nisten/shqiponja-59b-v1

Partial fine-tune (fewer layers) vs jondurbin/nontoxic-bagel-34b-v0.2. jondurbin/nontoxic-bagel-34b-v0.2 matches part of nisten/shqiponja-59b-v1 (543 of its 939 tensors; the rest of nisten/shqiponja-59b-v1 has no counterpart). On the shared part: Partial change: 14% of the weights changed (attention, lm_head, mlp, norms), in 8 of 60 layers.

Tensors changed: Embeddings 0/1, Norms 17/121, MLP 24/180, Attention 32/240, lm_head 1/1

2 derivatives checked

On the Hub · Downloads on Model Pulse