deepcrayon/AniMUL-v1
Partial fine-tune vs Qwen/Qwen3-Omni-30B-A3B-Instruct. Partial change: 67% of the weights changed (attention, lm_head, mlp, norms, other), in 48 of 48 layers.
Tensors changed: other 7/148, MLP 950/1557, Norms 307/562, Attention 448/696, Embeddings 0/21, lm_head 1/16
1 derivatives checked
- mradermacher/AniMUL-v1-GGUF: Not comparable yet