app.py

from fastapi import FastAPI
from pydantic import BaseModel
from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
import uvicorn

app = FastAPI()

model_id = "GoToCompany/gemma2-9b-cpt-sahabatai-v1-instruct"
tokenizer = AutoTokenizer.from_pretrained(model_id)
model = AutoModelForCausalLM.from_pretrained(model_id, torch_dtype=torch.float16, device_map="auto")

class ChatRequest(BaseModel):
    prompt: str
    max_new_tokens: int = 256
    temperature: float = 0.7
    top_p: float = 0.95

@app.post("/chat")
async def chat(request: ChatRequest):
    inputs = tokenizer(request.prompt, return_tensors="pt").to(model.device)
    outputs = model.generate(
        **inputs,
        max_new_tokens=request.max_new_tokens,
        temperature=request.temperature,
        top_p=request.top_p,
        do_sample=True,
        pad_token_id=tokenizer.eos_token_id,
    )
    result = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return {"response": result}

# This will only run locally or in Spaces, not if you import this module
if __name__ == "__main__":
    uvicorn.run(app, host="0.0.0.0", port=7860)