ArliAI/QwQ-32B-ArliAI-RpR-v4
LoRA-style vs Qwen/QwQ-32B. 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 64 of 64 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/129, MLP 192/192, Attention 256/448, lm_head 0/1
8 derivatives checked
- bartowski/ArliAI_QwQ-32B-ArliAI-RpR-v4-GGUF: Same weights, converted
- mradermacher/QwQ-32B-ArliAI-RpR-v4-i1-GGUF: Quantization
- ArliAI/QwQ-32B-ArliAI-RpR-v4-GGUF: Same weights, converted
- featherless-ai-quants/ArliAI-QwQ-32B-ArliAI-RpR-v4-GGUF: Quantization
- mradermacher/QwQ-32B-ArliAI-RpR-v4-GGUF: Quantization
- DevQuasar/ArliAI.QwQ-32B-ArliAI-RpR-v4-GGUF: Same weights, converted
- Phr00t/Phr00tyMix-v2-32B: Full fine-tune
- ParasiticRogue/RareBit-v2-32B: Full fine-tune