ehristoforu/Gemma2-9b-it-train6-plus_layer

Partial fine-tune vs ehristoforu/Gemma2-9b-it-train6. Partial change: 3% of the weights changed (attention, mlp, norms), in 1 of 33 layers.

Tensors changed: Embeddings 0/1, Norms 4/133, MLP 3/99, Attention 4/132

On the Hub · Downloads on Model Pulse