jaweed123/TinyJLLM-Instruct-DPO
Full fine-tune vs jaweed123/TinyJLLM-Instruct. Fine-tune: 100% of the weights changed, across 11 of 11 layers.
Tensors changed: Embeddings 1/1, Norms 23/23, MLP 33/33, Attention 44/44
Full fine-tune vs jaweed123/TinyJLLM-Instruct. Fine-tune: 100% of the weights changed, across 11 of 11 layers.
Tensors changed: Embeddings 1/1, Norms 23/23, MLP 33/33, Attention 44/44