mirror of
https://github.com/signerless/llm-checker.git
synced 2026-10-19 09:54:45 +00:00
### New Features: - **🎯 --limit flag**: Show multiple compatible models (e.g., --limit 5) - **✅ Installation detection**: Automatically detects already installed Ollama models - **🧠 Smart categories**: 7 use cases with typo tolerance (coding, creative, reasoning, multimodal, embeddings, talking, general) - **📊 Real size data**: Uses actual Ollama model sizes instead of estimates - **🚀 Intelligent Quick Start**: Shows "ollama run" for installed models, "ollama pull" for new ones ### Enhanced Algorithm: - Multi-objective ranking system with hardware-size matching - Penalizes models that underutilize high-end hardware - Filters out unreasonably large models (>25GB) from recommendations - Smart alternative suggestions without duplicate commands ### Improved UX: - Clean, emoji-free progress display - Ranked model display (#1, #2, #3...) - Status indicators: "Already installed in Ollama" vs "Available for installation" - Hardware-aware Quick Start commands - Comprehensive model classification system ### Technical Improvements: - Static database (177 models) with pre-classified categories - Removed dynamic cache updates for stability - Enhanced model matching algorithms - Cross-platform compatibility (macOS, Windows, Linux) - Improved error handling and typo tolerance Signed-off-by: signerless <109688848+signerless@users.noreply.github.com> |
||
|---|---|---|
| .. | ||
| compatibility.js | ||
| performance.js | ||