Tag
Z-lab releases DFlash for Qwen3.6-35B-A3B, a model fine-tuning/compression technique, with training complete and weights now available on GitHub and HuggingFace.