katanemo/Arch-Router-1.5B
Partial fine-tune vs Qwen/Qwen2.5-1.5B-Instruct. Partial change: 85% of the weights changed (attention, mlp, norms), in 28 of 28 layers.
Tensors changed: Embeddings 0/1, Norms 43/57, MLP 84/84, Attention 195/196
6 derivatives checked
- mradermacher/Arch-Router-1.5B-GGUF: Same weights, converted
- prithivMLmods/Arch-Router-1.5B-GGUF: Same weights, converted
- yasserrmd/Arch-Router-1.5B-gguf: Same weights, converted
- DevQuasar/katanemo.Arch-Router-1.5B-GGUF: Same weights, converted
- mlx-community/Arch-Router-1.5B-mlx-8Bit: Quantization
- tensorblock/katanemo_Arch-Router-1.5B-GGUF: Quantization
1 other repos hold byte-for-byte identical weight files.