Tivaphraen/Geryon-9B-v1

Partial fine-tune (fewer layers) vs Tesslate/OmniCoder-9B. Tivaphraen/Geryon-9B-v1 matches part of Tesslate/OmniCoder-9B (427 of its 760 tensors; the rest of Tesslate/OmniCoder-9B has no counterpart). On the shared part: Partial change: 89% of the weights changed (attention, lm_head, mlp, norms), in 32 of 32 layers.

Tensors changed: lm_head 1/1, Embeddings 0/1, Norms 56/105, Attention 176/224, MLP 96/96

3 derivatives checked

On the Hub · Downloads on Model Pulse