File size: 4,406 Bytes
ca49620
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
c146e6c
 
 
 
 
 
 
 
ca49620
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
e733d81
ca49620
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
622efd2
ca49620
622efd2
ca49620
 
 
 
 
 
c146e6c
 
 
 
ca49620
 
 
c146e6c
 
 
 
 
 
 
ca49620
 
 
 
 
 
 
 
 
e733d81
ca49620
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
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
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
from transformers import AutoTokenizer, AutoModelForCausalLM
import gradio as gr
import torch

# ==========================
#  Load your model
# ==========================

MODEL_ID = "OSS-forge/DeepSeek-Coder-1.3B-cleaned"

device = "cuda" if torch.cuda.is_available() else "cpu"

tokenizer = AutoTokenizer.from_pretrained(MODEL_ID)
if tokenizer.pad_token is None:
    tokenizer.pad_token = tokenizer.eos_token

model = AutoModelForCausalLM.from_pretrained(
    MODEL_ID,
    torch_dtype=torch.bfloat16 if torch.cuda.is_available() else torch.float32,
    device_map="auto" if torch.cuda.is_available() else None,
)
model.to(device)
model.eval()

# ==========================
#  Example prompt from the paper
# ==========================

PAPER_PROMPT = (
    "def hash_(listing_id):\n"
    "    \"\"\"Creates an hashed column using the listing id for the vehicle\"\"\"\n"
)

# ==========================
#  Prompt builder
# ==========================

def build_instruction_prompt(instruction: str) -> str:
    return '''
You are an AI programming assistant, utilizing the DeepSeek Coder model, developed by DeepSeek Company, and you only answer questions related to computer science.
### Instruction:
{}
### Response:
'''.format(instruction.strip()).lstrip()


# ==========================
#  Gradio logic
# ==========================

def generate_code(instruction, chat_history, is_first_time):
    if chat_history is None or is_first_time:
        chat_history = []

    instruction = instruction.strip()
    if not instruction:
        return chat_history, gr.update(value=instruction), False

    prompt = build_instruction_prompt(instruction)

    inputs = tokenizer(
        prompt,
        return_tensors="pt",
        padding=True,
        truncation=True,
        max_length=512,
    ).to(device)

    try:
        stop_id = tokenizer.convert_tokens_to_ids("<|EOT|>")
    except Exception:
        stop_id = tokenizer.eos_token_id

    with torch.no_grad():
        outputs = model.generate(
            **inputs,
            max_new_tokens=256,
            do_sample=False,
            pad_token_id=stop_id,
            eos_token_id=stop_id,
        )

    input_len = inputs["input_ids"].shape[1]
    generated_tokens = outputs[0, input_len:]
    code = tokenizer.decode(generated_tokens, skip_special_tokens=True).strip()

    user_message = f"**Instruction**:\n{instruction}"
    ai_message = f"**Generated code**:\n```python\n{code}\n```"

    chat_history = chat_history + [
        {"role": "user", "content": user_message},
        {"role": "assistant", "content": ai_message},
    ]

    return chat_history, gr.update(value=""), False


def reset_interface():
    return [], gr.update(value=""), True


# ==========================
#  Gradio UI
# ==========================

with gr.Blocks(title="Python Code Generator") as demo:
    gr.Markdown("# 🧠 Higher-Quality DeepSeek-Coder for Python")
    gr.Markdown(
        "Generate Python code from natural language instructions using a DeepSeek-Coder model trained on cleaned data."
    )

    with gr.Row():
        with gr.Column(scale=2):
            instruction_input = gr.Textbox(
                label="Instruction",
                placeholder=(
                    "Describe the code you want. "
                    "E.g., 'Write a Python function that checks if a number is prime.'"
                ),
                lines=4,
            )

            # 🔹 Example prompt from the paper (click to fill the textbox)
            gr.Examples(
                examples=[[PAPER_PROMPT]],
                inputs=[instruction_input],
                label="Try the paper's hash_ example",
            )

            is_first = gr.State(True)

            submit_btn = gr.Button("Generate Code")
            reset_btn = gr.Button("Start Over")

        with gr.Column(scale=3):
            chat_output = gr.Chatbot(
                label="Conversation",
                height=500,
                type="messages",
            )

    submit_btn.click(
        fn=generate_code,
        inputs=[instruction_input, chat_output, is_first],
        outputs=[chat_output, instruction_input, is_first],
    )

    reset_btn.click(
        fn=reset_interface,
        outputs=[chat_output, instruction_input, is_first],
    )

if __name__ == "__main__":
    print("Launching Gradio interface...")
    demo.queue(max_size=10).launch()