Deploy FastAPI Hugging Face Space without model files

Dockerfile

@@ -1,15 +1,11 @@
-FROM python:3.10-slim
-
-RUN useradd -m -u 1000 user
-USER user
-ENV HOME=/home/user \
-    PATH=/home/user/.local/bin:$PATH \
-    TRANSFORMERS_CACHE=/home/user/.cache/huggingface \
-    HUGGINGFACE_HUB_CACHE=/home/user/.cache/huggingface
+FROM python:3.10
 
 WORKDIR /app
+
 COPY requirements.txt .
-RUN pip install --no-cache-dir --user -r requirements.txt
-COPY . .
+RUN pip install --no-cache-dir -r requirements.txt
+
+COPY ./app ./app
 
+EXPOSE 7860
 CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "7860"]

app/main.py

@@ -1,33 +1,14 @@
-from fastapi import FastAPI, HTTPException
+from fastapi import FastAPI
 from pydantic import BaseModel
 from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
 import torch
-import os
-import logging
-
-# إعداد الـ logging
-logging.basicConfig(level=logging.INFO)
-logger = logging.getLogger(__name__)
 
 app = FastAPI(title="MGZON FLAN-T5 API")
 
-# Environment Variable (لو هتحتاج Token)
-HUGGING_FACE_TOKEN = os.getenv("HUGGING_FACE_TOKEN", None)
-
+# تحميل النموذج من Hugging Face مباشرة
 MODEL_NAME = "MGZON/mgzon-flan-t5-base"
-
-try:
-    tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME, use_auth_token=HUGGING_FACE_TOKEN)
-    model = AutoModelForSeq2SeqLM.from_pretrained(
-        MODEL_NAME,
-        use_auth_token=HUGGING_FACE_TOKEN,
-        torch_dtype=torch.float16,
-        device_map="auto"  # أو "cpu" لو مش فيه GPU
-    )
-    logger.info("Model and tokenizer loaded successfully")
-except Exception as e:
-    logger.error(f"Failed to load model: {e}")
-    raise
+tokenizer = AutoTokenizer.from_pretrained(MODEL_NAME)
+model = AutoModelForSeq2SeqLM.from_pretrained(MODEL_NAME, device_map="auto")
 
 class RequestText(BaseModel):
     text: str
@@ -37,25 +18,11 @@ class RequestText(BaseModel):
 async def health_check():
     return {"status": "healthy"}
 
-@app.post("/api/generate")
+@app.post("/api/generate/")
 async def generate(req: RequestText):
-    try:
-        inputs = tokenizer(req.text, return_tensors="pt", truncation=True, max_length=512).to(model.device)
-        outputs = model.generate(
-            **inputs,
-            max_length=req.max_length,
-            do_sample=True,
-            temperature=0.7,
-            top_p=0.9,
-            top_k=50,
-            pad_token_id=tokenizer.pad_token_id,
-            num_return_sequences=1
-        )
-        text = tokenizer.decode(outputs[0], skip_special_tokens=True)
-        return {"generated_text": text}
-    except Exception as e:
-        logger.error(f"Error generating text: {e}")
-        raise HTTPException(status_code=500, detail=str(e))
+    inputs = tokenizer(req.text, return_tensors="pt").to(model.device)
+    outputs = model.generate(**inputs, max_length=req.max_length)
+    return {"generated_text": tokenizer.decode(outputs[0], skip_special_tokens=True)}
 
 if __name__ == "__main__":
     import uvicorn

requirements.txt

@@ -3,3 +3,4 @@ uvicorn[standard]
 transformers
 torch
 accelerate
+pydantic