fix runtime error

Dockerfile

@@ -1,22 +1,27 @@
+# Use lightweight Python image
 FROM python:3.10-slim
 
-# Cài đặt các thư viện hệ thống cần thiết
+# Prevent interactive prompts during package install
+ENV DEBIAN_FRONTEND=noninteractive
+
+# Set working directory
+WORKDIR /app
+
+# Install basic dependencies
 RUN apt-get update && apt-get install -y \
     git \
     && rm -rf /var/lib/apt/lists/*
 
-# Tạo thư mục làm việc
-WORKDIR /app
-
-# Copy mã nguồn và cài đặt requirements
+# Copy requirements and install Python packages
 COPY requirements.txt .
-RUN pip install --no-cache-dir -r requirements.txt
+RUN pip install --no-cache-dir --upgrade pip \
+    && pip install --no-cache-dir -r requirements.txt
 
-COPY . .
+# Copy application code
+COPY app.py .
 
-# Tạo thư mục cache có quyền ghi
-RUN mkdir -p /tmp/hf-cache && chmod -R 777 /tmp/hf-cache
-ENV TRANSFORMERS_CACHE=/tmp/hf-cache
+# Expose Gradio default port
+EXPOSE 7860
 
-# Chạy server FastAPI
-CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "8000"]
+# Run the app
+CMD ["python", "app.py"]

app.py

@@ -1,26 +1,64 @@
-import os
-from fastapi import FastAPI, Request
-from pydantic import BaseModel
-from transformers import AutoTokenizer, AutoModelForSeq2SeqLM
+import gradio as gr
+from transformers import AutoTokenizer, AutoModelForCausalLM
 import torch
+from fastapi import FastAPI, Request
+import uvicorn
 
-# Đặt thư mục cache có quyền ghi
-os.makedirs("/tmp/hf-cache", exist_ok=True)
-os.environ["TRANSFORMERS_CACHE"] = "/tmp/hf-cache"
+from threading import Thread
 
-# Sử dụng model công khai
-model_name = "VietAI/vit5-base"
+# -------- Load model --------
+model_name = "VietAI/gpt-neo-1.3B-vietnamese-news"
 tokenizer = AutoTokenizer.from_pretrained(model_name)
-model = AutoModelForSeq2SeqLM.from_pretrained(model_name)
+model = AutoModelForCausalLM.from_pretrained(model_name)
+model.eval()
+device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+model.to(device)
+
+# -------- Inference function --------
+def generate_text(prompt, max_tokens=100, temperature=0.9, top_p=0.95):
+    inputs = tokenizer(prompt, return_tensors="pt").to(device)
+    outputs = model.generate(
+        **inputs,
+        max_new_tokens=max_tokens,
+        do_sample=True,
+        temperature=temperature,
+        top_p=top_p,
+        pad_token_id=tokenizer.eos_token_id,
+    )
+    return tokenizer.decode(outputs[0], skip_special_tokens=True)
 
+# -------- Gradio UI --------
+def launch_gradio():
+    with gr.Blocks() as demo:
+        gr.Markdown("## 🇻🇳 VietAI GPT-Neo 1.3B - Sinh văn bản tiếng Việt")
+        prompt = gr.Textbox(label="Prompt", placeholder="Nhập đoạn mở đầu văn bản...")
+        max_tokens = gr.Slider(10, 200, value=100, label="Số tokens sinh ra")
+        temperature = gr.Slider(0.1, 1.5, value=0.9, label="Temperature")
+        top_p = gr.Slider(0.1, 1.0, value=0.95, label="Top-p sampling")
+        output = gr.Textbox(label="Kết quả", lines=10)
+        btn = gr.Button("Sinh văn bản")
+        btn.click(fn=generate_text, inputs=[prompt, max_tokens, temperature, top_p], outputs=output)
+
+    demo.launch(server_name="0.0.0.0", server_port=7860, share=False)
+
+# -------- FastAPI for REST API --------
 app = FastAPI()
 
-class InputData(BaseModel):
-    input: str
+@app.post("/generate")
+async def generate(request: Request):
+    body = await request.json()
+    prompt = body.get("prompt", "")
+    max_tokens = body.get("max_tokens", 100)
+    temperature = body.get("temperature", 0.9)
+    top_p = body.get("top_p", 0.95)
+    output = generate_text(prompt, max_tokens, temperature, top_p)
+    return {"response": output}
+
+# -------- Start Gradio in background --------
+if __name__ == "__main__":
+    # Run Gradio in another thread
+    thread = Thread(target=launch_gradio)
+    thread.start()
 
-@app.post("/predict")
-async def predict(request: Request, data: InputData):
-    input_ids = tokenizer.encode(data.input, return_tensors="pt", max_length=512, truncation=True)
-    output_ids = model.generate(input_ids, max_length=128, num_beams=4, early_stopping=True)
-    output = tokenizer.decode(output_ids[0], skip_special_tokens=True)
-    return {"output": output}
+    # Start FastAPI
+    uvicorn.run(app, host="0.0.0.0", port=8000)

requirements.txt

@@ -1,4 +1,5 @@
-fastapi
+transformers>=4.40.0
+torch
+gradio>=4.26.0
 uvicorn
-transformers==4.40.0
-torch>=1.13.1
+fastapi