Qwen/Qwen3-14B-MLX-bf16

Full fine-tune vs Qwen/Qwen3-14B-Base. Fine-tune: 100% of the weights changed, across 40 of 40 layers.

Tensors changed: Embeddings 1/1, Norms 158/161, MLP 120/120, Attention 160/160, lm_head 1/1

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

On the Hub · Downloads on Model Pulse