CultriX/Qwen2.5-14B-Wernickev3
Full fine-tune vs CultriX/Qwen2.5-14B-Wernicke. Fine-tune: 100% of the weights changed, across 48 of 48 layers.
Tensors changed: Norms 97/97, MLP 144/144, Attention 336/336
7 derivatives checked
- mradermacher/Qwen2.5-14B-Wernickev3-GGUF: Quantization
- CultriX/Qwen2.5-14B-Brocav6: Full fine-tune
- CultriX/Qwen2.5-14B-Hyperionv4: Full fine-tune
- CultriX/Qwen2.5-14B-Unity: Full fine-tune
- CultriX/Qwen2.5-14B-BrocaV9: Full fine-tune
- CultriX/Qwen2.5-14B-Brocav3: Full fine-tune
- CultriX/Qwen2.5-14B-Hyperionv3: Full fine-tune