Soaperloafidksum/LOREA-cyber-coder-30B-A3B-v5.1

Partial fine-tune vs mlx-community/Qwen3-Coder-30B-A3B-Instruct-4bit. Partial change: 14% of the weights changed (attention, mlp), in 8 of 48 layers.

Tensors changed: Embeddings 0/3, Norms 0/193, MLP 81/576, Attention 84/576, lm_head 0/3

1 other repos hold byte-for-byte identical weight files.

On the Hub · Downloads on Model Pulse