DrRiceIO7/granite-4.0-micro-heretic
LoRA-style vs ibm-granite/granite-4.0-micro. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of o_proj, output_linear changed, in 22 of 40 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, Norms 0/81, Attention 22/160, MLP 6/80
2 derivatives checked
- mradermacher/granite-4.0-micro-heretic-i1-GGUF: Quantization
- mradermacher/granite-4.0-micro-heretic-GGUF: Same weights, converted