martinsu/tildeopen-30b-mu-instruct
Full fine-tune vs TildeAI/TildeOpen-30b. Fine-tune: 100% of the weights changed, across 60 of 60 layers.
Tensors changed: Norms 120/121, MLP 180/180, Attention 240/240
Full fine-tune vs TildeAI/TildeOpen-30b. Fine-tune: 100% of the weights changed, across 60 of 60 layers.
Tensors changed: Norms 120/121, MLP 180/180, Attention 240/240