tepirale/Ornith-Agents-A1-3.6-35B-A3B-dare_ties
Full fine-tune (fewer layers) vs InternScience/Agents-A1. tepirale/Ornith-Agents-A1-3.6-35B-A3B-dare_ties matches part of InternScience/Agents-A1 (946 of its 31666 tensors; the rest of InternScience/Agents-A1 has no counterpart). On the shared part: Fine-tune: 100% of the weights changed, across 40 of 40 layers.
Tensors changed: Embeddings 4/4, lm_head 1/1, Norms 241/241, Attention 375/388, MLP 309/312
3 derivatives checked
- tepirale/Ornith-Agents-A1-3.6-35B-A3B-MTP-GGUF: Same weights, converted
- tepirale/Ornith-Agents-A1-3.6-35B-A3B-GGUF-v2: Same weights, converted
- leonsarmiento/Ornith-Agents-A1-3.6-35B-A3B-dare_ties-6bit-XL-mlx: Quantization