Nexesenex/Llama_3.x_70b_SmarTricks_V1.01
Full fine-tune vs NexesMess/Llama_3.x_70b_SmarTricks_0.11. Fine-tune: 100% of the weights changed, across 80 of 80 layers.
Tensors changed: Norms 160/161, MLP 240/240, Attention 320/320
3 derivatives checked
- mradermacher/Llama_3.x_70b_SmarTricks_V1.01-i1-GGUF: Quantization
- mradermacher/Llama_3.x_70b_SmarTricks_V1.01-GGUF: Quantization
- NexesMess/Llama_3.x_70b_TessEssence_0.10c: Full fine-tune