Nexesenex/Llama_3.x_70b_SmarTracks_v1.30_flat
Full fine-tune vs Nexesenex/Llama_3.3_70b_DeepSeek_R1_Dropable_V1.01. Fine-tune: 100% of the weights changed, across 80 of 80 layers.
Tensors changed: Norms 161/161, MLP 240/240, Attention 320/320
2 derivatives checked
- mradermacher/Llama_3.x_70b_SmarTracks_v1.30_flat-i1-GGUF: Quantization
- mradermacher/Llama_3.x_70b_SmarTracks_v1.30_flat-GGUF: Quantization