ArliAI/DS-R1-Distill-70B-ArliAI-RpR-v4-Large
LoRA-style vs deepseek-ai/DeepSeek-R1-Distill-Llama-70B. 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: Embeddings 0/1, Attention 320/320, MLP 240/240, Norms 0/161, lm_head 0/1
4 derivatives checked
- mradermacher/DS-R1-Distill-70B-ArliAI-RpR-v4-Large-i1-GGUF: Quantization
- mradermacher/DS-R1-Distill-70B-ArliAI-RpR-v4-Large-GGUF: Quantization
- featherless-ai-quants/ArliAI-DS-R1-Distill-70B-ArliAI-RpR-v4-Large-GGUF: Quantization
- TareksLab/Mithril-Prose-LLaMa-70B: Full fine-tune
1 other repos hold byte-for-byte identical weight files.