yujiepan/gemma-4-e-tiny-random
Full fine-tune vs google/gemma-4-E4B-it. Fine-tune: 100% of the weights changed, across 4 of 4 layers.
Tensors changed: MLP 56/56, other 20/20, Attention 66/66, Norms 2/2
Full fine-tune vs google/gemma-4-E4B-it. Fine-tune: 100% of the weights changed, across 4 of 4 layers.
Tensors changed: MLP 56/56, other 20/20, Attention 66/66, Norms 2/2