serve_ministral.py

#!/usr/bin/env python3
"""
OpenAI-compatible API server for Ministral 14B with streaming support
"""

# Install dependencies first
import subprocess
import sys

def install_deps():
    # Check if torch with CUDA exists, only install if missing
    try:
        import torch
        need_torch = not torch.cuda.is_available()
    except ImportError:
        need_torch = True

    print("=== Installing dependencies ===")

    if need_torch:
        subprocess.check_call([sys.executable, "-m", "pip", "install", "-q", "torch"])

    # Install transformers from git for mistral3 support
    subprocess.check_call([sys.executable, "-m", "pip", "install", "-q",
                          "git+https://github.com/huggingface/transformers.git"])

    # Install other deps
    subprocess.check_call([sys.executable, "-m", "pip", "install", "-q",
                          "accelerate", "fastapi", "uvicorn", "pydantic", "sentencepiece", "protobuf"])

    print("=== Dependencies installed ===")

install_deps()

import torch
from transformers import AutoProcessor, AutoModelForImageTextToText, TextIteratorStreamer
from fastapi import FastAPI
from fastapi.responses import StreamingResponse
from pydantic import BaseModel
from typing import List, Optional
import uvicorn
import time
import traceback
import json
import asyncio
from threading import Thread

app = FastAPI()

def fix_bpe_tokens(text):
    """Fix BPE tokenization artifacts"""
    text = text.replace("Ġ", " ")
    text = text.replace("Ċ", "\n")
    text = text.replace("ĉ", "\t")
    text = text.replace("âĢĻ", "'")
    text = text.replace("âĢľ", '"')
    text = text.replace("âĢĿ", '"')
    text = text.replace("âĢĶ", "—")
    text = text.replace("âĢĵ", "–")
    text = text.replace("âĢ¦", "…")
    text = text.replace("âĢĺ", "'")
    return text

# Global model and tokenizer
model = None
processor = None

class Message(BaseModel):
    role: str
    content: str

class ChatRequest(BaseModel):
    model: str = "ministral-14b"
    messages: List[Message]
    max_tokens: Optional[int] = 2048
    temperature: Optional[float] = 0.7
    top_p: Optional[float] = 0.9
    top_k: Optional[int] = None
    min_p: Optional[float] = None  # "bottom_p" - minimum probability threshold
    typical_p: Optional[float] = None  # Typical decoding
    repetition_penalty: Optional[float] = None  # 1.0 = no penalty, >1.0 = penalize repeats
    no_repeat_ngram_size: Optional[int] = None  # Prevent n-gram repetition
    stream: Optional[bool] = False

@app.on_event("startup")
async def load_model():
    global model, processor
    print("Loading Ministral 14B...")

    model_id = "RoleModel/ministral-14b-merged-official"

    processor = AutoProcessor.from_pretrained(model_id, trust_remote_code=True)

    model = AutoModelForImageTextToText.from_pretrained(
        model_id,
        torch_dtype=torch.bfloat16,
        device_map="auto",
        trust_remote_code=True,
    )
    model.eval()
    print("Model loaded successfully!")

@app.post("/v1/chat/completions")
async def chat_completions(request: ChatRequest):
    global model, processor

    try:
        # Format messages
        messages = [{"role": m.role, "content": m.content} for m in request.messages]
        print(f"Processing {len(messages)} messages, stream={request.stream}")

        # Try to apply chat template
        try:
            chat_text = processor.apply_chat_template(
                messages,
                tokenize=False,
                add_generation_prompt=True
            )
        except Exception as e:
            print(f"Chat template error: {e}")
            chat_text = "<s>"
            for m in messages:
                if m["role"] == "system":
                    chat_text += f"[SYSTEM_PROMPT]{m['content']}[/SYSTEM_PROMPT]"
                elif m["role"] == "user":
                    chat_text += f"[INST]{m['content']}[/INST]"
                elif m["role"] == "assistant":
                    chat_text += f"{m['content']}</s>"

        # Tokenize
        inputs = processor.tokenizer(chat_text, return_tensors="pt").to(model.device)
        input_len = inputs["input_ids"].shape[1]
        print(f"Input tokens: {input_len}")

        if request.stream:
            # Streaming response
            async def generate_stream():
                streamer = TextIteratorStreamer(
                    processor.tokenizer,
                    skip_prompt=True,
                    skip_special_tokens=True
                )

                generation_kwargs = {
                    **inputs,
                    "max_new_tokens": request.max_tokens,
                    "temperature": request.temperature if request.temperature and request.temperature > 0 else 1.0,
                    "top_p": request.top_p if request.top_p else 0.9,
                    "do_sample": request.temperature is not None and request.temperature > 0,
                    "pad_token_id": processor.tokenizer.eos_token_id,
                    "streamer": streamer,
                }
                # Add optional parameters if provided
                if request.top_k is not None:
                    generation_kwargs["top_k"] = request.top_k
                if request.min_p is not None:
                    generation_kwargs["min_p"] = request.min_p
                if request.typical_p is not None:
                    generation_kwargs["typical_p"] = request.typical_p
                if request.repetition_penalty is not None:
                    generation_kwargs["repetition_penalty"] = request.repetition_penalty
                if request.no_repeat_ngram_size is not None:
                    generation_kwargs["no_repeat_ngram_size"] = request.no_repeat_ngram_size

                thread = Thread(target=model.generate, kwargs=generation_kwargs)
                thread.start()

                response_id = f"chatcmpl-{int(time.time())}"

                for text in streamer:
                    if text:
                        text = fix_bpe_tokens(text)
                        chunk = {
                            "id": response_id,
                            "object": "chat.completion.chunk",
                            "created": int(time.time()),
                            "model": request.model,
                            "choices": [{
                                "index": 0,
                                "delta": {"content": text},
                                "finish_reason": None
                            }]
                        }
                        yield f"data: {json.dumps(chunk)}\n\n"
                        await asyncio.sleep(0)

                # Send final chunk
                final_chunk = {
                    "id": response_id,
                    "object": "chat.completion.chunk",
                    "created": int(time.time()),
                    "model": request.model,
                    "choices": [{
                        "index": 0,
                        "delta": {},
                        "finish_reason": "stop"
                    }]
                }
                yield f"data: {json.dumps(final_chunk)}\n\n"
                yield "data: [DONE]\n\n"

                thread.join()

            return StreamingResponse(
                generate_stream(),
                media_type="text/event-stream",
                headers={
                    "Cache-Control": "no-cache, no-store, must-revalidate",
                    "Connection": "keep-alive",
                    "X-Accel-Buffering": "no",  # Disable nginx buffering
                    "Transfer-Encoding": "chunked",
                }
            )
        else:
            # Non-streaming response
            generation_kwargs = {
                **inputs,
                "max_new_tokens": request.max_tokens,
                "temperature": request.temperature if request.temperature and request.temperature > 0 else 1.0,
                "top_p": request.top_p if request.top_p else 0.9,
                "do_sample": request.temperature is not None and request.temperature > 0,
                "pad_token_id": processor.tokenizer.eos_token_id,
            }
            # Add optional parameters if provided
            if request.top_k is not None:
                generation_kwargs["top_k"] = request.top_k
            if request.min_p is not None:
                generation_kwargs["min_p"] = request.min_p
            if request.typical_p is not None:
                generation_kwargs["typical_p"] = request.typical_p
            if request.repetition_penalty is not None:
                generation_kwargs["repetition_penalty"] = request.repetition_penalty
            if request.no_repeat_ngram_size is not None:
                generation_kwargs["no_repeat_ngram_size"] = request.no_repeat_ngram_size

            with torch.no_grad():
                outputs = model.generate(**generation_kwargs)

            new_tokens = outputs[0][input_len:]
            response_text = processor.tokenizer.decode(
                new_tokens,
                skip_special_tokens=True,
                clean_up_tokenization_spaces=True
            )
            response_text = fix_bpe_tokens(response_text)
            print(f"Generated {len(new_tokens)} tokens")

            return {
                "id": f"chatcmpl-{int(time.time())}",
                "object": "chat.completion",
                "created": int(time.time()),
                "model": request.model,
                "choices": [{
                    "index": 0,
                    "message": {"role": "assistant", "content": response_text},
                    "finish_reason": "stop"
                }],
                "usage": {
                    "prompt_tokens": input_len,
                    "completion_tokens": len(new_tokens),
                    "total_tokens": input_len + len(new_tokens)
                }
            }
    except Exception as e:
        print(f"Error: {e}")
        traceback.print_exc()
        raise

@app.get("/v1/models")
async def list_models():
    return {
        "object": "list",
        "data": [{"id": "ministral-14b", "object": "model", "owned_by": "rolemodel"}]
    }

@app.get("/health")
async def health():
    return {"status": "ok"}

if __name__ == "__main__":
    uvicorn.run(app, host="0.0.0.0", port=8000)