Spaces:
Running
Running
File size: 3,781 Bytes
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 |
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()
# ==========================
# 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=512,
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("# 🧠 Python Code Generator")
gr.Markdown(
"Generate Python code from natural language instructions using your Hugging Face model."
)
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,
)
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,
)
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()
|