app.py

import os
os.environ["HF_HOME"] = "/tmp"
os.environ["HF_TRANSFORMERS_CACHE"] = "/tmp/huggingface"
from fastapi import FastAPI, Request
from transformers import AutoModelForCausalLM, AutoTokenizer
from peft import PeftModel
import torch

app = FastAPI()

base_model_id = "unsloth/gemma-3n-E2B"
lora_adapter_id = "Rudra1729/lora-adapters"

base_model = AutoModelForCausalLM.from_pretrained(base_model_id, torch_dtype=torch.float16)
tokenizer = AutoTokenizer.from_pretrained(base_model_id)
model = PeftModel.from_pretrained(base_model, lora_adapter_id)

@app.get("/")
def read_root():
    return {"message": "Nidaan-AI FastAPI backend is running. Use /generate to interact with the model."}

@app.post("/generate")
async def generate_text(request: Request):
    body = await request.json()
    prompt = body.get("prompt", "")
    inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
    outputs = model.generate(**inputs, max_new_tokens=256)
    result = tokenizer.decode(outputs[0], skip_special_tokens=True)
    return {"response": result}


if __name__ == "__main__":
    import uvicorn
    port = int(os.environ.get("PORT", 8080))  # Use 8080 by default for Cloud Run
    uvicorn.run(app, host="0.0.0.0", port=port)