ProCreations/grug-27b
LoRA-style vs Qwen/Qwen3.6-27B. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of down_proj, gate_proj, in_proj_a, in_proj_b, in_proj_qkv, in_proj_z, 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: Embeddings 0/4, MLP 192/304, Attention 304/556, lm_head 0/1, Norms 0/319
7 derivatives checked
- ProCreations/grug-27b-gguf: Quantization
- ProCreations/grug-27b-qat-q4-gguf: Quantization
- bartowski/ProCreations_grug-27b-GGUF: LoRA-style
- mradermacher/grug-27b-i1-GGUF: Quantization
- mradermacher/grug-27b-GGUF: Quantization
- ProCreations/grug-27b-mtp: Same weights, repacked
- Lasimeri/GrugCap-27B: LoRA-style