WPS-Qingqiu/OmniAlign
Full fine-tune vs Alibaba-NLP/gte-multilingual-mlm-base. Fine-tune: 100% of the weights changed, across 12 of 12 layers.
Tensors changed: Embeddings 4/4, Attention 72/72, MLP 60/60
Full fine-tune vs Alibaba-NLP/gte-multilingual-mlm-base. Fine-tune: 100% of the weights changed, across 12 of 12 layers.
Tensors changed: Embeddings 4/4, Attention 72/72, MLP 60/60