DatasetLwaziANV

← All examples Model summary →

Modelfacebook/omniASR-CTC-1BOpen5facebook/omniASR-CTC-300MOpen10facebook/omniASR-CTC-3BOpen1facebook/omniASR-LLM-1BOpen2facebook/omniASR-LLM-300MOpen5facebook/omniASR-LLM-3BOpen1facebook/omniASR-LLM-7BOpen0dsfsi/anv-whisper-large-v3-turbo-anv-zulOpen4dsfsi/anv-whisper-small-anv-zulu-first-batchOpen6openai/whisper-large-v3Open36openai/whisper-smallOpen148sitwala/whisper-large-anv-sotOpen51sitwala/whisper-large-v3-anv-sotOpen56sitwala/whisper-large-v3-turbo-anv-zul-150hOpen4sitwala/whisper-large-v3-turbo-anv-zul-250hOpen7sitwala/whisper-large-v3-turbo-anv-zul-50hOpen5

What did the model get wrong?

sitwala/whisper-large-v3-anv-sotOpen   Example 19 of 100  ·  sample index 18

✅ What was actually said (isiZulu reference)
Ngenxayokuthiiziphilincincizihlanganisainingileziphiliezingunongqamuzanazazozontathuizidlangalazempilolokhukwenzazingafani.IsidlangalasesithathuesiyiNswelandenisihlanganisazonkeiziphiliezingumangqamuzanakanyenezinanjazanenezinambiezingunongqamuzanaoyigciwane.
🤖 What this model heard
ngengwaeokotsiesephilengwengweesehlanganisaenengwelesephileezengkunongamosanazazozongtatshoesehlangalazempelolehokwetsazengafaneesehlangalasesetshatsoesengtswelangdenesehlanganisazonkeesephileezengkunongamosanakanyetsonezinangjazanenezinang
Correct
Wrong (different word)
Missed (skipped)
Extra (added)
1
Correct
25
Wrong
0
Missed
31
Extra
4%
Words right
2.154
WER
0.479
CER
34.44s
Duration
0.3s
Latency