Delta-Vector/Austral-70B-Preview
LoRA-style vs Sao10K/Llama-3.3-70B-Vulpecula-r1. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, k_proj, o_proj, q_proj, up_proj, v_proj changed, in 80 of 80 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Norms 0/161, Embeddings 0/1, MLP 240/240, Attention 320/320
5 derivatives checked
- mradermacher/Austral-70B-Preview-i1-GGUF: Quantization
- mradermacher/Austral-70B-Preview-GGUF: Quantization
- KaraKaraWarehouse/Llama-EveningMirai-Moonwalker-MS-3.3-70B: Full fine-tune
- Delta-Vector/Austral-70B-Winton: LoRA-style
- KaraKaraWarehouse/Llama-EveningMirai-Moonwalker-3.3-70B: Full fine-tune