neopolita/Qwen3.8-Flash-Next-102B-A5B-Niwaki-v2.4-3bit-mlx
Partial fine-tune vs mlx-community/Qwen3.8-Flash-Next-4bit. Partial change: 5% of the weights changed (mlp), in 31 of 31 layers.
Tensors changed: Embeddings 0/6, Attention 0/812, Norms 0/120, MLP 279/744, other 0/7