
"act_data_type": "float",
|
| 478 |
+
"bits": 16,
|
| 479 |
+
"data_type": "float"
|
| 480 |
+
},
|
| 481 |
+
"model.layers.24.self_attn.o_proj": {
|
| 482 |
+
"act_bits": 16,
|
| 483 |
+
"act_data_type": "float",
|
| 484 |
+
"bits": 16,
|
| 485 |
+
"data_type": "float"
|
| 486 |
+
},
|
| 487 |
+
"model.layers.24.self_attn.q_proj": {
|
| 488 |
+
"act_bits": 16,
|
| 489 |
+
"act_data_type": "float",
|
| 490 |
+
"bits": 16,
|
| 491 |
+
"data_type": "float"
|
| 492 |
+
},
|
| 493 |
+
"model.layers.24.self_attn.v_proj": {
|
| 494 |
+
"act_bits": 16,
|
| 495 |
+
"act_data_type": "float",
|
| 496 |
+
"bits": 16,
|
| 497 |
+
"data_type": "float"
|
| 498 |
+
},
|
| 499 |
+
"model.layers.25.self_attn.k_proj": {
|
| 500 |
+
"act_bits": 16,
|
| 501 |
+
"act_data_type": "float",
|
| 502 |
+
"bits": 16,
|
| 503 |
+
"data_type": "float"
|
| 504 |
+
},
|
| 505 |
+
"model.layers.25.self_attn.o_proj": {
|
| 506 |
+
"act_bits": 16,
|
| 507 |
+
"act_data_type": "float",
|
| 508 |
+
"bits": 16,
|
| 509 |
+
"data_type": "float"
|
| 510 |
+
},
|
| 511 |
+
"model.layers.25.self_attn.q_proj": {
|
| 512 |
+
"act_bits": 16,
|
| 513 |
+
"act_data_type": "float",
|
| 514 |
+
"bits": 16,
|
| 515 |
+
"data_type": "float"
|
| 516 |
+
},
|
| 517 |
+
"model.layers.25.self_attn.v_proj": {
|
| 518 |
+
"act_bits": 16,
|
| 519 |
+
"act_data_type": "float",
|
| 520 |
+
"bits": 16,
|
| 521 |
+
"data_type": "float"
|
| 522 |
+
},
|
| 523 |
+
"model.layers.26.self_attn.k_proj": {
|
| 524 |
+
"act_bits": 16,
|
| 525 |
+
"act_data_type": "float",
|
| 526 |
+
"bits": 16,
|
| 527 |
+
"data_type": "float"
|
| 528 |
+
},
|
| 529 |
+
"model.layers.26.self_attn.o_proj": {
|
| 530 |
+
"act_bits": 16,
|
| 531 |
+
"act_data_type": "float",
|
| 532 |
+
"bits": 16,
|
| 533 |
+
"data_type": "float"
|
| 534 |
+
},
|
| 535 |
+
"model.layers.26.self_attn.q_proj": {
|
| 536 |
+
"act_bits": 16,
|
| 537 |
+
"act_data_type": "float",
|
| 538 |
+
"bits": 16,
|
| 539 |
+
"data_type": "float"
|
| 540 |
+
},
|
| 541 |
+
"model.layers.26.self_attn.v_proj": {
|
| 542 |
+
"act_bits": 16,
|
| 543 |
+
"act_data_type": "float",
|
| 544 |
+
"bits": 16,
|
| 545 |
+
"data_type": "float"
|
| 546 |
+
},
|
| 547 |
+
"model.layers.27.self_attn.k_proj": {
|
| 548 |
+
"act_bits": 16,
|
| 549 |
+
"act_data_type": "float",
|
| 550 |
+
"bits": 16,
|
| 551 |
+
"data_type": "float"
|
| 552 |
+
},
|
| 553 |
+
"model.layers.27.self_attn.o_proj": {
|
| 554 |
+
"act_bits": 16,
|
| 555 |
+
"act_data_type": "float",
|
| 556 |
+
"bits": 16,
|
| 557 |
+
"data_type": "float"
|
| 558 |
+
},
|
| 559 |
+
"model.layers.27.self_attn.q_proj": {
|
| 560 |
+
"act_bits": 16,
|
| 561 |
+
"act_data_type": "float",
|
| 562 |
+
"bits": 16,
|
| 563 |
+
"data_type": "float"
|
| 564 |
+
},
|
| 565 |
+
"model.layers.27.self_attn.v_proj": {
|
| 566 |
+
"act_bits": 16,
|
| 567 |
+
"act_data_type": "float",
|
| 568 |
+
"bits": 16,
|
| 569 |
+
"data_type": "float"
|
| 570 |
+
},
|
| 571 |
+
"model.layers.3.self_attn.k_proj": {
|
| 572 |
+
"act_bits": 16,
|
| 573 |
+
"act_data_type": "float",
|
| 574 |
+
"bits": 16,
|
| 575 |
+
"data_type": "float"
|
| 576 |
+
},
|
| 577 |
+
"model.layers.3.self_attn.o_proj": {
|
| 578 |
+
"act_bits": 16,
|
| 579 |
+
"act_data_type": "float",
|
| 580 |
+
"bits": 16,
|
| 581 |
+
"data_type": "float"
|
| 582 |
+
},
|
| 583 |
+
"model.layers.3.self_attn.q_proj": {
|
| 584 |
+
"act_bits": 16,
|
| 585 |
+
"act_data_type": "float",
|
| 586 |
+
"bits": 16,
|
| 587 |
+
"data_type": "float"
|
| 588 |
+
},
|
| 589 |
+
"model.layers.3.self_attn.v_proj": {
|
| 590 |
+
"act_bits": 16,
|
| 591 |
+
"act_data_type": "float",
|
| 592 |
+
"bits": 16,
|
| 593 |
+
"data_type": "float"
|
| 594 |
+
},
|
| 595 |
+
"model.layers.4.self_attn.k_proj": {
|
| 596 |
+
"act_bits": 16,
|
| 597 |
+
"act_data_type": "float",
|
| 598 |
+
"bits": 16,
|
| 599 |
+
"data_type": "float"
|
| 600 |
+
},
|
| 601 |
+
"model.layers.4.self_attn.o_proj": {
|
| 602 |
+
"act_bits": 16,
|
| 603 |
+
"act_data_type": "float",
|
| 604 |
+
"bits": 16,
|
| 605 |
+
"data_type": "float"
|
| 606 |
+
},
|
| 607 |
+
"model.layers.4.self_attn.q_proj": {
|
| 608 |
+
"act_bits": 16,
|
| 609 |
+
"act_data_type": "float",
|
| 610 |
+
"bits": 16,
|
| 611 |
+
"data_type": "float"
|
| 612 |
+
},
|
| 613 |
+
"model.layers.4.self_attn.v_proj": {
|
| 614 |
+
"act_bits": 16,
|
| 615 |
+
"act_data_type": "float",
|
| 616 |
+
"bits": 16,
|
| 617 |
+
"data_type": "float"
|
| 618 |
+
},
|
| 619 |
+
"model.layers.5.self_attn.k_proj": {
|
| 620 |
+
"act_bits": 16,
|
| 621 |
+
"act_data_type": "float",
|
| 622 |
+
"bits": 16,
|
| 623 |
+
"data_type": "float"
|
| 624 |
+
},
|
| 625 |
+
"model.layers.5.self_attn.o_proj": {
|
| 626 |
+
"act_bits": 16,
|
| 627 |
+
"act_data_type": "float",
|
| 628 |
+
"bits": 16,
|
| 629 |
+
"data_type": "float"
|
| 630 |
+
},
|
| 631 |
+
"model.layers.5.self_attn.q_proj": {
|
| 632 |
+
"act_bits": 16,
|
| 633 |
+
"act_data_type": "float",
|
| 634 |
+
"bits": 16,
|
| 635 |
+
"data_type": "float"
|
| 636 |
+
},
|
| 637 |
+
"model.layers.5.self_attn.v_proj": {
|
| 638 |
+
"act_bits": 16,
|
| 639 |
+
"act_data_type": "float",
|
| 640 |
+
"bits": 16,
|
| 641 |
+
"data_type": "float"
|
| 642 |
+
},
|
| 643 |
+
"model.layers.6.self_attn.k_proj": {
|
| 644 |
+
"act_bits": 16,
|
| 645 |
+
"act_data_type": "float",
|
| 646 |
+
"bits": 16,
|
| 647 |
+
"data_type": "float"
|
| 648 |
+
},
|
| 649 |
+
"model.layers.6.self_attn.o_proj": {
|
| 650 |
+
"act_bits": 16,
|
| 651 |
+
"act_data_type": "float",
|
| 652 |
+
"bits": 16,
|
| 653 |
+
"data_type": "float"
|
| 654 |
+
},
|
| 655 |
+
"model.layers.6.self_attn.q_proj": {
|
| 656 |
+
"act_bits": 16,
|
| 657 |
+
"act_data_type": "float",
|
| 658 |
+
"bits": 16,
|
| 659 |
+
"data_type": "float"
|
| 660 |
+
},
|
| 661 |
+
"model.layers.6.self_attn.v_proj": {
|
| 662 |
+
"act_bits": 16,
|
| 663 |
+
"act_data_type": "float",
|
| 664 |
+
"bits": 16,
|
| 665 |
+
"data_type": "float"
|
| 666 |
+
},
|
| 667 |
+
"model.layers.7.self_attn.k_proj": {
|
| 668 |
+
"act_bits": 16,
|
| 669 |
+
"act_data_type": "float",
|
| 670 |
+
"bits": 16,
|
| 671 |
+
"data_type": "float"
|
| 672 |
+
},
|
| 673 |
+
"model.layers.7.self_attn.o_proj": {
|
| 674 |
+
"act_bits": 16,
|
| 675 |
+
"act_data_type": "float",
|
| 676 |
+
"bits": 16,
|
| 677 |
+
"data_type": "float"
|
| 678 |
+
},
|
| 679 |
+
"model.layers.7.self_attn.q_proj": {
|
| 680 |
+
"act_bits": 16,
|
| 681 |
+
"act_data_type": "float",
|
| 682 |
+
"bits": 16,
|
| 683 |
+
"data_type": "float"
|
| 684 |
+
},
|
| 685 |
+
"model.layers.7.self_attn.v_proj": {
|
| 686 |
+
"act_bits": 16,
|
| 687 |
+
"act_data_type": "float",
|
| 688 |
+
"bits": 16,
|
| 689 |
+
"data_type": "float"
|
| 690 |
+
},
|
| 691 |
+
"model.layers.8.self_attn.k_proj": {
|
| 692 |
+
"act_bits": 16,
|
| 693 |
+
"act_data_type": "float",
|
| 694 |
+
"bits": 16,
|
| 695 |
+
"data_type": "float"
|
| 696 |
+
},
|
| 697 |
+
"model.layers.8.self_attn.o_proj": {
|
| 698 |
+
"act_bits": 16,
|
| 699 |
+
"act_data_type": "float",
|
| 700 |
+
"bits": 16,
|
| 701 |
+
"data_type": "float"
|
| 702 |
+
},
|
| 703 |
+
"model.layers.8.self_attn.q_proj": {
|
| 704 |
+
"act_bits": 16,
|
| 705 |
+
"act_data_type": "float",
|
| 706 |
+
"bits": 16,
|
| 707 |
+
"data_type": "float"
|
| 708 |
+
},
|
| 709 |
+
"model.layers.8.self_attn.v_proj": {
|
| 710 |
+
"act_bits": 16,
|
| 711 |
+
"act_data_type": "float",
|
| 712 |
+
"bits": 16,
|
| 713 |
+
"data_type": "float"
|
| 714 |
+
},
|
| 715 |
+
"model.layers.9.self_attn.k_proj": {
|
| 716 |
+
"act_bits": 16,
|
| 717 |
+
"act_data_type": "float",
|
| 718 |
+
"bits": 16,
|
| 719 |
+
"data_type": "float"
|
| 720 |
+
},
|
| 721 |
+
"model.layers.9.self_attn.o_proj": {
|
| 722 |
+
"act_bits": 16,
|
| 723 |
+
"act_data_type": "float",
|
| 724 |
+
"bits": 16,
|
| 725 |
+
"data_type": "float"
|
| 726 |
+
},
|
| 727 |
+
"model.layers.9.self_attn.q_proj": {
|
| 728 |
+
"act_bits": 16,
|
| 729 |
+
"act_data_type": "float",
|
| 730 |
+
"bits": 16,
|
| 731 |
+
"data_type": "float"
|
| 732 |
+
},
|
| 733 |
+
"model.layers.9.self_attn.v_proj": {
|
| 734 |
+
"act_bits": 16,
|
| 735 |
+
"act_data_type": "float",
|
| 736 |
+
"bits": 16,
|
| 737 |
+
"data_type": "float"
|
| 738 |
+
}
|
| 739 |
+
},
|
| 740 |
+
"group_size": 16,
|
| 741 |
+
"low_gpu_mem_usage": true,
|
| 742 |
+
"packing_format": "auto_round:llm_compressor",
|
| 743 |
+
"quant_method": "auto-round",
|
| 744 |
+
"sym": true
|
| 745 |
+
},
|
| 746 |
+
"rms_norm_eps": 1e-06,
|
| 747 |
+
"rope_parameters": {
|
| 748 |
+
"rope_theta": 10000,
|
| 749 |
+
"rope_type": "default"
|
| 750 |
+
},
|
| 751 |
+
"sliding_window": null,
|
| 752 |
+
"tie_word_embeddings": false,
|
| 753 |
+
"transformers_version": "5.12.1",
|
| 754 |
+
"use_cache": true,
|
| 755 |
+
"use_mrope": false,
|
| 756 |
+
"use_sliding_window": false,
|
| 757 |
+
"vocab_size": 151936
|
| 758 |
+
}
|
generation_config.json
ADDED
|
@@ -0,0 +1,9 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"_from_model_config": true,
|
| 3 |
+
"bos_token_id": 151646,
|
| 4 |
+
"do_sample": true,
|
| 5 |
+
"eos_token_id": 151643,
|
| 6 |
+
"temperature": 0.6,
|
| 7 |
+
"top_p": 0.95,
|
| 8 |
+
"transformers_version": "5.12.1"
|
| 9 |
+
}
|
model.safetensors
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:b2a181186f5acba7e1d2016ca7c8280f05b7a52893703af4898c9b5b64326e49
|
| 3 |
+
size 1892414080
|
quantization_config.json
ADDED
|
@@ -0,0 +1,696 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"bits": 4,
|
| 3 |
+
"act_bits": 4,
|
| 4 |
+
"data_type": "nv_fp",
|
| 5 |
+
"act_data_type": "nv_fp4_with_static_gs",
|
| 6 |
+
"group_size": 16,
|
| 7 |
+
"act_group_size": 16,
|
| 8 |
+
"sym": true,
|
| 9 |
+
"act_sym": true,
|
| 10 |
+
"act_dynamic": true,
|
| 11 |
+
"low_gpu_mem_usage": true,
|
| 12 |
+
"autoround_version": "0.13.1",
|
| 13 |
+
"block_name_to_quantize": "model.layers",
|
| 14 |
+
"quant_method": "auto-round",
|
| 15 |
+
"packing_format": "auto_round:llm_compressor",
|
| 16 |
+
"extra_config": {
|
| 17 |
+
"model.layers.0.self_attn.q_proj": {
|
| 18 |
+
"bits": 16,
|
| 19 |
+
"data_type": "float",
|
| 20 |
+
"act_bits": 16,
|
| 21 |
+
"act_data_type": "float"
|
| 22 |
+
},
|
| 23 |
+
"model.layers.0.self_attn.k_proj": {
|
| 24 |
+
"bits": 16,
|
| 25 |
+
"data_type": "float",
|
| 26 |
+
"act_bits": 16,
|
| 27 |
+
"act_data_type": "float"
|
| 28 |
+
},
|
| 29 |
+
"model.layers.0.self_attn.v_proj": {
|
| 30 |
+
"bits": 16,
|
| 31 |
+
"data_type": "float",
|
| 32 |
+
"act_bits": 16,
|
| 33 |
+
"act_data_type": "float"
|
| 34 |
+
},
|
| 35 |
+
"model.layers.0.self_attn.o_proj": {
|
| 36 |
+
"bits": 16,
|
| 37 |
+
"data_type": "float",
|
| 38 |
+
"act_bits": 16,
|
| 39 |
+
"act_data_type": "float"
|
| 40 |
+
},
|
| 41 |
+
"model.layers.1.self_attn.q_proj": {
|
| 42 |
+
"bits": 16,
|
| 43 |
+
"data_type": "float",
|
| 44 |
+
"act_bits": 16,
|
| 45 |
+
"act_data_type": "float"
|
| 46 |
+
},
|
| 47 |
+
"model.layers.1.self_attn.k_proj": {
|
| 48 |
+
"bits": 16,
|
| 49 |
+
"data_type": "float",
|
| 50 |
+
"act_bits": 16,
|
| 51 |
+
"act_data_type": "float"
|
| 52 |
+
},
|
| 53 |
+
"model.layers.1.self_attn.v_proj": {
|
| 54 |
+
"bits": 16,
|
| 55 |
+
"data_type": "float",
|
| 56 |
+
"act_bits": 16,
|
| 57 |
+
"act_data_type": "float"
|
| 58 |
+
},
|
| 59 |
+
"model.layers.1.self_attn.o_proj": {
|
| 60 |
+
"bits": 16,
|
| 61 |
+
"data_type": "float",
|
| 62 |
+
"act_bits": 16,
|
| 63 |
+
"act_data_type": "float"
|
| 64 |
+
},
|
| 65 |
+
"model.layers.2.self_attn.q_proj": {
|
| 66 |
+
"bits": 16,
|
| 67 |
+
"data_type": "float",
|
| 68 |
+
"act_bits": 16,
|
| 69 |
+
"act_data_type": "float"
|
| 70 |
+
},
|
| 71 |
+
"model.layers.2.self_attn.k_proj": {
|
| 72 |
+
"bits": 16,
|
| 73 |
+
"data_type": "float",
|
| 74 |
+
"act_bits": 16,
|
| 75 |
+
"act_data_type": "float"
|
| 76 |
+
},
|
| 77 |
+
"model.layers.2.self_attn.v_proj": {
|
| 78 |
+
"bits": 16,
|
| 79 |
+
"data_type": "float",
|
| 80 |
+
"act_bits": 16,
|
| 81 |
+
"act_data_type": "float"
|
| 82 |
+
},
|
| 83 |
+
"model.layers.2.self_attn.o_proj": {
|
| 84 |
+
"bits": 16,
|
| 85 |
+
"data_type": "float",
|
| 86 |
+
"act_bits": 16,
|
| 87 |
+
"act_data_type": "float"
|
| 88 |
+
},
|
| 89 |
+
"model.layers.3.self_attn.q_proj": {
|
| 90 |
+
"bits": 16,
|
| 91 |
+
"data_type": "float",
|
| 92 |
+
"act_bits": 16,
|
| 93 |
+
"act_data_type": "float"
|
| 94 |
+
},
|
| 95 |
+
"model.layers.3.self_attn.k_proj": {
|
| 96 |
+
"bits": 16,
|
| 97 |
+
"data_type": "float",
|
| 98 |
+
"act_bits": 16,
|
| 99 |
+
"act_data_type": "float"
|
| 100 |
+
},
|
| 101 |
+
"model.layers.3.self_attn.v_proj": {
|
| 102 |
+
"bits": 16,
|
| 103 |
+
"data_type": "float",
|
| 104 |
+
"act_bits": 16,
|
| 105 |
+
"act_data_type": "float"
|
| 106 |
+
},
|
| 107 |
+
"model.layers.3.self_attn.o_proj": {
|
| 108 |
+
"bits": 16,
|
| 109 |
+
"data_type": "float",
|
| 110 |
+
"act_bits": 16,
|
| 111 |
+
"act_data_type": "float"
|
| 112 |
+
},
|
| 113 |
+
"model.layers.4.self_attn.q_proj": {
|
| 114 |
+
"bits": 16,
|
| 115 |
+
"data_type": "float",
|
| 116 |
+
"act_bits": 16,
|
| 117 |
+
"act_data_type": "float"
|
| 118 |
+
},
|
| 119 |
+
"model.layers.4.self_attn.k_proj": {
|
| 120 |
+
"bits": 16,
|
| 121 |
+
"data_type": "float",
|
| 122 |
+
"act_bits": 16,
|
| 123 |
+
"act_data_type": "float"
|
| 124 |
+
},
|
| 125 |
+
"model.layers.4.self_attn.v_proj": {
|
| 126 |
+
"bits": 16,
|
| 127 |
+
"data_type": "float",
|
| 128 |
+
"act_bits": 16,
|
| 129 |
+
"act_data_type": "float"
|
| 130 |
+
},
|
| 131 |
+
"model.layers.4.self_attn.o_proj": {
|
| 132 |
+
"bits": 16,
|
| 133 |
+
"data_type": "float",
|
| 134 |
+
"act_bits": 16,
|
| 135 |
+
"act_data_type": "float"
|
| 136 |
+
},
|
| 137 |
+
"model.layers.5.self_attn.q_proj": {
|
| 138 |
+
"bits": 16,
|
| 139 |
+
"data_type": "float",
|
| 140 |
+
"act_bits": 16,
|
| 141 |
+
"act_data_type": "float"
|
| 142 |
+
},
|
| 143 |
+
"model.layers.5.self_attn.k_proj": {
|
| 144 |
+
"bits": 16,
|
| 145 |
+
"data_type": "float",
|
| 146 |
+
"act_bits": 16,
|
| 147 |
+
"act_data_type": "float"
|
| 148 |
+
},
|
| 149 |
+
"model.layers.5.self_attn.v_proj": {
|
| 150 |
+
"bits": 16,
|
| 151 |
+
"data_type": "float",
|
| 152 |
+
"act_bits": 16,
|
| 153 |
+
"act_data_type": "float"
|
| 154 |
+
},
|
| 155 |
+
"model.layers.5.self_attn.o_proj": {
|
| 156 |
+
"bits": 16,
|
| 157 |
+
"data_type": "float",
|
| 158 |
+
"act_bits": 16,
|
| 159 |
+
"act_data_type": "float"
|
| 160 |
+
},
|
| 161 |
+
"model.layers.6.self_attn.q_proj": {
|
| 162 |
+
"bits": 16,
|
| 163 |
+
"data_type": "float",
|
| 164 |
+
"act_bits": 16,
|
| 165 |
+
"act_data_type": "float"
|
| 166 |
+
},
|
| 167 |
+
"model.layers.6.self_attn.k_proj": {
|
| 168 |
+
"bits": 16,
|
| 169 |
+
"data_type": "float",
|
| 170 |
+
"act_bits": 16,
|
| 171 |
+
"act_data_type": "float"
|
| 172 |
+
},
|
| 173 |
+
"model.layers.6.self_attn.v_proj": {
|
| 174 |
+
"bits": 16,
|
| 175 |
+
"data_type": "float",
|
| 176 |
+
"act_bits": 16,
|
| 177 |
+
"act_data_type": "float"
|
| 178 |
+
},
|
| 179 |
+
"model.layers.6.self_attn.o_proj": {
|
| 180 |
+
"bits": 16,
|
| 181 |
+
"data_type": "float",
|
| 182 |
+
"act_bits": 16,
|
| 183 |
+
"act_data_type": "float"
|
| 184 |
+
},
|
| 185 |
+
"model.layers.7.self_attn.q_proj": {
|
| 186 |
+
"bits": 16,
|
| 187 |
+
"data_type": "float",
|
| 188 |
+
"act_bits": 16,
|
| 189 |
+
"act_data_type": "float"
|
| 190 |
+
},
|
| 191 |
+
"model.layers.7.self_attn.k_proj": {
|
| 192 |
+
"bits": 16,
|
| 193 |
+
"data_type": "float",
|
| 194 |
+
"act_bits": 16,
|
| 195 |
+
"act_data_type": "float"
|
| 196 |
+
},
|
| 197 |
+
"model.layers.7.self_attn.v_proj": {
|
| 198 |
+
"bits": 16,
|
| 199 |
+
"data_type": "float",
|
| 200 |
+
"act_bits": 16,
|
| 201 |
+
"act_data_type": "float"
|
| 202 |
+
},
|
| 203 |
+
"model.layers.7.self_attn.o_proj": {
|
| 204 |
+
"bits": 16,
|
| 205 |
+
"data_type": "float",
|
| 206 |
+
"act_bits": 16,
|
| 207 |
+
"act_data_type": "float"
|
| 208 |
+
},
|
| 209 |
+
"model.layers.8.self_attn.q_proj": {
|
| 210 |
+
"bits": 16,
|
| 211 |
+
"data_type": "float",
|
| 212 |
+
"act_bits": 16,
|
| 213 |
+
"act_data_type": "float"
|
| 214 |
+
},
|
| 215 |
+
"model.layers.8.self_attn.k_proj": {
|
| 216 |
+
"bits": 16,
|
| 217 |
+
"data_type": "float",
|
| 218 |
+
"act_bits": 16,
|
| 219 |
+
"act_data_type": "float"
|
| 220 |
+
},
|
| 221 |
+
"model.layers.8.self_attn.v_proj": {
|
| 222 |
+
"bits": 16,
|
| 223 |
+
"data_type": "float",
|
| 224 |
+
"act_bits": 16,
|
| 225 |
+
"act_data_type": "float"
|
| 226 |
+
},
|
| 227 |
+
"model.layers.8.self_attn.o_proj": {
|
| 228 |
+
"bits": 16,
|
| 229 |
+
"data_type": "float",
|
| 230 |
+
"act_bits": 16,
|
| 231 |
+
"act_data_type": "float"
|
| 232 |
+
},
|
| 233 |
+
"model.layers.9.self_attn.q_proj": {
|
| 234 |
+
"bits": 16,
|
| 235 |
+
"data_type": "float",
|
| 236 |
+
"act_bits": 16,
|
| 237 |
+
"act_data_type": "float"
|
| 238 |
+
},
|
| 239 |
+
"model.layers.9.self_attn.k_proj": {
|
| 240 |
+
"bits": 16,
|
| 241 |
+
"data_type": "float",
|
| 242 |
+
"act_bits": 16,
|
| 243 |
+
"act_data_type": "float"
|
| 244 |
+
},
|
| 245 |
+
"model.layers.9.self_attn.v_proj": {
|
| 246 |
+
"bits": 16,
|
| 247 |
+
"data_type": "float",
|
| 248 |
+
"act_bits": 16,
|
| 249 |
+
"act_data_type": "float"
|
| 250 |
+
},
|
| 251 |
+
"model.layers.9.self_attn.o_proj": {
|
| 252 |
+
"bits": 16,
|
| 253 |
+
"data_type": "float",
|
| 254 |
+
"act_bits": 16,
|
| 255 |
+
"act_data_type": "float"
|
| 256 |
+
},
|
| 257 |
+
"model.layers.10.self_attn.q_proj": {
|
| 258 |
+
"bits": 16,
|
| 259 |
+
"data_type": "float",
|
| 260 |
+
"act_bits": 16,
|
| 261 |
+
"act_data_type": "float"
|
| 262 |
+
},
|
| 263 |
+
"model.layers.10.self_attn.k_proj": {
|
| 264 |
+
"bits": 16,
|
| 265 |
+
"data_type": "float",
|
| 266 |
+
"act_bits": 16,
|
| 267 |
+
"act_data_type": "float"
|
| 268 |
+
},
|
| 269 |
+
"model.layers.10.self_attn.v_proj": {
|
| 270 |
+
"bits": 16,
|
| 271 |
+
"data_type": "float",
|
| 272 |
+
"act_bits": 16,
|
| 273 |
+
"act_data_type": "float"
|
| 274 |
+
},
|
| 275 |
+
"model.layers.10.self_attn.o_proj": {
|
| 276 |
+
"bits": 16,
|
| 277 |
+
"data_type": "float",
|
| 278 |
+
"act_bits": 16,
|
| 279 |
+
"act_data_type": "float"
|
| 280 |
+
},
|
| 281 |
+
"model.layers.11.self_attn.q_proj": {
|
| 282 |
+
"bits": 16,
|
| 283 |
+
"data_type": "float",
|
| 284 |
+
"act_bits": 16,
|
| 285 |
+
"act_data_type": "float"
|
| 286 |
+
},
|
| 287 |
+
"model.layers.11.self_attn.k_proj": {
|
| 288 |
+
"bits": 16,
|
| 289 |
+
"data_type": "float",
|
| 290 |
+
"act_bits": 16,
|
| 291 |
+
"act_data_type": "float"
|
| 292 |
+
},
|
| 293 |
+
"model.layers.11.self_attn.v_proj": {
|
| 294 |
+
"bits": 16,
|
| 295 |
+
"data_type": "float",
|
| 296 |
+
"act_bits": 16,
|
| 297 |
+
"act_data_type": "float"
|
| 298 |
+
},
|
| 299 |
+
"model.layers.11.self_attn.o_proj": {
|
| 300 |
+
"bits": 16,
|
| 301 |
+
"data_type": "float",
|
| 302 |
+
"act_bits": 16,
|
| 303 |
+
"act_data_type": "float"
|
| 304 |
+
},
|
| 305 |
+
"model.layers.12.self_attn.q_proj": {
|
| 306 |
+
"bits": 16,
|
| 307 |
+
"data_type": "float",
|
| 308 |
+
"act_bits": 16,
|
| 309 |
+
"act_data_type": "float"
|
| 310 |
+
},
|
| 311 |
+
"model.layers.12.self_attn.k_proj": {
|
| 312 |
+
"bits": 16,
|
| 313 |
+
"data_type": "float",
|
| 314 |
+
"act_bits": 16,
|
| 315 |
+
"act_data_type": "float"
|
| 316 |
+
},
|
| 317 |
+
"model.layers.12.self_attn.v_proj": {
|
| 318 |
+
"bits": 16,
|
| 319 |
+
"data_type": "float",
|
| 320 |
+
"act_bits": 16,
|
| 321 |
+
"act_data_type": "float"
|
| 322 |
+
},
|
| 323 |
+
"model.layers.12.self_attn.o_proj": {
|
| 324 |
+
"bits": 16,
|
| 325 |
+
"data_type": "float",
|
| 326 |
+
"act_bits": 16,
|
| 327 |
+
"act_data_type": "float"
|
| 328 |
+
},
|
| 329 |
+
"model.layers.13.self_attn.q_proj": {
|
| 330 |
+
"bits": 16,
|
| 331 |
+
"data_type": "float",
|
| 332 |
+
"act_bits": 16,
|
| 333 |
+
"act_data_type": "float"
|
| 334 |
+
},
|
| 335 |
+
"model.layers.13.self_attn.k_proj": {
|
| 336 |
+
"bits": 16,
|
| 337 |
+
"data_type": "float",
|
| 338 |
+
"act_bits": 16,
|
| 339 |
+
"act_data_type": "float"
|
| 340 |
+
},
|
| 341 |
+
"model.layers.13.self_attn.v_proj": {
|
| 342 |
+
"bits": 16,
|
| 343 |
+
"data_type": "float",
|
| 344 |
+
"act_bits": 16,
|
| 345 |
+
"act_data_type": "float"
|
| 346 |
+
},
|
| 347 |
+
"model.layers.13.self_attn.o_proj": {
|
| 348 |
+
"bits": 16,
|
| 349 |
+
"data_type": "float",
|
| 350 |
+
"act_bits": 16,
|
| 351 |
+
"act_data_type": "float"
|
| 352 |
+
},
|
| 353 |
+
"model.layers.14.self_attn.q_proj": {
|
| 354 |
+
"bits": 16,
|
| 355 |
+
"data_type": "float",
|
| 356 |
+
"act_bits": 16,
|
| 357 |
+
"act_data_type": "float"
|
| 358 |
+
},
|
| 359 |
+
"model.layers.14.self_attn.k_proj": {
|
| 360 |
+
"bits": 16,
|
| 361 |
+
"data_type": "float",
|
| 362 |
+
"act_bits": 16,
|
| 363 |
+
"act_data_type": "float"
|
| 364 |
+
},
|
| 365 |
+
"model.layers.14.self_attn.v_proj": {
|
| 366 |
+
"bits": 16,
|
| 367 |
+
"data_type": "float",
|
| 368 |
+
"act_bits": 16,
|
| 369 |
+
"act_data_type": "float"
|
| 370 |
+
},
|
| 371 |
+
"model.layers.14.self_attn.o_proj": {
|
| 372 |
+
"bits": 16,
|
| 373 |
+
"data_type": "float",
|
| 374 |
+
"act_bits": 16,
|
| 375 |
+
"act_data_type": "float"
|
| 376 |
+
},
|
| 377 |
+
"model.layers.15.self_attn.q_proj": {
|
| 378 |
+
"bits": 16,
|
| 379 |
+
"data_type": "float",
|
| 380 |
+
"act_bits": 16,
|
| 381 |
+
"act_data_type": "float"
|
| 382 |
+
},
|
| 383 |
+
"model.layers.15.self_attn.k_proj": {
|
| 384 |
+
"bits": 16,
|
| 385 |
+
"data_type": "float",
|
| 386 |
+
"act_bits": 16,
|
| 387 |
+
"act_data_type": "float"
|
| 388 |
+
},
|
| 389 |
+
"model.layers.15.self_attn.v_proj": {
|
| 390 |
+
"bits": 16,
|
| 391 |
+
"data_type": "float",
|
| 392 |
+
"act_bits": 16,
|
| 393 |
+
"act_data_type": "float"
|
| 394 |
+
},
|
| 395 |
+
"model.layers.15.self_attn.o_proj": {
|
| 396 |
+
"bits": 16,
|
| 397 |
+
"data_type": "float",
|
| 398 |
+
"act_bits": 16,
|
| 399 |
+
"act_data_type": "float"
|
| 400 |
+
},
|
| 401 |
+
"model.layers.16.self_attn.q_proj": {
|
| 402 |
+
"bits": 16,
|
| 403 |
+
"data_type": "float",
|
| 404 |
+
"act_bits": 16,
|
| 405 |
+
"act_data_type": "float"
|
| 406 |
+
},
|
| 407 |
+
"model.layers.16.self_attn.k_proj": {
|
| 408 |
+
"bits": 16,
|
| 409 |
+
"data_type": "float",
|
| 410 |
+
"act_bits": 16,
|
| 411 |
+
"act_data_type": "float"
|
| 412 |
+
},
|
| 413 |
+
"model.layers.16.self_attn.v_proj": {
|
| 414 |
+
"bits": 16,
|
| 415 |
+
"data_type": "float",
|
| 416 |
+
"act_bits": 16,
|
| 417 |
+
"act_data_type": "float"
|
| 418 |
+
},
|
| 419 |
+
"model.layers.16.self_attn.o_proj": {
|
| 420 |
+
"bits": 16,
|
| 421 |
+
"data_type": "float",
|
| 422 |
+
"act_bits": 16,
|
| 423 |
+
"act_data_type": "float"
|
| 424 |
+
},
|
| 425 |
+
"model.layers.17.self_attn.q_proj": {
|
| 426 |
+
"bits": 16,
|
| 427 |
+
"data_type": "float",
|
| 428 |
+
"act_bits": 16,
|
| 429 |
+
"act_data_type": "float"
|
| 430 |
+
},
|
| 431 |
+
"model.layers.17.self_attn.k_proj": {
|
| 432 |
+
"bits": 16,
|
| 433 |
+
"data_type": "float",
|
| 434 |
+
"act_bits": 16,
|
| 435 |
+
"act_data_type": "float"
|
| 436 |
+
},
|
| 437 |
+
"model.layers.17.self_attn.v_proj": {
|
| 438 |
+
"bits": 16,
|
| 439 |
+
"data_type": "float",
|
| 440 |
+
"act_bits": 16,
|
| 441 |
+
"act_data_type": "float"
|
| 442 |
+
},
|
| 443 |
+
"model.layers.17.self_attn.o_proj": {
|
| 444 |
+
"bits": 16,
|
| 445 |
+
"data_type": "float",
|
| 446 |
+
"act_bits": 16,
|
| 447 |
+
"act_data_type": "float"
|
| 448 |
+
},
|
| 449 |
+
"model.layers.18.self_attn.q_proj": {
|
| 450 |
+
"bits": 16,
|
| 451 |
+
"data_type": "float",
|
| 452 |
+
"act_bits": 16,
|
| 453 |
+
"act_data_type": "float"
|
| 454 |
+
},
|
| 455 |
+
"model.layers.18.self_attn.k_proj": {
|
| 456 |
+
"bits": 16,
|
| 457 |
+
"data_type": "float",
|
| 458 |
+
"act_bits": 16,
|
| 459 |
+
"act_data_type": "float"
|
| 460 |
+
},
|
| 461 |
+
"model.layers.18.self_attn.v_proj": {
|
| 462 |
+
"bits": 16,
|
| 463 |
+
"data_type": "float",
|
| 464 |
+
"act_bits": 16,
|
| 465 |
+
"act_data_type": "float"
|
| 466 |
+
},
|
| 467 |
+
"model.layers.18.self_attn.o_proj": {
|
| 468 |
+
"bits": 16,
|
| 469 |
+
"data_type": "float",
|
| 470 |
+
"act_bits": 16,
|
| 471 |
+
"act_data_type": "float"
|
| 472 |
+
},
|
| 473 |
+
"model.layers.19.self_attn.q_proj": {
|
| 474 |
+
"bits": 16,
|
| 475 |
+
"data_type": "float",
|
| 476 |
+
"act_bits": 16,
|
| 477 |
+
"act_data_type": "float"
|
| 478 |
+
},
|
| 479 |
+
"model.layers.19.self_attn.k_proj": {
|
| 480 |
+
"bits": 16,
|
| 481 |
+
"data_type": "float",
|
| 482 |
+
"act_bits": 16,
|
| 483 |
+
"act_data_type": "float"
|
| 484 |
+
},
|
| 485 |
+
"model.layers.19.self_attn.v_proj": {
|
| 486 |
+
"bits": 16,
|
| 487 |
+
"data_type": "float",
|
| 488 |
+
"act_bits": 16,
|
| 489 |
+
"act_data_type": "float"
|
| 490 |
+
},
|
| 491 |
+
"model.layers.19.self_attn.o_proj": {
|
| 492 |
+
"bits": 16,
|
| 493 |
+
"data_type": "float",
|
| 494 |
+
"act_bits": 16,
|
| 495 |
+
"act_data_type": "float"
|
| 496 |
+
},
|
| 497 |
+
"model.layers.20.self_attn.q_proj": {
|
| 498 |
+
"bits": 16,
|
| 499 |
+
"data_type": "float",
|
| 500 |
+
"act_bits": 16,
|
| 501 |
+
"act_data_type": "float"
|
| 502 |
+
},
|
| 503 |
+
"model.layers.20.self_attn.k_proj": {
|
| 504 |
+
"bits": 16,
|
| 505 |
+
"data_type": "float",
|
| 506 |
+
"act_bits": 16,
|
| 507 |
+
"act_data_type": "float"
|
| 508 |
+
},
|
| 509 |
+
"model.layers.20.self_attn.v_proj": {
|
| 510 |
+
"bits": 16,
|
| 511 |
+
"data_type": "float",
|
| 512 |
+
"act_bits": 16,
|
| 513 |
+
"act_data_type": "float"
|
| 514 |
+
},
|
| 515 |
+
"model.layers.20.self_attn.o_proj": {
|
| 516 |
+
"bits": 16,
|
| 517 |
+
"data_type": "float",
|
| 518 |
+
"act_bits": 16,
|
| 519 |
+
"act_data_type": "float"
|
| 520 |
+
},
|
| 521 |
+
"model.layers.21.self_attn.q_proj": {
|
| 522 |
+
"bits": 16,
|
| 523 |
+
"data_type": "float",
|
| 524 |
+
"act_bits": 16,
|
| 525 |
+
"act_data_type": "float"
|
| 526 |
+
},
|
| 527 |
+
"model.layers.21.self_attn.k_proj": {
|
| 528 |
+
"bits": 16,
|
| 529 |
+
"data_type": "float",
|
| 530 |
+
"act_bits": 16,
|
| 531 |
+
"act_data_type": "float"
|
| 532 |
+
},
|
| 533 |
+
"model.layers.21.self_attn.v_proj": {
|
| 534 |
+
"bits": 16,
|
| 535 |
+
"data_type": "float",
|
| 536 |
+
"act_bits": 16,
|
| 537 |
+
"act_data_type": "float"
|
| 538 |
+
},
|
| 539 |
+
"model.layers.21.self_attn.o_proj": {
|
| 540 |
+
"bits": 16,
|
| 541 |
+
"data_type": "float",
|
| 542 |
+
"act_bits": 16,
|
| 543 |
+
"act_data_type": "float"
|
| 544 |
+
},
|
| 545 |
+
"model.layers.22.self_attn.q_proj": {
|
| 546 |
+
"bits": 16,
|
| 547 |
+
"data_type": "float",
|
| 548 |
+
"act_bits": 16,
|
| 549 |
+
"act_data_type": "float"
|
| 550 |
+
},
|
| 551 |
+
"model.layers.22.self_attn.k_proj": {
|
| 552 |
+
"bits": 16,
|
| 553 |
+
"data_type": "float",
|
| 554 |
+
"act_bits": 16,
|
| 555 |
+
"act_data_type": "float"
|
| 556 |
+
},
|
| 557 |
+
"model.layers.22.self_attn.v_proj": {
|
| 558 |
+
"bits": 16,
|
| 559 |
+
"data_type": "float",
|
| 560 |
+
"act_bits": 16,
|
| 561 |
+
"act_data_type": "float"
|
| 562 |
+
},
|
| 563 |
+
"model.layers.22.self_attn.o_proj": {
|
| 564 |
+
"bits": 16,
|
| 565 |
+
"data_type": "float",
|
| 566 |
+
"act_bits": 16,
|
| 567 |
+
"act_data_type": "float"
|
| 568 |
+
},
|
| 569 |
+
"model.layers.23.self_attn.q_proj": {
|
| 570 |
+
"bits": 16,
|
| 571 |
+
"data_type": "float",
|
| 572 |
+
"act_bits": 16,
|
| 573 |
+
"act_data_type": "float"
|
| 574 |
+
},
|
| 575 |
+
"model.layers.23.self_attn.k_proj": {
|
| 576 |
+
"bits": 16,
|
| 577 |
+
"data_type": "float",
|
| 578 |
+
"act_bits": 16,
|
| 579 |
+
"act_data_type": "float"
|
| 580 |
+
},
|
| 581 |
+
"model.layers.23.self_attn.v_proj": {
|
| 582 |
+
"bits": 16,
|
| 583 |
+
"data_type": "float",
|
| 584 |
+
"act_bits": 16,
|
| 585 |
+
"act_data_type": "float"
|
| 586 |
+
},
|
| 587 |
+
"model.layers.23.self_attn.o_proj": {
|
| 588 |
+
"bits": 16,
|
| 589 |
+
"data_type": "float",
|
| 590 |
+
"act_bits": 16,
|
| 591 |
+
"act_data_type": "float"
|
| 592 |
+
},
|
| 593 |
+
"model.layers.24.self_attn.q_proj": {
|
| 594 |
+
"bits": 16,
|
| 595 |
+
"data_type": "float",
|
| 596 |
+
"act_bits": 16,
|
| 597 |
+
"act_data_type": "float"
|
| 598 |
+
},
|
| 599 |
+
"model.layers.24.self_attn.k_proj": {
|
| 600 |
+
"bits": 16,
|
| 601 |
+
"data_type": "float",
|
| 602 |
+
"act_bits": 16,
|
| 603 |
+
"act_data_type": "float"
|
| 604 |
+
},
|
| 605 |
+
"model.layers.24.self_attn.v_proj": {
|
| 606 |
+
"bits": 16,
|
| 607 |
+
"data_type": "float",
|
| 608 |
+
"act_bits": 16,
|
| 609 |
+
"act_data_type": "float"
|
| 610 |
+
},
|
| 611 |
+
"model.layers.24.self_attn.o_proj": {
|
| 612 |
+
"bits": 16,
|
| 613 |
+
"data_type": "float",
|
| 614 |
+
"act_bits": 16,
|
| 615 |
+
"act_data_type": "float"
|
| 616 |
+
},
|
| 617 |
+
"model.layers.25.self_attn.q_proj": {
|
| 618 |
+
"bits": 16,
|
| 619 |
+
"data_type": "float",
|
| 620 |
+
"act_bits": 16,
|
| 621 |
+
"act_data_type": "float"
|
| 622 |
+
},
|
| 623 |
+
"model.layers.25.self_attn.k_proj": {
|
| 624 |
+
"bits": 16,
|
| 625 |
+
"data_type": "float",
|
| 626 |
+
"act_bits": 16,
|
| 627 |
+
"act_data_type": "float"
|
| 628 |
+
},
|
| 629 |
+
"model.layers.25.self_attn.v_proj": {
|
| 630 |
+
"bits": 16,
|
| 631 |
+
"data_type": "float",
|
| 632 |
+
"act_bits": 16,
|
| 633 |
+
"act_data_type": "float"
|
| 634 |
+
},
|
| 635 |
+
"model.layers.25.self_attn.o_proj": {
|
| 636 |
+
"bits": 16,
|
| 637 |
+
"data_type": "float",
|
| 638 |
+
"act_bits": 16,
|
| 639 |
+
"act_data_type": "float"
|
| 640 |
+
},
|
| 641 |
+
"model.layers.26.self_attn.q_proj": {
|
| 642 |
+
"bits": 16,
|
| 643 |
+
"data_type": "float",
|
| 644 |
+
"act_bits": 16,
|
| 645 |
+
"act_data_type": "float"
|
| 646 |
+
},
|
| 647 |
+
"model.layers.26.self_attn.k_proj": {
|
| 648 |
+
"bits": 16,
|
| 649 |
+
"data_type": "float",
|
| 650 |
+
"act_bits": 16,
|
| 651 |
+
"act_data_type": "float"
|
| 652 |
+
},
|
| 653 |
+
"model.layers.26.self_attn.v_proj": {
|
| 654 |
+
"bits": 16,
|
| 655 |
+
"data_type": "float",
|
| 656 |
+
"act_bits": 16,
|
| 657 |
+
"act_data_type": "float"
|
| 658 |
+
},
|
| 659 |
+
"model.layers.26.self_attn.o_proj": {
|
| 660 |
+
"bits": 16,
|
| 661 |
+
"data_type": "float",
|
| 662 |
+
"act_bits": 16,
|
| 663 |
+
"act_data_type": "float"
|
| 664 |
+
},
|
| 665 |
+
"model.layers.27.self_attn.q_proj": {
|
| 666 |
+
"bits": 16,
|
| 667 |
+
"data_type": "float",
|
| 668 |
+
"act_bits": 16,
|
| 669 |
+
"act_data_type": "float"
|
| 670 |
+
},
|
| 671 |
+
"model.layers.27.self_attn.k_proj": {
|
| 672 |
+
"bits": 16,
|
| 673 |
+
"data_type": "float",
|
| 674 |
+
"act_bits": 16,
|
| 675 |
+
"act_data_type": "float"
|
| 676 |
+
},
|
| 677 |
+
"model.layers.27.self_attn.v_proj": {
|
| 678 |
+
"bits": 16,
|
| 679 |
+
"data_type": "float",
|
| 680 |
+
"act_bits": 16,
|
| 681 |
+
"act_data_type": "float"
|
| 682 |
+
},
|
| 683 |
+
"model.layers.27.self_attn.o_proj": {
|
| 684 |
+
"bits": 16,
|
| 685 |
+
"data_type": "float",
|
| 686 |
+
"act_bits": 16,
|
| 687 |
+
"act_data_type": "float"
|
| 688 |
+
},
|
| 689 |
+
".*self_attn.*": {
|
| 690 |
+
"bits": 16,
|
| 691 |
+
"data_type": "float",
|
| 692 |
+
"act_bits": 16,
|
| 693 |
+
"act_data_type": "float"
|
| 694 |
+
}
|
| 695 |
+
}
|
| 696 |
+
}
|
tokenizer.json
ADDED
|
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:f624f8136bc57382156e7595e8ea404a66fbce12e3181b933ec2459bda3a71be
|
| 3 |
+
size 11421907
|
tokenizer_config.json
ADDED
|
@@ -0,0 +1,15 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
{
|
| 2 |
+
"add_prefix_space": null,
|
| 3 |
+
"backend": "tokenizers",
|
| 4 |
+
"bos_token": "<｜begin▁of▁sentence｜>",
|
| 5 |
+
"clean_up_tokenization_spaces": false,
|
| 6 |
+
"eos_token": "<｜end▁of▁sentence｜>",
|
| 7 |
+
"is_local": false,
|
| 8 |
+
"legacy": true,
|
| 9 |
+
"local_files_only": false,
|
| 10 |
+
"model_max_length": 16384,
|
| 11 |
+
"pad_token": "<｜end▁of▁sentence｜>",
|
| 12 |
+
"sp_model_kwargs": {},
|
| 13 |
+
"tokenizer_class": "Qwen2Tokenizer",
|
| 14 |
+
"unk_token": null
|
| 15 |
+
}
|