โ† All examples

ModelEthio-ASR Amharic (600M)3Ethio-ASR Multi (300M)12Ethio-ASR Multi (600M)5Ethio-ASR Multi (94M)6Meta MMS-1B10OmniASR-CTC (1B)4OmniASR-CTC (300M)4OmniASR-CTC (3B)3OmniASR-LLM (1B)4OmniASR-LLM (300M)4OmniASR-LLM (3B)5OpenAI Whisper-small10
Audio qualityClean Studio6Ambient Room Noise7Harsh Environment9Distant / Low-Volume Mic7Wideband Phone (VoLTE)9Randomized Phone Channel12Narrowband 2G + Packet Loss18

What did the model get wrong?

Ethio-ASR Multi (300M) Randomized Phone Channel   Example 12 of 20  ยท  speaker: Female

๐ŸŽง Listen โ€” Randomized Phone Channel
Phone-call audio with random extra distortions layered on top.
โœ… What was actually said
แŠ แˆแŠ• แ‹ซแˆˆแˆแ‰ต แŠ แ‹ณแˆ› แŠแ‹ แ‹ˆแ‹ฐ แŒ…แ‰กแ‰ฒ แˆˆแˆตแˆซ แˆแˆ„แ‹ต แˆตแˆˆแˆ†แŠ แŠแ‹
๐Ÿค– What this model heard
amh แŠ แˆ แ‹ซแˆˆแ‹ แ‰ณแ‹ณแ‰ฃ แŠแ‹ แ‰ แ‹ฐ แ‰น แ‰กแ‰ต แˆ‹ แˆตแˆซ แ‹ตแˆ„แ‹ฑแˆต แˆŠแˆ†แŠ แ‹ซแˆˆ
Correct
Wrong (different word)
Missed (skipped)
Extra (added)
1
Correct
9
Wrong
0
Missed
3
Extra
10%
Words right
Out of 10 spoken words, this model got 1 right (10%). Errors: 9 words wrong; 3 extra.

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