multimodal-reasoning-lab/Anole-Zebra-CoT

Full fine-tune vs GAIR/Anole-7b. Fine-tune: 99% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 1/2, Norms 193/317, MLP 96/96, Attention 128/144, lm_head 1/1, other 0/152

2 derivatives checked

• mradermacher/Anole-Zebra-CoT-i1-GGUF: Full fine-tune (fewer layers)
• mradermacher/Anole-Zebra-CoT-GGUF: Partial fine-tune (fewer layers)

On the Hub · Downloads on Model Pulse