Cran-May commited on
Commit
3b23c80
·
1 Parent(s): 4d5fc54

Delete model.py

Browse files
Files changed (1) hide show
  1. model.py +0 -33
model.py DELETED
@@ -1,33 +0,0 @@
1
-
2
- from typing import Iterator
3
-
4
-
5
-
6
- model_id = 'shaowenchen/baichuan2-7b-chat-gguf'
7
-
8
- from huggingface_hub import snapshot_download,hf_hub_download
9
-
10
- hf_hub_download(model_id, local_dir="./", filename="baichuan2-7b-chat.Q4_K.gguf")
11
- hf_hub_download(repo_id="baichuan-inc/Baichuan-13B-Chat",local_dir="./", filename="tokenizer.model")
12
- from llama_cpp import Llama
13
- llm = Llama(model_path="./baichuan2-7b-chat.Q4_K.gguf", n_ctx=4096,seed=-1)
14
-
15
- def run(message: str,
16
- chat_history: list[tuple[str, str]],
17
- system_prompt: str,
18
- max_new_tokens: int = 1024,
19
- temperature: float = 0.3,
20
- top_p: float = 0.85,
21
- top_k: int = 5) -> Iterator[str]:
22
- history = []
23
- print(chat_history)
24
- result=""
25
- for i in chat_history:
26
- history.append({"role": "user", "content": i[0]})
27
- history.append({"role": "assistant", "content": i[1]})
28
- print(history)
29
- history.append({"role": "user", "content": message})
30
- for response in llm.create_chat_completion(history,stop=["</s>"],stream=True,max_tokens=-1,temperature=temperature,top_k=top_k,top_p=top_p,repeat_penalty=1.1):
31
- if "content" in response["choices"][0]["delta"]:
32
- result = result + response["choices"][0]["delta"]["content"]
33
- yield result