Jackrong/Qwopus3.5-27B-v3
LoRA-style vs unsloth/Qwen3.5-27B. 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, out_proj, q_proj, up_proj, v_proj changed, in 64 of 64 layers; norms, biases and embeddings are untouched.
Tensors changed: lm_head 0/1, Embeddings 0/4, MLP 192/304, Attention 112/556, Norms 0/319
9 derivatives checked
- mradermacher/Qwopus3.5-27B-v3-i1-GGUF: Quantization
- mradermacher/Qwopus3.5-27B-v3-GGUF: Quantization
- croll83/Qwopus3.5-27B-v3-Abliterated: LoRA-style
- h34v7/Jackrong-Qwopus3.5-27B-v3-GGUF: Quantization
- Jackrong/MLX-Qwopus3.5-27B-v3-4bit: Quantization
- Stabhappy/Qwopus-3.5-27B-V3-Stabbed-GGUF: Full fine-tune
- YTan2000/Qwopus3.5-27B-v3-TQ3_4S: Quantization
- QuantTrio/Qwopus3.5-27B-v3-AWQ: Quantization
- dangell7/Condor-27B: Full fine-tune
1 other repos hold byte-for-byte identical weight files.