blascotobasco/LFM2-16B-A2B
LoRA-style (fewer layers) vs MuXodious/LFM2-8B-A1B-absolute-heresy-MPOA. MuXodious/LFM2-8B-A1B-absolute-heresy-MPOA matches part of blascotobasco/LFM2-16B-A2B (2258 of its 4414 tensors; the rest of blascotobasco/LFM2-16B-A2B has no counterpart). On the shared part: Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of w1, w2, w3 changed, in 22 of 24 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/2, other 0/54, MLP 1418/1424, Norms 0/60, Attention 0/24
2 derivatives checked
- mradermacher/LFM2-16B-A2B-i1-GGUF: Quantization
- mradermacher/LFM2-16B-A2B-GGUF: Quantization