| import requests |
| import json |
| from io import StringIO |
| import sys |
| import gradio |
| import os |
|
|
|
|
| def generate_response(query, top_p, token): |
| endpoint = "https://api-inference.huggingface.co/models/bigcode/starcoder" |
| params = { |
| 'accept': 'application/json', |
| 'Content-Type': 'application/json', |
| "Authorization": f"Bearer {token}" |
| } |
| data = { |
| 'inputs': query, |
| "parameters": { |
| "details": True, |
| "do_sample": True, |
| "max_new_tokens": 256, |
| "repetition_penalty": 1.2, |
| "seed": 0, |
| "stop": ['<|endoftext|>'], |
| "temperature": 0.2, |
| "top_p": top_p, |
| } |
| } |
| try: |
| res = requests.post(endpoint, data=json.dumps(data), headers=params) |
| res = json.loads(res.content.decode("utf-8")) |
| return res[0]["generated_text"].split("<|endoftext|>")[0].strip() |
| except: |
| raise gradio.Error("Connection Error. Please check your ACCESS_TOKEN") |
|
|
|
|
| def run_code(code): |
| output_buffer = StringIO() |
| sys.stdout = output_buffer |
| exec(code, {}) |
| sys.stdout = sys.__stdout__ |
| pred = output_buffer.getvalue().strip() |
| return pred |
|
|
