Kimi K3 Draft Collection released: EAGLE-3, DFlash2 and DSpark draft models trained on GB200
hongyangzh · x · 2026-09-03
lightseekorg released the Kimi K3 Draft Collection, a set of three draft models for speculative decoding — EAGLE-3, DFlash2 and DSpark — trained with TorchSpec and vLLM on NVIDIA GB200 hardware. The team also published the data recipes used for training, with further details available on their blog.
Related event: Three Open-Source Draft Models for Kimi K3 Released(3 posts)→
More from Infra
- Corporate token spending growth collapses from 42% to 0.2% as hyperscaler ROIC falls to 20% — SumitGup · 2026-10-21
- Qualcomm plans worldwide performance-based layoffs as it chases $5B AI data center revenue — rohanpaul_ai · 2026-10-21
- vLLM at PyTorchCon 2026: KV cache and kernel optimizations across TPU, AMD, and NVIDIA — PyTorch · 2026-10-21
- antirez to port DeepSeek Infer kernels to DGX Spark's SM121 for DwarfStar — antirez · 2026-10-21
- Ternary-Bonsai-2-9B: 2-bit ternary vision-language model trends on Hugging Face for on-device use — prism-ml · 2026-10-21
- General Compute x Positron: GPUs Handle Prefill, Memory-First Chips Tackle Decode-Bound LLM Inference — cneuralnetwork · 2026-10-21