aloobun/d-mxbai-L8-embed
Full fine-tune (fewer layers) vs mixedbread-ai/mxbai-embed-large-v1. aloobun/d-mxbai-L8-embed matches part of mixedbread-ai/mxbai-embed-large-v1 (135 of its 391 tensors; the rest of mixedbread-ai/mxbai-embed-large-v1 has no counterpart). On the shared part: Fine-tune: 99% of the weights changed, across 8 of 8 layers.
Tensors changed: Embeddings 5/5, Norms 32/32, Attention 64/64, other 32/34
1 derivatives checked
- aloobun/indic-mxbai-L8-embed: Full fine-tune