armand0e/Granite-4.2-30B-Fable-Distill
LoRA-style vs ibm-granite/granite-4.2-30b. 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: lm_head 0/1, Embeddings 0/1, Norms 0/129, MLP 192/192, Attention 251/256
2 derivatives checked
- mradermacher/Granite-4.2-30B-Fable-Distill-i1-GGUF: Quantization
- mradermacher/Granite-4.2-30B-Fable-Distill-GGUF: Quantization