Spaces:
Build error
Build error
| import json | |
| import markdown | |
| from typing import Callable, List, Dict, Any, Generator | |
| from functools import partial | |
| import fastapi | |
| import uvicorn | |
| from fastapi import HTTPException, Depends, Request | |
| from fastapi.responses import HTMLResponse | |
| from fastapi.middleware.cors import CORSMiddleware | |
| from sse_starlette.sse import EventSourceResponse | |
| from anyio import create_memory_object_stream | |
| from anyio.to_thread import run_sync | |
| from ctransformers import AutoModelForCausalLM | |
| from pydantic import BaseModel | |
| llm = AutoModelForCausalLM.from_pretrained("TheBloke/WizardCoder-15B-1.0-GGML", | |
| model_file="WizardCoder-15B-1.0.ggmlv3.q4_0.bin", | |
| model_type="starcoder", | |
| threads=8) | |
| app = fastapi.FastAPI(title="🪄WizardCoder💫") | |
| app.add_middleware( | |
| CORSMiddleware, | |
| allow_origins=["*"], | |
| allow_credentials=True, | |
| allow_methods=["*"], | |
| allow_headers=["*"], | |
| ) | |
| async def index(): | |
| html_content = """ | |
| <html> | |
| <head> | |
| </head> | |
| <body style="font-family:system-ui"> | |
| <h2><a href="https://huggingface.co./TheBloke/WizardCoder-15B-1.0-GGML">wizardcoder-ggml</a></h2> | |
| <h2><a href="https://nerddisco-wizardcoder-15b-1-0-ggmlv3-q4-0.hf.space/docs">FastAPI Docs</a></h2> | |
| </body> | |
| </html> | |
| """ | |
| return HTMLResponse(content=html_content, status_code=200) | |
| class ChatCompletionRequestV0(BaseModel): | |
| prompt: str | |
| class Message(BaseModel): | |
| role: str | |
| content: str | |
| class ChatCompletionRequest(BaseModel): | |
| messages: List[Message] | |
| max_tokens: int = 1024 | |
| async def completion(request: ChatCompletionRequestV0, response_mode=None): | |
| response = llm(request.prompt) | |
| return response | |
| async def generate_response(chat_chunks, llm): | |
| for chat_chunk in chat_chunks: | |
| response = { | |
| 'choices': [ | |
| { | |
| 'message': { | |
| 'role': 'system', | |
| 'content': llm.detokenize(chat_chunk) | |
| }, | |
| 'finish_reason': 'stop' if llm.is_eos_token(chat_chunk) else 'unknown' | |
| } | |
| ] | |
| } | |
| yield dict(data=json.dumps(response)) | |
| yield dict(data="[DONE]") | |
| async def chat(request: ChatCompletionRequest): | |
| combined_messages = ' '.join([message.content for message in request.messages]) | |
| tokens = llm.tokenize(combined_messages) | |
| try: | |
| chat_chunks = llm.generate(tokens) | |
| except Exception as e: | |
| raise HTTPException(status_code=500, detail=str(e)) | |
| return EventSourceResponse(generate_response(chat_chunks, llm)) | |
| async def stream_response(tokens, llm): | |
| try: | |
| iterator: Generator = llm.generate(tokens) | |
| for chat_chunk in iterator: | |
| response = { | |
| 'choices': [ | |
| { | |
| 'message': { | |
| 'role': 'system', | |
| 'content': llm.detokenize(chat_chunk) | |
| }, | |
| 'finish_reason': 'stop' if llm.is_eos_token(chat_chunk) else 'unknown' | |
| } | |
| ] | |
| } | |
| yield dict(data=json.dumps(response)) | |
| yield dict(data="[DONE]") | |
| except Exception as e: | |
| print(f"Exception in event publisher: {str(e)}") | |
| async def chatV2_endpoint(request: Request, body: ChatCompletionRequest): | |
| combined_messages = ' '.join([message.content for message in body.messages]) | |
| tokens = llm.tokenize(combined_messages) | |
| return EventSourceResponse(stream_response(tokens, llm)) | |
| async def chat(request: ChatCompletionRequestV0, response_mode=None): | |
| tokens = llm.tokenize(request.prompt) | |
| async def server_sent_events(chat_chunks, llm): | |
| for chat_chunk in llm.generate(chat_chunks): | |
| yield dict(data=json.dumps(llm.detokenize(chat_chunk))) | |
| yield dict(data="[DONE]") | |
| return EventSourceResponse(server_sent_events(tokens, llm)) | |
| if __name__ == "__main__": | |
| uvicorn.run(app, host="0.0.0.0", port=8000) | |