BELLE-2/Belle-whisper-large-v3-zh-punct
LoRA-style vs BELLE-2/Belle-whisper-large-v3-zh. Looks like a merged LoRA (or a fine-tune with frozen norms): only the weight matrices of fc1, fc2, k_proj, out_proj, q_proj, v_proj changed, in 32 of 32 layers; norms, biases and embeddings are untouched.
Tensors changed: Embeddings 0/3, Norms 0/324, Attention 384/672, MLP 128/256, other 0/4
1 other repos hold byte-for-byte identical weight files.