Upload app.py with huggingface_hub

app.py

@@ -20,21 +20,21 @@ def create_summary_metrics():
     metrics = data['metrics']
 
     summary = f"""
-# CHIMERA Performance Summary
-
-## Overall Metrics
-- **Average Speedup:** {metrics['average_speedup']:.1f}x faster than baseline
-- **Maximum Speedup:** {metrics['max_speedup']:.1f}x (best case)
-- **Average Latency:** {metrics['average_latency_ms']:.2f}ms
-- **Energy Efficiency:** {metrics['average_energy_joules']:.3f}J per operation
-- **Efficiency Score:** {metrics['average_efficiency']:.1f} ops/J
-
-## Architecture Advantages
-- **Framework Size:** {metrics['framework_size_mb']}MB (99.6% smaller than PyTorch)
-- **Memory Footprint:** {metrics['memory_footprint_mb']}MB (88.7% reduction)
-- **All-in-One GPU:** No CPU/RAM usage - pure GPU processing
-- **Universal Hardware:** Works on NVIDIA, AMD, Intel, Apple M1/M2
-"""
+    # CHIMERA Performance Summary
+
+    ## Overall Metrics
+    - **Average Speedup:** {metrics['average_speedup']:.1f}x faster than baseline
+    - **Maximum Speedup:** {metrics['max_speedup']:.1f}x (best case)
+    - **Average Latency:** {metrics['average_latency_ms']:.2f}ms
+    - **Energy Efficiency:** {metrics['average_energy_joules']:.3f}J per operation
+    - **Efficiency Score:** {metrics['average_efficiency']:.1f} ops/J
+
+    ## Architecture Advantages
+    - **Framework Size:** {metrics['framework_size_mb']}MB (99.6% smaller than PyTorch)
+    - **Memory Footprint:** {metrics['memory_footprint_mb']}MB (88.7% reduction)
+    - **All-in-One GPU:** No CPU/RAM usage - pure GPU processing
+    - **Universal Hardware:** Works on NVIDIA, AMD, Intel, Apple M1/M2
+    """
 
     return summary
 
@@ -58,8 +58,7 @@ def create_speedup_chart():
         xaxis_title='Benchmark Task',
         yaxis_title='Speedup Factor (x)',
         yaxis_type='log',
-        height=500,
-        xaxis_tickangle=-45
+        height=500
     )
 
     return fig
@@ -90,8 +89,7 @@ def create_latency_comparison():
         yaxis_title='Latency (ms)',
         yaxis_type='log',
         barmode='group',
-        height=500,
-        xaxis_tickangle=-45
+        height=500
     )
 
     return fig
@@ -105,13 +103,13 @@ def create_energy_efficiency_chart():
         x='energy_joules',
         y='efficiency_score',
         size='speedup_factor',
-        color='benchmark_suite',
+        color='benchmark_name',
         hover_data=['task_name', 'latency_ms', 'power_watts'],
         title='Energy Efficiency: Lower Energy + Higher Efficiency = Better',
         labels={
             'energy_joules': 'Energy Consumption (J)',
             'efficiency_score': 'Efficiency Score (ops/J)',
-            'benchmark_suite': 'Benchmark Suite'
+            'benchmark_name': 'Benchmark'
         }
     )
 
@@ -119,12 +117,46 @@ def create_energy_efficiency_chart():
 
     return fig
 
+def create_hardware_scaling_chart():
+    """Create hardware scalability visualization"""
+    # Filter scalability benchmarks
+    scaling_df = pd.DataFrame([
+        b for b in data['benchmarks']
+        if 'Scalability' in b['benchmark_name']
+    ])
+
+    if len(scaling_df) == 0:
+        return go.Figure().update_layout(title="No scalability data available")
+
+    fig = go.Figure()
+
+    for platform in scaling_df['hardware_platform'].unique():
+        platform_data = scaling_df[scaling_df['hardware_platform'] == platform]
+
+        fig.add_trace(go.Bar(
+            name=platform,
+            x=['Latency', 'Power'],
+            y=[
+                platform_data['latency_ms'].values[0],
+                platform_data['power_watts'].values[0]
+            ]
+        ))
+
+    fig.update_layout(
+        title='Hardware Scalability: CHIMERA Performance Across Platforms',
+        yaxis_title='Value',
+        barmode='group',
+        height=500
+    )
+
+    return fig
+
 def get_detailed_table():
     """Create detailed results table"""
     df = pd.DataFrame(data['benchmarks'])
 
     table_df = df[[
-        'benchmark_suite', 'task_name', 'latency_ms', 'throughput_qps',
+        'benchmark_name', 'task_name', 'latency_ms', 'throughput_qps',
         'speedup_factor', 'energy_joules', 'efficiency_score', 'hardware_platform'
     ]].copy()
 
@@ -149,56 +181,81 @@ with gr.Blocks(title="CHIMERA Benchmark Dashboard", theme=gr.themes.Soft()) as d
         gr.Markdown(create_summary_metrics())
 
     with gr.Tab("Performance"):
-        with gr.Row():
-            gr.Plot(create_speedup_chart())
-        with gr.Row():
-            gr.Plot(create_latency_comparison())
+        gr.Plot(create_speedup_chart())
+        gr.Plot(create_latency_comparison())
 
     with gr.Tab("Energy Efficiency"):
         gr.Plot(create_energy_efficiency_chart())
         gr.Markdown("""
-## Energy Efficiency Analysis
+        ## Energy Efficiency Analysis
+
+        CHIMERA achieves exceptional energy efficiency through:
+        - **All-in-one GPU processing** - No CPU/RAM overhead
+        - **Holographic memory** - Data stays in GPU textures
+        - **Frame-by-frame simulation** - Efficient neuromorphic computation
+        - **Minimal framework size** - 10MB vs 2.5GB for PyTorch
+
+        **Average energy savings: 92.7% vs baseline frameworks**
+        """)
+
+    with gr.Tab("Hardware Scalability"):
+        gr.Plot(create_hardware_scaling_chart())
+        gr.Markdown("""
+        ## Universal Hardware Support
 
-CHIMERA achieves exceptional energy efficiency through:
-- **All-in-one GPU processing** - No CPU/RAM overhead
-- **Holographic memory** - Data stays in GPU textures
-- **Frame-by-frame simulation** - Efficient neuromorphic computation
-- **Minimal framework size** - 10MB vs 2.5GB for PyTorch
+        CHIMERA works on any GPU with OpenGL 4.3+:
+        - NVIDIA GeForce/RTX (CUDA 11.0+)
+        - AMD Radeon (OpenGL 4.6)
+        - Intel UHD/Iris (OpenGL 4.5)
+        - Apple M1/M2 (Metal backend)
+        - Raspberry Pi 4 (OpenGL 3.3)
 
-**Average energy savings: 92.7% vs baseline frameworks**
-""")
+        **No vendor lock-in - truly universal AI acceleration**
+        """)
 
     with gr.Tab("Detailed Results"):
         gr.Dataframe(get_detailed_table(), interactive=True)
 
     with gr.Tab("About"):
         gr.Markdown(f"""
-## About CHIMERA
-
-CHIMERA is a revolutionary all-in-one GPU architecture for artificial intelligence:
-
-### Key Innovations
-1. **Everything as Images** - All processing happens as frame-by-frame GPU textures
-2. **Living Brain** - Evolutionary cellular automaton simulates neuromorphic intelligence
-3. **Holographic Memory** - Memory integrated within GPU textures (no RAM needed)
-4. **Pure GPU** - Zero CPU usage during inference
-5. **Universal** - Works on any GPU hardware
-
-### Performance Highlights
-- Average {data['metrics']['average_speedup']:.1f}x speedup
-- 88.7% memory reduction
-- 92.7% energy savings
-- 10MB framework (vs 2.5GB PyTorch)
-
-### Repository
-- GitHub: [CHIMERA Architecture](https://github.com/Agnuxo1/CHIMERA-Revolutionary-AI-Architecture)
-- Author: Francisco Angulo de Lafuente
-- Version: {data['model_name']}
-
-### Public Benchmarks
-- OpenML Dataset: [Dataset 47101](https://www.openml.org/d/47101)
-- Weights & Biases: [Dashboard](https://wandb.ai/lareliquia-angulo-agnuxo/chimera-public-benchmarks)
-""")
+        ## About CHIMERA
+
+        CHIMERA is a revolutionary all-in-one GPU architecture for artificial intelligence:
+
+        ### Key Innovations
+        1. **Everything as Images** - All processing happens as frame-by-frame GPU textures
+        2. **Living Brain** - Evolutionary cellular automaton simulates neuromorphic intelligence
+        3. **Holographic Memory** - Memory integrated within GPU textures (no RAM needed)
+        4. **Pure GPU** - Zero CPU usage during inference
+        5. **Universal** - Works on any GPU hardware
+
+        ### Architecture Principles
+        - **Neuromorphic simulation** in every frame
+        - **Cellular automaton** creates emergent intelligence
+        - **Holographic encoding** for efficient memory
+        - **OpenGL compute shaders** for universal compatibility
+
+        ### Performance Highlights
+        - Average {data['metrics']['average_speedup']:.1f}x speedup
+        - 88.7% memory reduction
+        - 92.7% energy savings
+        - 10MB framework (vs 2.5GB PyTorch)
+
+        ### Repository
+        - GitHub: [CHIMERA Architecture](https://github.com/Agnuxo1/CHIMERA-Revolutionary-AI-Architecture)
+        - Author: Francisco Angulo de Lafuente
+        - Version: {data['model_name']}
+
+        ### Citation
+        ```
+        @software{{chimera2025,
+          title={{CHIMERA: All-in-One GPU Neuromorphic Architecture}},
+          author={{Angulo de Lafuente, Francisco}},
+          year={{2025}},
+          url={{https://github.com/Agnuxo1/CHIMERA-Revolutionary-AI-Architecture}}
+        }}
+        ```
+        """)
 
 if __name__ == "__main__":
     demo.launch()