aiuser3993/AMD-OLMo-1B-IT
LoRA-style vs amd/AMD-OLMo-1B-SFT. 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 16 of 16 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/1, MLP 48/48, Attention 64/64
2 derivatives checked
- mradermacher/AMD-OLMo-1B-IT-i1-GGUF: Quantization
- mradermacher/AMD-OLMo-1B-IT-GGUF: Same weights, converted