fix: use simple pipeline with our fine-tuned model and A100

README.md

@@ -1,65 +1,46 @@
 ---
-title: Math Problem Solver Demo
+title: Mathematics Problem Solver
 emoji: 🧮
 colorFrom: blue
-colorTo: purple
+colorTo: green
 sdk: gradio
-sdk_version: 5.8.0
+sdk_version: 4.8.0
 app_file: app.py
 pinned: false
+hardware:
+  accelerator: a100
+  gpu: true
+python_packages:
+  - "torch>=2.0.0"
+  - "transformers>=4.30.0"
+  - "accelerate>=0.20.0"
+  - "peft==0.5.0"
+  - "numpy>=1.21.0"
 ---
 
-# Mathematics Problem Solver Demo
+# Mathematics Problem Solver
 
-This demo showcases a comparison between base and fine-tuned language models in solving mathematical problems. It features real-time performance monitoring and supports multiple types of math problems.
+This Space demonstrates our fine-tuned math model for solving various mathematical problems, with a focus on derivatives. Compare solutions between:
 
-## Models Used
-
-- Base Model: [LlaMA 3.2 1B](https://huggingface.co/Alexis-Az/Math-Problem-LlaMA-3.2-1B-GGUF)
-- Fine-tuned Model: [SmolLM2 1.7B](https://huggingface.co/Alexis-Az/Math-Problem-LlaMA-3.2-1.7B-GGUF)
+1. Base Model: HuggingFaceTB/SmolLM2-1.7B-Instruct
+2. Our Fine-tuned Model: Joash2024/Math-SmolLM2-1.7B
 
 ## Features
 
-- 🔢 Multiple problem types:
-  - Addition operations
-  - Root finding
-  - Derivatives
-  - Custom problems
-- 📊 Real-time performance metrics:
+- Side-by-side comparison of base and fine-tuned models
+- Performance monitoring:
   - Response times
   - Success rates
   - Problem type distribution
-- 🔄 Side-by-side model comparison
-- ⚡ Example problems included
-
-## How to Use
-
-1. Select a problem type from the dropdown menu
-2. Enter your math problem in the input field
-3. Click "Solve" to see solutions from both models
-4. Compare the results and view performance metrics
-
-## Example Problems
-
-Try these sample problems:
-
-- Derivative: "Find the derivative of x^2 + 3x"
-- Root Finding: "What is the square root of 144?"
-- Addition: "Calculate 235 + 567"
-
-## Performance Monitoring
-
-The interface includes a live dashboard showing:
-
-- Average response times for each model
-- Success rates comparison
-- Distribution of problem types solved
-- Real-time performance metrics
-
-## Project Details
-
-This demo is part of a larger project comparing LLM performance on mathematical problems. The models have been fine-tuned on a custom dataset of mathematical problems to improve their problem-solving capabilities.
+- Support for various problems:
+  - Derivatives
+  - Addition
+  - Roots
+  - Custom problems
 
-## Credits
+## Technical Details
 
-Models provided by [Alexis-Az](https://huggingface.co/Alexis-Az)
+- A100 GPU acceleration
+- Float16 precision for efficient inference
+- LaTeX notation support
+- Real-time performance tracking

app.py

@@ -5,8 +5,8 @@ import numpy as np
 from monitoring import PerformanceMonitor, measure_time
 
 # Model IDs
-BASE_MODEL_ID = "Alexis-Az/Math-Problem-LlaMA-3.2-1B-GGUF"
-FINETUNED_MODEL_ID = "Alexis-Az/Math-Problem-LlaMA-3.2-1.7B-GGUF"
+BASE_MODEL_ID = "HuggingFaceTB/SmolLM2-1.7B-Instruct"  # Base model
+FINETUNED_MODEL_ID = "Joash2024/Math-SmolLM2-1.7B"     # Our fine-tuned model
 
 # Initialize performance monitor
 monitor = PerformanceMonitor()
@@ -92,7 +92,7 @@ def solve_problem(problem, problem_type):
 # Create Gradio interface
 with gr.Blocks(title="Mathematics Problem Solver") as demo:
     gr.Markdown("# Mathematics Problem Solver")
-    gr.Markdown("Compare solutions between base (1B) and fine-tuned (1.7B) models")
+    gr.Markdown("Compare solutions between base and fine-tuned models")
     
     with gr.Row():
         with gr.Column():
@@ -109,11 +109,11 @@ with gr.Blocks(title="Mathematics Problem Solver") as demo:
     
     with gr.Row():
         with gr.Column():
-            gr.Markdown("### Base Model (1B)")
+            gr.Markdown("### Base Model")
             base_output = gr.Textbox(label="Base Model Solution", lines=5)
         
         with gr.Column():
-            gr.Markdown("### Fine-tuned Model (1.7B)")
+            gr.Markdown("### Fine-tuned Model")
             finetuned_output = gr.Textbox(label="Fine-tuned Model Solution", lines=5)
     
     # Performance metrics display
@@ -126,6 +126,11 @@ with gr.Blocks(title="Mathematics Problem Solver") as demo:
             ["Find the derivative of x^2 + 3x", "Derivative"],
             ["What is the square root of 144?", "Root Finding"],
             ["Calculate 235 + 567", "Addition"],
+            ["\\sin{\\left(x\\right)}", "Derivative"],
+            ["e^x", "Derivative"],
+            ["\\frac{1}{x}", "Derivative"],
+            ["x^3 + 2x", "Derivative"],
+            ["\\cos{\\left(x^2\\right)}", "Derivative"]
         ],
         inputs=[problem_input, problem_type],
         outputs=[base_output, finetuned_output, metrics_display],