Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,45 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
# llm-jp-13b-instruct-full-dolly-oasst-v1.0-GPTQ-calib-ja-1k
|
| 6 |
+
llm-jpさんが公開している、[llm-jp-13b-instruct-full-dolly-oasst-v1.0](https://huggingface.co/llm-jp/llm-jp-13b-instruct-full-dolly-oasst-v1.0)を、
|
| 7 |
+
日本語のキャリブレーションセットで生成したGPTQモデルになります。
|
| 8 |
+
|
| 9 |
+
キャリブレーションセットは[izumi-lab/wikipedia-ja-20230720](https://huggingface.co/datasets/izumi-lab/wikipedia-ja-20230720)から、
|
| 10 |
+
1kほどランダムサンプリングしたものと、
|
| 11 |
+
[ELYZA-tasks-100](https://huggingface.co/datasets/elyza/ELYZA-tasks-100)のinput/outputを計200ほど追加しています。
|
| 12 |
+
[mmnga/wikipedia-ja-20230720-1k](https://huggingface.co/datasets/mmnga/wikipedia-ja-20230720-1k)
|
| 13 |
+
|
| 14 |
+
モデル一覧
|
| 15 |
+
[mmnga/llm-jp-13b-v1.0-4bit-g128-GPTQ-calib-ja-1k](https://huggingface.co/mmnga/llm-jp-13b-v1.0-4bit-g128-GPTQ-calib-ja-1k)
|
| 16 |
+
[mmnga/llm-jp-13b-instruct-full-jaster-dolly-oasst-v1.0-GPTQ-calib-ja-1k](https://huggingface.co/mmnga/llm-jp-13b-instruct-full-jaster-dolly-oasst-v1.0-GPTQ-calib-ja-1k)
|
| 17 |
+
[mmnga/llm-jp-13b-instruct-full-dolly-oasst-v1.0-GPTQ-calib-ja-1k](https://huggingface.co/mmnga/llm-jp-13b-instruct-full-dolly-oasst-v1.0-GPTQ-calib-ja-1k)
|
| 18 |
+
|
| 19 |
+
GGUF版
|
| 20 |
+
[mmnga/llm-jp-13b-v1.0-gguf](https://huggingface.co/mmnga/llm-jp-13b-v1.0-gguf)
|
| 21 |
+
[mmnga/llm-jp-13b-instruct-full-jaster-dolly-oasst-v1.0-gguf](https://huggingface.co/mmnga/llm-jp-13b-instruct-full-jaster-dolly-oasst-v1.0-gguf)
|
| 22 |
+
[mmnga/llm-jp-13b-instruct-full-dolly-oasst-v1.0-gguf](https://huggingface.co/mmnga/llm-jp-13b-instruct-full-dolly-oasst-v1.0-gguf)
|
| 23 |
+
|
| 24 |
+
# Usage
|
| 25 |
+
|
| 26 |
+
~~~Bash
|
| 27 |
+
pip install auto-gptq transformers
|
| 28 |
+
~~~
|
| 29 |
+
|
| 30 |
+
~~~python
|
| 31 |
+
from auto_gptq import AutoGPTQForCausalLM, BaseQuantizeConfig
|
| 32 |
+
from transformers import AutoTokenizer
|
| 33 |
+
|
| 34 |
+
model_name_or_path = "mmnga/llm-jp-13b-instruct-full-jaster-dolly-oasst-v1.0-GPTQ-calib-ja-1k"
|
| 35 |
+
|
| 36 |
+
# Tokenizer
|
| 37 |
+
tokenizer = AutoTokenizer.from_pretrained(model_name_or_path, trust_remote_code=True)
|
| 38 |
+
|
| 39 |
+
# Model
|
| 40 |
+
model = AutoGPTQForCausalLM.from_quantized(model_name_or_path, use_safetensors=True, device="cuda:0", use_auth_token=False)
|
| 41 |
+
|
| 42 |
+
#Your test prompt
|
| 43 |
+
prompt = """今日の晩御飯のレシピの作り方を教えて ### 回答："""
|
| 44 |
+
print(tokenizer.decode(model.generate(**tokenizer(prompt, return_tensors="pt",add_special_tokens=False).to(model.device), max_new_tokens=100,do_sample=True,top_p=0.95,temperature=0.7)[0]))
|
| 45 |
+
~~~
|