rostlabs/rost-1b-instruct-v2
Full fine-tune vs rostlabs/rost-1b-base. Fine-tune: 100% of the weights changed, across 24 of 24 layers.
Tensors changed: other 5/5, lm_head 1/1, Attention 108/108, MLP 48/48, Embeddings 13/13
Full fine-tune vs rostlabs/rost-1b-base. Fine-tune: 100% of the weights changed, across 24 of 24 layers.
Tensors changed: other 5/5, lm_head 1/1, Attention 108/108, MLP 48/48, Embeddings 13/13