app/main.py

import os
import torch
from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
from transformers import AutoModelForCausalLM, AutoTokenizer

# Set Hugging Face cache directory to avoid permission issues in Docker
os.environ["HF_HOME"] = "/app/huggingface_cache"
os.environ["TRANSFORMERS_CACHE"] = "/app/huggingface_cache"

app = FastAPI()

class TextGenerationRequest(BaseModel):
    prompt: str
    max_length: int = 100
    temperature: float = 0.7

# Load model and tokenizer (Force CPU)
model_name = "unsloth/Qwen2.5-7B-bnb-4bit"
tokenizer = AutoTokenizer.from_pretrained(model_name, trust_remote_code=True)

model = AutoModelForCausalLM.from_pretrained(
    model_name,
    trust_remote_code=True,
    torch_dtype=torch.float32,  # Use float32 for CPU
    device_map={"": "cpu"}  # Ensure CPU usage
)

@app.get("/")
def api_home():
    return {"detail": "Welcome to FastAPI TextGen API!"}

@app.post("/generate")
async def generate_text(request: TextGenerationRequest):
    try:
        inputs = tokenizer(request.prompt, return_tensors="pt").to("cpu")  # Ensure CPU usage
        outputs = model.generate(
            inputs.input_ids,
            max_length=request.max_length,
            temperature=request.temperature,
            do_sample=True,
        )
        generated_text = tokenizer.decode(outputs[0], skip_special_tokens=True)
        return {"generated_text": generated_text}
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))