rohansheth/tiro-qwen3-asr-1.7b-v2

Partial fine-tune vs Qwen/Qwen3-ASR-1.7B-hf. Partial change: 16% of the weights changed (attention, mlp, norms, other), in 24 of 28 layers.

Tensors changed: other 11/11, MLP 96/180, Norms 58/211, Attention 192/304, Embeddings 0/1

On the Hub · Downloads on Model Pulse