ā† All examples

ModelEthio-ASR Amharic (600M)5Ethio-ASR Multi (300M)1Ethio-ASR Multi (600M)3Ethio-ASR Multi (94M)1Meta MMS-1B11OmniASR-CTC (1B)0OmniASR-CTC (300M)0OmniASR-CTC (3B)0OmniASR-LLM (1B)0OmniASR-LLM (300M)0OmniASR-LLM (3B)0OpenAI Whisper-small11
Audio qualityClean Studio11Ambient Room Noise11Harsh Environment11Distant / Low-Volume Mic11Wideband Phone (VoLTE)11Randomized Phone Channel11Narrowband 2G + Packet Loss12

What did the model get wrong?

Meta MMS-1B Ambient Room Noise   Example 16 of 20  Ā·  speaker: Male

šŸŽ§ Listen ā€” Ambient Room Noise
Realistic everyday background noise ā€” like a normal room or cafĆ©.
āœ… What was actually said
į‰įŒ„įˆ© įˆµį‹µįˆµį‰µ įˆ°į‰£į‰µ į‹˜įŒ įŠ įˆ¶įˆµį‰µ įˆįˆˆį‰µ įˆµį‹µįˆµį‰µ įˆ°į‰£į‰µ įˆµį‹µįˆµį‰µ įŠ įˆ«į‰µ įŠį‹
šŸ¤– What this model heard
į‰³įˆ˜ į‰  įˆ į‰µ
Correct
Wrong (different word)
Missed (skipped)
Extra (added)
0
Correct
4
Wrong
7
Missed
0
Extra
0%
Words right
Out of 11 spoken words, this model got 0 right (0%). Errors: 4 words wrong; 7 missed.

Try switching Audio quality ā€” see the same model degrade as the audio gets noisier.