asa-global-sft-best

Headline global SFT model from the bachelor thesis Automatic Speech Assessment (Schmidt-Svejstrup & Lindahl, DTU 2026). Descriptive MOS-captioning, Table 5.x (in-domain MSE 0.193, BLEU 25.51). Fine-tuned from Qwen2-Audio-7B. Code: https://github.com/DavidLindahl/automatic-speech-assessment

Downloads last month
26
Safetensors
Model size
8B params
Tensor type
BF16
·
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support