Simplify app.py and update configuration for ZeroGPU

README.md

@@ -4,9 +4,11 @@ emoji: 🤖
 colorFrom: blue
 colorTo: green
 sdk: gradio
-sdk_version: 5.8.0
+sdk_version: 4.0.0
 app_file: app.py
 pinned: false
+hf_oauth: false
+hardware: a10g-small
 ---
 
 # Code Review Assistant
@@ -23,8 +25,15 @@ An automated code review system powered by Gemma-2b that provides intelligent co
 
 ### LLMOps Integration
 - Uses Gemma-2b for intelligent code analysis
-- Provides detailed code analysis
-- Generates actionable suggestions
+- Tracks model performance and accuracy
+- Monitors response times and token usage
+- Optimized with ZeroGPU for efficient inference
+
+### Performance Monitoring
+- Real-time metrics dashboard
+- Review history tracking
+- Response time monitoring
+- Usage statistics
 
 ### User Interface
 - Simple and intuitive Gradio interface
@@ -39,23 +48,24 @@ The following environment variables need to be set in your Hugging Face Space:
 - `HUGGING_FACE_TOKEN`: Your Hugging Face API token (required)
 - `MODEL_NAME`: google/gemma-2b-it (default)
 
+## Hardware Configuration
+
+This Space uses:
+- Runtime: ZeroGPU
+- Hardware: A10G Small
+- Memory: Optimized for efficient model inference
+
 ## Usage
 
 1. Enter your code in the text box
 2. Select the programming language from the dropdown
-3. Click "Submit" to get the review
+3. Click "Submit for Review"
 4. View the detailed analysis including:
    - Critical issues
    - Suggested improvements
    - Best practices
    - Security considerations
 
-## Example Code
-
-Try the included example code snippets to see how the review system works:
-- Python function example
-- JavaScript array processing example
-
 ## Model Details
 
 This application uses the Gemma-2b-it model from Google, which is:
@@ -66,4 +76,4 @@ This application uses the Gemma-2b-it model from Google, which is:
 
 ## License
 
-This project is licensed under the MIT License.
+This project is licensed under the MIT License.

app.py

@@ -7,9 +7,6 @@ import logging
 from datetime import datetime
 import json
 from typing import List, Dict
-from huggingface_hub import HfApi
-from huggingface_hub.spaces import SpaceHardware, SpaceStage
-from huggingface_hub.spaces.space_sdk import SpaceRuntime
 
 # Configure logging
 logging.basicConfig(level=logging.INFO)
@@ -19,16 +16,6 @@ logger = logging.getLogger(__name__)
 HF_TOKEN = os.getenv("HUGGING_FACE_TOKEN")
 MODEL_NAME = os.getenv("MODEL_NAME", "google/gemma-2b-it")
 
-# Initialize Hugging Face API
-api = HfApi()
-
-# Space hardware configuration
-space_config = {
-    "hardware": SpaceHardware.A10G_SMALL,
-    "stage": SpaceStage.RUNTIME,
-    "runtime": SpaceRuntime.ZEROGPU
-}
-
 class Review:
     def __init__(self, code: str, language: str, suggestions: str):
         self.code = code
@@ -41,7 +28,7 @@ class CodeReviewer:
     def __init__(self):
         self.model = None
         self.tokenizer = None
-        self.device = None
+        self.device = torch.device("cuda" if torch.cuda.is_available() else "cpu")
         self.review_history: List[Review] = []
         self.metrics = {
             'total_reviews': 0,
@@ -64,17 +51,14 @@ class CodeReviewer:
             )
             
             logger.info("Loading model...")
-            # Initialize with ZeroGPU configuration
             self.model = AutoModelForCausalLM.from_pretrained(
                 MODEL_NAME,
                 token=HF_TOKEN,
                 device_map="auto",
                 torch_dtype=torch.float16,
                 trust_remote_code=True,
-                low_cpu_mem_usage=True,
-                use_zerogpu=True  # Enable ZeroGPU
+                low_cpu_mem_usage=True
             )
-            self.device = next(self.model.parameters()).device
             logger.info(f"Model loaded successfully on {self.device}")
         except Exception as e:
             logger.error(f"Error initializing model: {e}")
@@ -105,7 +89,7 @@ Code:
                 truncation=True,
                 max_length=512,
                 padding=True
-            ).to(self.device)  # Move inputs to GPU
+            ).to(self.device)
             
             with torch.no_grad():
                 outputs = self.model.generate(
@@ -179,12 +163,8 @@ Code:
             'Device': str(self.device)
         }
 
-# Initialize reviewer with ZeroGPU
-@space_config
-def create_reviewer():
-    return CodeReviewer()
-
-reviewer = create_reviewer()
+# Initialize reviewer
+reviewer = CodeReviewer()
 
 # Create Gradio interface
 with gr.Blocks(theme=gr.themes.Soft()) as iface: