yasserrmd/SciReason-LFM2-2.6B
LoRA-style vs LiquidAI/LFM2-2.6B. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of k_proj, out_proj, q_proj, v_proj, w1, w2, w3 changed, in 30 of 30 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/2, other 0/66, MLP 90/90, Norms 0/76, Attention 32/32
2 derivatives checked
- mradermacher/SciReason-LFM2-2.6B-i1-GGUF: Quantization
- mradermacher/SciReason-LFM2-2.6B-GGUF: Same weights, converted