Text Generation
Transformers
PyTorch
Chinese
English
llama
text2text-generation
text-generation-inference
Instructions to use ClueAI/ChatYuan-7B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use ClueAI/ChatYuan-7B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="ClueAI/ChatYuan-7B")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("ClueAI/ChatYuan-7B") model = AutoModelForCausalLM.from_pretrained("ClueAI/ChatYuan-7B", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use ClueAI/ChatYuan-7B with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "ClueAI/ChatYuan-7B" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "ClueAI/ChatYuan-7B", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/ClueAI/ChatYuan-7B
- SGLang
How to use ClueAI/ChatYuan-7B with SGLang:
Install from pip and serve model
# Install SGLang from pip: pip install sglang # Start the SGLang server: python3 -m sglang.launch_server \ --model-path "ClueAI/ChatYuan-7B" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "ClueAI/ChatYuan-7B", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker images
docker run --gpus all \ --shm-size 32g \ -p 30000:30000 \ -v ~/.cache/huggingface:/root/.cache/huggingface \ --env "HF_TOKEN=<secret>" \ --ipc=host \ lmsysorg/sglang:latest \ python3 -m sglang.launch_server \ --model-path "ClueAI/ChatYuan-7B" \ --host 0.0.0.0 \ --port 30000 # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:30000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "ClueAI/ChatYuan-7B", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use ClueAI/ChatYuan-7B with Docker Model Runner:
docker model run hf.co/ClueAI/ChatYuan-7B
Update README.md
Browse files
README.md
CHANGED
|
@@ -16,7 +16,7 @@ ChatYuan-7B是一个支持中英双语的功能型对话语言大模型。它是
|
|
| 16 |
## 使用方式
|
| 17 |
为了遵守LLaMA模型许可证,我们将ChatYuan-7B权重发布为增量权重。您可以将我们的增量权重与原始的LLaMA权重相加,得到ChatYuan-7B权重。
|
| 18 |
|
| 19 |
-
1. 通过原始[LLaMA-7B](https://github.com/facebookresearch/llama)生成LLaMA的hf模型(LLaMA-7B-HF),可以参考[指导](https://huggingface.co/docs/transformers/main/model_doc/llama)
|
| 20 |
2. 合并LLaMA-7B的hf模型和ChatYuan-7B模型
|
| 21 |
### 合并脚本
|
| 22 |
```shell
|
|
@@ -38,7 +38,7 @@ tokenizer = AutoTokenizer.from_pretrained(ckpt)
|
|
| 38 |
## 推理方式
|
| 39 |
|
| 40 |
```python
|
| 41 |
-
prompt = "用户:
|
| 42 |
input_ids = tokenizer(prompt, return_tensors="pt").input_ids.to(device)
|
| 43 |
generate_ids = model.generate(input_ids, max_new_tokens=1024, do_sample = True, temperature = 0.7)
|
| 44 |
output = tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
|
|
|
|
| 16 |
## 使用方式
|
| 17 |
为了遵守LLaMA模型许可证,我们将ChatYuan-7B权重发布为增量权重。您可以将我们的增量权重与原始的LLaMA权重相加,得到ChatYuan-7B权重。
|
| 18 |
|
| 19 |
+
1. 通过原始[LLaMA-7B](https://github.com/facebookresearch/llama)生成LLaMA的hf模型(LLaMA-7B-HF),可以参考[指导](https://huggingface.co/docs/transformers/main/model_doc/llama), 也可以直接使用[llama-7b-hf](https://huggingface.co/decapoda-research/llama-7b-hf)
|
| 20 |
2. 合并LLaMA-7B的hf模型和ChatYuan-7B模型
|
| 21 |
### 合并脚本
|
| 22 |
```shell
|
|
|
|
| 38 |
## 推理方式
|
| 39 |
|
| 40 |
```python
|
| 41 |
+
prompt = "用户:怎么让自己精力充沛,列5点建议\n小元: "
|
| 42 |
input_ids = tokenizer(prompt, return_tensors="pt").input_ids.to(device)
|
| 43 |
generate_ids = model.generate(input_ids, max_new_tokens=1024, do_sample = True, temperature = 0.7)
|
| 44 |
output = tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
|