yanolja/YanoljaNEXT-EEVE-Instruct-2.8B
Full fine-tune vs yanolja/YanoljaNEXT-EEVE-2.8B. Fine-tune: 95% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 0/1, Norms 0/66, MLP 64/128, Attention 128/256, lm_head 2/2
Full fine-tune vs yanolja/YanoljaNEXT-EEVE-2.8B. Fine-tune: 95% of the weights changed, across 32 of 32 layers.
Tensors changed: Embeddings 0/1, Norms 0/66, MLP 64/128, Attention 128/256, lm_head 2/2