Vikhrmodels/QVikhr-2.5-1.5B-Instruct-r
Full fine-tune vs Qwen/Qwen2.5-1.5B. Fine-tune: 100% of the weights changed, across 28 of 28 layers.
Tensors changed: Norms 57/57, MLP 84/84, Attention 196/196
6 derivatives checked
- Vikhrmodels/QVikhr-2.5-1.5B-Instruct-r_GGUF: Same weights, converted
- csabakecskemeti/Vikhrmodels.QVikhr-2.5-1.5B-Instruct-r-GGUF: Same weights, converted
- mradermacher/QVikhr-2.5-1.5B-Instruct-r-GGUF: Same weights, converted
- tensorblock/QVikhr-2.5-1.5B-Instruct-r-GGUF: Quantization
- MaziyarPanahi/QVikhr-2.5-1.5B-Instruct-r-GGUF: Same weights, converted
- RomanNeobutov/PhysicsOnBooks: LoRA-style