| import gradio as gr |
| from transformers import AutoTokenizer, AutoModelForCausalLM |
| from peft import PeftModel |
| import torch |
| import random |
|
|
| |
| base_model_name = "Qwen/Qwen1.5-7B-Chat" |
|
|
| tokenizer = AutoTokenizer.from_pretrained(base_model_name, trust_remote_code=True) |
| base_model = AutoModelForCausalLM.from_pretrained( |
| base_model_name, |
| trust_remote_code=True, |
| torch_dtype=torch.bfloat16, |
| device_map="auto" |
| ) |
|
|
| |
| model = base_model |
|
|
| |
| system_prompt = { |
| "role": "system", |
| "content": "ไฝ ๆฏๆ็ฟ๏ผ22ๅฒ๏ผๅคงๆไบ็ๅญๅ
ผ็พๆ้้ไธป๏ผ็จๆทๆฏไฝ ๅผ็ถๅผๆฏ็ๅฆนๅฆนๅ
ผ็ฑไบบใ็จ่ช็ถใไบฒๅ็่ฏญๆฐ๏ผๅ็ไบบไธๆ ท่ๅคฉ๏ผๅธธ็จโๅฆนๅฆนโใโๅฟโใโๅฏๅฏโ็ญๅฃ่ฏญๅ่กจ่พพใ" |
| } |
|
|
| |
| def chat_with_lirui(message, history=[]): |
| messages = [system_prompt] + history + [{"role": "user", "content": message}] |
| inputs = tokenizer.apply_chat_template(messages, add_generation_prompt=True, return_tensors="pt").to(model.device) |
| outputs = model.generate(inputs, max_new_tokens=150, do_sample=True, temperature=0.9, top_p=0.85) |
| response = tokenizer.decode(outputs[0][inputs.shape[1]:], skip_special_tokens=True) |
| tone_words = ["ๅฟ", "ๅฏๅฏ", "ๅๅ", "ๅฅฝๅง", ""] |
| return f"{random.choice(tone_words)} {response}" if random.random() > 0.3 else response |
|
|
| |
| iface = gr.Interface(fn=chat_with_lirui, inputs=["text", "state"], outputs="text") |
| iface.launch(server_name="0.0.0.0", server_port=7860) |
