MuXodious/gemma-3n-E4B-it-less-heretic
Partial fine-tune vs google/gemma-3n-E4B-it. Partial change: 11% of the weights changed (attention, mlp, other), in 26 of 35 layers.
Tensors changed: Attention 4/363, Embeddings 0/25, Norms 0/489, other 3/687, MLP 26/112
2 derivatives checked
- mradermacher/gemma-3n-E4B-it-less-heretic-i1-GGUF: Quantization (fewer layers)
- mradermacher/gemma-3n-E4B-it-less-heretic-GGUF: Same weights, converted (fewer layers)