MERALION-AudioLLM & MERALION-AudioLLM-Intruction-Following
#5
by
jiada16011994
- opened
I tried the demo and found that MERALION-AudioLLM (more accurate) performs better than MERALION-AudioLLM-Intruction-Following in term of transcribing. Are these two different models, or are they just configured differently using the same model?
Hi, they are two models with different weights. MERaLiON-AudioLLM should perform better on transcription and some accent/gender recognition tasks, while the instruction following version could handle more flexible tasks. We will release more details soon!
I see! Will you be releasing the other model as well?