Update src/streamlit_app.py

src/streamlit_app.py

@@ -12,11 +12,8 @@ import numpy as np
 import re
 import plotly.express as px
 import plotly.graph_objects as go
-from plotly.subplots import make_subplots
-from typing import Optional, Tuple, List, Dict, Any
+from typing import Optional, Tuple, List, Dict
 import json
-from datetime import datetime, timedelta
-import time
 
 # ✅ MUST be the first Streamlit command
 st.set_page_config(
@@ -25,60 +22,6 @@ st.set_page_config(
     page_icon="🧠",
     initial_sidebar_state="expanded"
 )
-# Custom CSS for better styling
-st.markdown("""
-<style>
-    .main-header {
-        font-size: 3rem;
-        font-weight: bold;
-        text-align: center;
-        background: linear-gradient(90deg, #667eea 0%, #764ba2 100%);
-        -webkit-background-clip: text;
-        -webkit-text-fill-color: transparent;
-        margin-bottom: 2rem;
-    }
-    .metric-card {
-        background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
-        padding: 1rem;
-        border-radius: 10px;
-        color: white;
-        margin: 0.5rem 0;
-    }
-    .model-card {
-        border: 2px solid #e0e0e0;
-        border-radius: 10px;
-        padding: 1rem;
-        margin: 0.5rem 0;
-        background: white;
-        box-shadow: 0 2px 4px rgba(0,0,0,0.1);
-    }
-    .performance-badge {
-        padding: 0.25rem 0.5rem;
-        border-radius: 15px;
-        font-size: 0.8rem;
-        font-weight: bold;
-        color: white;
-    }
-    .badge-ultra-high { background: #8B5CF6; }
-    .badge-high { background: #3B82F6; }
-    .badge-good { background: #10B981; }
-    .badge-moderate { background: #F59E0B; }
-    .badge-low { background: #EF4444; }
-    .badge-ultra-low { background: #6B7280; }
-    .stTabs [data-baseweb="tab-list"] {
-        gap: 2px;
-    }
-    .stTabs [data-baseweb="tab"] {
-        height: 50px;
-        padding-left: 20px;
-        padding-right: 20px;
-        background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
-        border-radius: 10px 10px 0 0;
-        color: white;
-        font-weight: bold;
-    }
-</style>
-""", unsafe_allow_html=True)
 
 # Enhanced data loading with error handling
 @st.cache_data
@@ -130,14 +73,42 @@ def extract_numeric_ram(ram) -> Optional[int]:
 
 # Quantization options and size calculations
 QUANTIZATION_FORMATS = {
-    "FP16": {"multiplier": 1.0, "description": "Full precision, best quality", "icon": "🔥"},
-    "8-bit": {"multiplier": 0.5, "description": "50% smaller, good quality", "icon": "⚡"},
-    "4-bit": {"multiplier": 0.25, "description": "75% smaller, acceptable quality", "icon": "💎"},
-    "2-bit": {"multiplier": 0.125, "description": "87.5% smaller, experimental", "icon": "🧪"}
+    "FP16": {
+        "multiplier": 1.0, 
+        "description": "Full precision, best quality", 
+        "icon": "🔥",
+        "quality": "Excellent",
+        "speed": "Moderate",
+        "memory_efficiency": "Low"
+    },
+    "8-bit": {
+        "multiplier": 0.5, 
+        "description": "50% smaller, good quality", 
+        "icon": "⚡",
+        "quality": "Very Good",
+        "speed": "Good",
+        "memory_efficiency": "Good"
+    },
+    "4-bit": {
+        "multiplier": 0.25, 
+        "description": "75% smaller, acceptable quality", 
+        "icon": "💎",
+        "quality": "Good",
+        "speed": "Very Good",
+        "memory_efficiency": "Excellent"
+    },
+    "2-bit": {
+        "multiplier": 0.125, 
+        "description": "87.5% smaller, experimental", 
+        "icon": "🧪",
+        "quality": "Fair",
+        "speed": "Excellent",
+        "memory_efficiency": "Outstanding"
+    }
 }
 
 def calculate_quantized_size(base_size_str, quant_format):
-    """Calculate quantized model size"""
+    """Calculate quantized model size with better formatting"""
     size_match = re.search(r'(\d+\.?\d*)', base_size_str)
     if not size_match:
         return base_size_str
@@ -148,7 +119,13 @@ def calculate_quantized_size(base_size_str, quant_format):
     multiplier = QUANTIZATION_FORMATS[quant_format]["multiplier"]
     new_size = base_size * multiplier
     
-    return f"{new_size:.1f}{unit}"
+    # Smart unit conversion
+    if unit.upper() == "GB" and new_size < 1:
+        return f"{new_size * 1024:.0f}MB"
+    elif unit.upper() == "MB" and new_size > 1024:
+        return f"{new_size / 1024:.1f}GB"
+    else:
+        return f"{new_size:.1f}{unit}"
 
 # Enhanced LLM database with more models and metadata
 LLM_DATABASE = {
@@ -271,20 +248,21 @@ LLM_DATABASE = {
 }
 
 # GPU compatibility database
+# Enhanced GPU compatibility database with more details
 GPU_DATABASE = {
-    "RTX 3060": {"vram": 8, "performance": "mid", "architecture": "Ampere"},
-    "RTX 3070": {"vram": 8, "performance": "high", "architecture": "Ampere"},
-    "RTX 3080": {"vram": 10, "performance": "high", "architecture": "Ampere"},
-    "RTX 3090": {"vram": 24, "performance": "ultra", "architecture": "Ampere"},
-    "RTX 4060": {"vram": 8, "performance": "mid", "architecture": "Ada Lovelace"},
-    "RTX 4070": {"vram": 12, "performance": "high", "architecture": "Ada Lovelace"},
-    "RTX 4080": {"vram": 16, "performance": "ultra", "architecture": "Ada Lovelace"},
-    "RTX 4090": {"vram": 24, "performance": "ultra", "architecture": "Ada Lovelace"},
-    "Apple M1": {"vram": 8, "performance": "mid", "architecture": "Apple Silicon"},
-    "Apple M2": {"vram": 16, "performance": "high", "architecture": "Apple Silicon"},
-    "Apple M3": {"vram": 24, "performance": "ultra", "architecture": "Apple Silicon"},
-    "RX 6700 XT": {"vram": 12, "performance": "mid", "architecture": "RDNA 2"},
-    "RX 7900 XTX": {"vram": 24, "performance": "ultra", "architecture": "RDNA 3"},
+    "RTX 3060": {"vram": 8, "performance": "mid", "architecture": "Ampere", "tensor_cores": "2nd gen", "memory_bandwidth": "360 GB/s"},
+    "RTX 3070": {"vram": 8, "performance": "high", "architecture": "Ampere", "tensor_cores": "2nd gen", "memory_bandwidth": "448 GB/s"},
+    "RTX 3080": {"vram": 10, "performance": "high", "architecture": "Ampere", "tensor_cores": "2nd gen", "memory_bandwidth": "760 GB/s"},
+    "RTX 3090": {"vram": 24, "performance": "ultra", "architecture": "Ampere", "tensor_cores": "2nd gen", "memory_bandwidth": "936 GB/s"},
+    "RTX 4060": {"vram": 8, "performance": "mid", "architecture": "Ada Lovelace", "tensor_cores": "4th gen", "memory_bandwidth": "272 GB/s"},
+    "RTX 4070": {"vram": 12, "performance": "high", "architecture": "Ada Lovelace", "tensor_cores": "4th gen", "memory_bandwidth": "504 GB/s"},
+    "RTX 4080": {"vram": 16, "performance": "ultra", "architecture": "Ada Lovelace", "tensor_cores": "4th gen", "memory_bandwidth": "716 GB/s"},
+    "RTX 4090": {"vram": 24, "performance": "ultra", "architecture": "Ada Lovelace", "tensor_cores": "4th gen", "memory_bandwidth": "1008 GB/s"},
+    "Apple M1": {"vram": 8, "performance": "mid", "architecture": "Apple Silicon", "tensor_cores": "None", "memory_bandwidth": "68.25 GB/s"},
+    "Apple M2": {"vram": 16, "performance": "high", "architecture": "Apple Silicon", "tensor_cores": "None", "memory_bandwidth": "100 GB/s"},
+    "Apple M3": {"vram": 24, "performance": "ultra", "architecture": "Apple Silicon", "tensor_cores": "None", "memory_bandwidth": "150 GB/s"},
+    "RX 6700 XT": {"vram": 12, "performance": "mid", "architecture": "RDNA 2", "tensor_cores": "None", "memory_bandwidth": "384 GB/s"},
+    "RX 7900 XTX": {"vram": 24, "performance": "ultra", "architecture": "RDNA 3", "tensor_cores": "None", "memory_bandwidth": "960 GB/s"},
 }
 
 def get_gpu_recommendations(gpu_name, ram_gb):