aloobun/indic-mxbai-L8-embed
Full fine-tune vs aloobun/d-mxbai-L8-embed. Fine-tune: 99% of the weights changed, across 8 of 8 layers.
Tensors changed: Embeddings 4/5, Norms 32/32, Attention 64/64, other 32/34
Full fine-tune vs aloobun/d-mxbai-L8-embed. Fine-tune: 99% of the weights changed, across 8 of 8 layers.
Tensors changed: Embeddings 4/5, Norms 32/32, Attention 64/64, other 32/34