Sakalti/Saba1.5-Pro-3B
Partial fine-tune (fewer layers) vs Sakalti/Saba1.5-Pro. Sakalti/Saba1.5-Pro matches part of Sakalti/Saba1.5-Pro-3B (338 of its 627 tensors; the rest of Sakalti/Saba1.5-Pro-3B has no counterpart). On the shared part: Partial change: 79% of the weights changed (attention, mlp, norms), in 26 of 28 layers.
Tensors changed: Embeddings 0/1, Norms 52/57, MLP 78/84, Attention 182/196
2 derivatives checked
- mradermacher/Saba1.5-Pro-3B-i1-GGUF: Quantization
- mradermacher/Saba1.5-Pro-3B-GGUF: Same weights, converted