mlx-community/DeepSeek-V4-Pro-Qwen3.5-4B-bf16
Partial fine-tune vs Jackrong/DeepSeek-V4-Pro-Qwen3.5-4B. Partial change: 0% of the weights changed (norms), in 32 of 32 layers.
Tensors changed: Embeddings 0/1, MLP 0/96, Attention 0/200, Norms 81/105
Partial fine-tune vs Jackrong/DeepSeek-V4-Pro-Qwen3.5-4B. Partial change: 0% of the weights changed (norms), in 32 of 32 layers.
Tensors changed: Embeddings 0/1, MLP 0/96, Attention 0/200, Norms 81/105