drowzeys/keys-GLM-5.3-speculator.dspark-ft2
Full fine-tune vs RedHatAI/GLM-5.3-speculator.dspark. Fine-tune: 97% of the weights changed, across 3 of 3 layers.
Tensors changed: other 3/5, Norms 8/14, MLP 9/9, Attention 12/12
Full fine-tune vs RedHatAI/GLM-5.3-speculator.dspark. Fine-tune: 97% of the weights changed, across 3 of 3 layers.
Tensors changed: other 3/5, Norms 8/14, MLP 9/9, Attention 12/12