← All examples

ModelEthio-ASR Amharic (600M)1Ethio-ASR Multi (300M)4Ethio-ASR Multi (600M)3Ethio-ASR Multi (94M)4Meta MMS-1B9OmniASR-CTC (1B)2OmniASR-CTC (300M)7OmniASR-CTC (3B)0OmniASR-LLM (1B)7OmniASR-LLM (300M)6OmniASR-LLM (3B)2OpenAI Whisper-small19
Audio qualityClean Studio7Ambient Room Noise7Harsh Environment7Distant / Low-Volume Mic7Wideband Phone (VoLTE)7Randomized Phone Channel7Narrowband 2G + Packet Loss9

What did the model get wrong?

OmniASR-CTC (300M) Ambient Room Noise   Example 6 of 20  Β·  speaker: Male

🎧 Listen β€” Ambient Room Noise
Realistic everyday background noise β€” like a normal room or cafΓ©.
βœ… What was actually said
αˆ˜α‰Έ αŠα‹ α‹¨αˆšαˆ΅α‰°αŠ«αŠ¨αˆˆα‹ ሡራ αˆ΅αˆ‹αˆˆα‰₯ኝ α‰ΆαˆŽ α‰’αˆ†αŠ•αˆαŠ αŒ₯ሩ αŠα‹
πŸ€– What this model heard
αˆ˜α‰ΈαŠ• α‹¨αˆšαˆ΅α‰°αŠ«αŠ¨αˆˆα‹ ሡራ αˆ΅αˆ‹αˆˆ α‰’αŠ α‰°αˆ†αˆŽ α‰’αˆ†αŠ ልኝ αŒ₯ሩ αŠα‹
Correct
Wrong (different word)
Missed (skipped)
Extra (added)
4
Correct
4
Wrong
1
Missed
2
Extra
44%
Words right
Out of 9 spoken words, this model got 4 right (44%). Errors: 4 words wrong; 1 missed; 2 extra.

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