froggeric/WestLake-10.7B-v2
Partial fine-tune (fewer layers) vs senseable/WestLake-7B-v2. senseable/WestLake-7B-v2 matches part of froggeric/WestLake-10.7B-v2 (291 of its 435 tensors; the rest of froggeric/WestLake-10.7B-v2 has no counterpart). On the shared part: Partial change: 69% of the weights changed (attention, mlp, norms), in 23 of 32 layers.
Tensors changed: Embeddings 0/1, Norms 46/65, MLP 69/96, Attention 92/128, lm_head 0/1
3 derivatives checked
- backyardai/WestLake-10.7B-v2-GGUF: Same weights, converted
- mradermacher/WestLake-10.7B-v2-GGUF: Quantization
- jsfs11/WestTemptressTensor-10.7B-v0.2a-SLERP: Full fine-tune
1 other repos hold byte-for-byte identical weight files.