BetaRare
ARahim3/mlx-dspark
Up to 4× faster LLM decoding on Apple Silicon, lossless. Native MLX port of DeepSeek's DSpark & z-lab's DFlash speculative decoding — Gemma-4, Qwen3.8, Muse-Glimmer, Nemotron, LFM2.5, Ornith-1.0, ternary Bonsai-27B.
- Owner
- ARahim3 (personal account)
- Stars
- Stars: 734
- Forks
- Forks: 63
- Language
- Python
- License
- MIT
- Created
- 2026
- Topics
- apple-siliconclaude-codecodexdflash2dsparkinference-enginellm-inferencelocal-llmmacosmetal
- Tier
- Beta (700 to 1,999 ★) What this tier means
More repos by ARahim3
All 5 repos by ARahim3- ARahim3/DigUpStars: 6,144SwiftStable
- ARahim3/mlx-tuneStars: 1,457PythonBeta
- ARahim3/kaggle-tpu-labStars: 758PythonBeta
- ARahim3/LekhoStars: 128SwiftPrototype
Nearby in Python
All Python repos- facebookresearch/EUPEStars: 734PythonBeta
- yzhao062/agent-styleStars: 734PythonBeta
- chrissotraidis/kartpadStars: 734PythonBeta
- itdxer/neupyStars: 734PythonBeta
- czheo/syntax_sugar_pythonStars: 734PythonBeta
- Cadene/vqa.pytorchStars: 734PythonBeta