marcuscedricridia/Cheng-2-Base
Full fine-tune vs Qwen/Qwen2.5-14B. Fine-tune: 99% of the weights changed, across 48 of 48 layers.
Tensors changed: Norms 71/97, MLP 143/144, Attention 324/336
2 derivatives checked
- mradermacher/Cheng-2-Base-i1-GGUF: Quantization
- mradermacher/Cheng-2-Base-GGUF: Quantization