169 4 days ago

Qwen/Qwen3.6-27B + 3 Qwen3.6 fine-tunes with MLP-passthrough surgery - MTP quants

8 models