MediaTek-Research/Breeze-ASR-26

Full fine-tune vs openai/whisper-large-v2. Fine-tune: 100% of the weights changed, across 32 of 32 layers.

Tensors changed: Embeddings 2/3, Norms 324/324, Attention 672/672, MLP 256/256, other 4/4

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

On the Hub · Downloads on Model Pulse