Blackroot/Mirai-70B-1.0
Full fine-tune vs Bllossom/llama-3-Korean-Bllossom-70B. Fine-tune: 100% of the weights changed, across 80 of 80 layers.
Tensors changed: Embeddings 1/1, Norms 161/161, MLP 240/240, Attention 320/320, lm_head 1/1
5 derivatives checked
- mradermacher/Mirai-70B-1.0-i1-GGUF: Quantization
- mradermacher/Mirai-70B-1.0-GGUF: Quantization
- KaraKaraModel/Llama-SoraEToAtetaTegami-3.1X-70B: Full fine-tune
- KaraKaraWarehouse/Llama-MiraiFanfare-3.3-70B: Full fine-tune
- mergekit-community/Llama-SoraEToAtetaTegami-3.1X-PROPER-70B: Full fine-tune