File size: 2,090 Bytes
c311821
9b2c83e
c311821
a447070
9b2c83e
890d27d
9b2c83e
c311821
9b2c83e
 
 
df78555
c311821
df78555
9b2c83e
c311821
 
 
 
 
 
 
 
 
 
 
 
 
 
7665895
 
 
df78555
7665895
c311821
 
 
 
 
9b2c83e
cb9cef0
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
4239ac9
cb9cef0
 
4239ac9
cb9cef0
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
import spaces
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
import gradio as gr

model_name = "Qwen/Qwen2.5-Coder-14B-Instruct"

# Load model and tokenizer (outside the function for efficiency)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    torch_dtype="auto",
   # device_map="auto",
    trust_remote_code=True # Add this line for Qwen models
).to('cuda')

tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True) # Add this line for Qwen models

@spaces.GPU(required=True)
def generate_code(prompt):
    messages = [
        {"role": "system", "content": "You are Qwen, created by Alibaba Cloud. You are a helpful assistant."},
        {"role": "user", "content": prompt}
    ]
    text = tokenizer.apply_chat_template(
        messages,
        tokenize=False,
        add_generation_prompt=True
    )
    model_inputs = tokenizer([text], return_tensors="pt").to(model.device)
    with torch.no_grad():
        generated_ids = model.generate(
            **model_inputs,
            max_new_tokens=768
        )
    generated_ids = [
        output_ids[len(input_ids):] for input_ids, output_ids in zip(model_inputs.input_ids, generated_ids)
    ]
    response = tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
    return response

with gr.Blocks(title="Qwen 14b") as demo:  # Updated title
    with gr.Tab("Code Chat"):
        run_button = gr.Button("Run", scale=0)
        prompt = gr.Text(
            label="Prompt",
            show_label=False,
            max_lines=1,
            placeholder="Enter your prompt",
            container=False,
        )
        result = gr.Text(
            label="Result",
            show_label=False,
            max_lines=100,
            container=False,
        )
        gr.on(
            triggers=[
                run_button.click,
            ],
      #  api_name="generate",  # Add this line
            fn=generate_code,
            inputs=[
                prompt,
            ],
            outputs=[result],
        )

demo.launch(share=False)