altslate/JugnuLM-110M-R2plus
Full fine-tune vs altslate/JugnuLM-110M. Fine-tune: 100% of the weights changed, across 23 of 23 layers.
Tensors changed: Embeddings 1/1, Norms 93/93, MLP 69/69, Attention 92/92
Full fine-tune vs altslate/JugnuLM-110M. Fine-tune: 100% of the weights changed, across 23 of 23 layers.
Tensors changed: Embeddings 1/1, Norms 93/93, MLP 69/69, Attention 92/92