cross-encoder/ms-marco-MiniLM-L4-v2
Full fine-tune (fewer layers) vs cross-encoder/ms-marco-MiniLM-L12-v2. cross-encoder/ms-marco-MiniLM-L4-v2 matches part of cross-encoder/ms-marco-MiniLM-L12-v2 (74 of its 202 tensors; the rest of cross-encoder/ms-marco-MiniLM-L12-v2 has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 4 of 4 layers.
Tensors changed: Embeddings 5/6, Norms 16/16, Attention 32/32, other 19/20
1 derivatives checked
- mradermacher/ms-marco-MiniLM-L4-v2-GGUF: Not comparable yet
2 other repos hold byte-for-byte identical weight files.