Whisper

This is an implementation of Whisper built on top of the two most best improvements to the original Whisper algorithm.

The biggest improvements are around hallucinations, timestamp accuracy, and ability to detect filler words.

Picking the right settings

decode_boost is basically the option that switches between which implementation to use during the decoding process. Setting it to true will use the whisper-timestamped implementation and setting it to false will use the stable-ts implementation. whisper-timestamped is slower but more accurate specifically on timestamps and filler words while stable-ts is faster but less accurate on timestamps. However, it is key to note that both of these approaches make significant improvements to the timestamps you might get from the original Whisper implementation or other implementations such as WhisperX.
Enabling speaker_diarization returns speaker IDs for each word in the transcript. This is useful if you want to know who said what.
Enabling speed_boost will use smaller versions of Whisper with either decoding approach. This is useful if you want to get results faster and don't mind sacrificing some accuracy. When decode_boost is set to true, the base model is used and when decode_boost is set to false, the whisper-v3-turbo model is used.
initial_prompt basically allows you to include uncommon words that might appear in your audio. This is useful if you want to improve the accuracy and spelling on these types of words.

Languages

Whisper supports 99 total languages. You may enter a language code into the language parameter in case you know the language of the original audio already. If you don't know the language of the original audio, you may leave the language parameter blank and Whisper will automatically detect the language of the original audio. If you want to see the full list of supported languages, you may refer to the table below.

en (English)
zh (Chinese)
de (German)
es (Spanish)
ru (Russian)
ko (Korean)
fr (French)
ja (Japanese)
pt (Portuguese)
tr (Turkish)
pl (Polish)
ca (Catalan)
nl (Dutch)
ar (Arabic)
sv (Swedish)
it (Italian)
id (Indonesian)
hi (Hindi)
fi (Finnish)
vi (Vietnamese)
he (Hebrew)
uk (Ukrainian)
el (Greek)
ms (Malay)
cs (Czech)
ro (Romanian)
da (Danish)
hu (Hungarian)
ta (Tamil)
no (Norwegian)
th (Thai)
ur (Urdu)
hr (Croatian)
bg (Bulgarian)
lt (Lithuanian)
la (Latin)
mi (Maori)
ml (Malayalam)
cy (Welsh)
sk (Slovak)
te (Telugu)
fa (Persian)
lv (Latvian)
bn (Bengali)
sr (Serbian)
az (Azerbaijani)
sl (Slovenian)
kn (Kannada)
et (Estonian)
mk (Macedonian)
br (Breton)
eu (Basque)
is (Icelandic)
hy (Armenian)
ne (Nepali)
mn (Mongolian)
bs (Bosnian)
kk (Kazakh)
sq (Albanian)
sw (Swahili)
gl (Galician)
mr (Marathi)
pa (Punjabi)
si (Sinhala)
km (Khmer)
sn (Shona)
yo (Yoruba)
so (Somali)
af (Afrikaans)
oc (Occitan)
ka (Georgian)
be (Belarusian)
tg (Tajik)
sd (Sindhi)
gu (Gujarati)
am (Amharic)
yi (Yiddish)
lo (Lao)
uz (Uzbek)
fo (Faroese)
ps (Pashto)
tk (Turkmen)
nn (Nynorsk)
mt (Maltese)
sa (Sanskrit)
lb (Luxembourgish)
my (Myanmar)
bo (Tibetan)
tl (Tagalog)
mg (Malagasy)
as (Assamese)
tt (Tatar)
haw (Hawaiian)
ln (Lingala)
ha (Hausa)
ba (Bashkir)
jw (Javanese)
su (Sundanese)
yue (Cantonese)
my (Burmese)
ca (Valencian)
nl (Flemish)
ht (Haitian)
lb (Letzeburgesch)
ps (Pushto)
pa (Panjabi)
ro (Moldavian)
si (Sinhalese)
es (Castilian)
zh (Mandarin)

MORE EXAMPLES

No examples found. If you own this app, try going to a job and saving it as an example.