Hcompany/NeoMME-800M-Retriever
Full fine-tune vs Hcompany/NeoMME-800M. Fine-tune: 94% of the weights changed, across 20 of 20 layers.
Tensors changed: Embeddings 8/8, other 17/20, MLP 36/40, Attention 120/120
2 derivatives checked
- Hcompany/NeoMME-800M-Retriever-ST-dense: Same weights, repacked
- Hcompany/NeoMME-800M-Retriever-ST-late: Same weights, repacked