Update app.py
Browse files
app.py
CHANGED
@@ -3,7 +3,7 @@ from transformers import AutoTokenizer, AutoModelForCausalLM
|
|
3 |
import torch
|
4 |
import spaces
|
5 |
|
6 |
-
|
7 |
model_name = "Zhihu-ai/Zhi-writing-dsr1-14b"
|
8 |
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
|
9 |
model = AutoModelForCausalLM.from_pretrained(
|
@@ -12,10 +12,9 @@ model = AutoModelForCausalLM.from_pretrained(
|
|
12 |
device_map="auto",
|
13 |
trust_remote_code=True
|
14 |
)
|
15 |
-
|
16 |
@spaces.GPU()
|
17 |
def predict(message, history):
|
18 |
-
|
19 |
history_text = ""
|
20 |
for human, assistant in history:
|
21 |
history_text += f"Human: {human}\nAssistant: {assistant}\n"
|
@@ -23,40 +22,29 @@ def predict(message, history):
|
|
23 |
|
24 |
# 生成回复
|
25 |
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
|
26 |
-
|
27 |
-
# 使用流式生成
|
28 |
-
for response in model.generate(
|
29 |
**inputs,
|
30 |
max_new_tokens=10000,
|
31 |
do_sample=True,
|
32 |
temperature=0.7,
|
33 |
top_p=0.9,
|
34 |
repetition_penalty=1.1,
|
35 |
-
pad_token_id=tokenizer.eos_token_id
|
36 |
-
|
37 |
-
)
|
38 |
-
yield response.strip()
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
with gr.Blocks(theme=gr.themes.Soft()) as demo:
|
43 |
-
gr.Markdown("# Zhi-writing-dsr1-14")
|
44 |
-
gr.Markdown("这是一个基于Zhi-writing-dsr1-14的文章生成器")
|
45 |
|
46 |
-
|
47 |
-
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
|
52 |
-
|
53 |
-
|
54 |
-
|
55 |
-
|
56 |
-
return "", chat_history
|
57 |
-
|
58 |
-
msg.submit(respond, [msg, chatbot], [msg, chatbot])
|
59 |
-
clear.click(lambda: None, None, chatbot, queue=False)
|
60 |
|
61 |
if __name__ == "__main__":
|
62 |
demo.launch(share=True)
|
|
|
|
|
|
3 |
import torch
|
4 |
import spaces
|
5 |
|
6 |
+
# 加载模型和分词器
|
7 |
model_name = "Zhihu-ai/Zhi-writing-dsr1-14b"
|
8 |
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)
|
9 |
model = AutoModelForCausalLM.from_pretrained(
|
|
|
12 |
device_map="auto",
|
13 |
trust_remote_code=True
|
14 |
)
|
|
|
15 |
@spaces.GPU()
|
16 |
def predict(message, history):
|
17 |
+
# 构建输入
|
18 |
history_text = ""
|
19 |
for human, assistant in history:
|
20 |
history_text += f"Human: {human}\nAssistant: {assistant}\n"
|
|
|
22 |
|
23 |
# 生成回复
|
24 |
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
|
25 |
+
outputs = model.generate(
|
|
|
|
|
26 |
**inputs,
|
27 |
max_new_tokens=10000,
|
28 |
do_sample=True,
|
29 |
temperature=0.7,
|
30 |
top_p=0.9,
|
31 |
repetition_penalty=1.1,
|
32 |
+
pad_token_id=tokenizer.eos_token_id
|
33 |
+
)
|
34 |
+
response = tokenizer.decode(outputs[0][inputs.input_ids.shape[1]:], skip_special_tokens=True)
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
35 |
|
36 |
+
return response.strip()
|
37 |
+
|
38 |
+
# 创建Gradio界面
|
39 |
+
demo = gr.ChatInterface(
|
40 |
+
predict,
|
41 |
+
title="测试Zhi-writing-dsr1-14b",
|
42 |
+
description="Zhihu-ai/Zhi-writing-dsr1-14b",
|
43 |
+
examples=["鲁迅口吻写五百字,描述桔猫的可爱!", "桔了个仔是谁", "介绍自己"],
|
44 |
+
theme=gr.themes.Soft()
|
45 |
+
)
|
|
|
|
|
|
|
|
|
46 |
|
47 |
if __name__ == "__main__":
|
48 |
demo.launch(share=True)
|
49 |
+
|
50 |
+
|