Create README.md

README.md

@@ -0,0 +1,36 @@
+---
+language:
+- zh
+- en
+pipeline_tag: text2text-generation
+---
+
+You can see more detail in this [repo](https://github.com/clue-ai/ChatYuan-7B)
+
+
+# How to use
+```python
+from transformers import LlamaForCausalLM, AutoTokenizer
+import torch
+
+ckpt = "tiansz/ChatYuan-7B-merge"
+device = torch.device('cuda')
+model = LlamaForCausalLM.from_pretrained(ckpt)
+tokenizer = AutoTokenizer.from_pretrained(ckpt)
+
+def answer(prompt):
+  prompt = f"用户：{prompt}\n小元："
+  input_ids = tokenizer(prompt, return_tensors="pt").input_ids.to(device)
+  generate_ids = model.generate(input_ids, max_new_tokens=1024, do_sample = True, temperature = 0.7)
+  output = tokenizer.batch_decode(generate_ids, skip_special_tokens=True, clean_up_tokenization_spaces=False)[0]
+  response = output[len(prompt):]
+  return response
+
+result = answer("你好")
+print(result)
+```
+
+
+# License
+- [ChatYuan-7B](https://github.com/clue-ai/ChatYuan-7B)
+- [llama](https://github.com/facebookresearch/llama)