Unsloth releases Kimi K3 quantizations, shrinking 1.56TB to 594GB
Unsloth released 8/4/2/1-bit GGUF versions of Kimi K3: Q8 is lossless at 1.56TB, Q4 at 1.51TB, Q2 at 861GB, and Q1 cuts to 594GB while retaining 78.9% accuracy.
1 source
Daily brief
Get tomorrow's AI brief in your inbox
More stories today
- China AI Chip Designer Moore Threads Plans Hong Kong Listing
- Ineffable Intelligence founder David Silver pledges company sale to charity
- ChatGPT Finance announced to help you save money
- Evaluation benchmarks released for model performance testing
- Resource explains LLM tokenization, Byte Pair Encoding, and attention math