List of all possible model options

[blainek-ew]

I've only been able to glean a few model options from the discussion group. Is there a comprehensive list of available models to choose from? If not it might be good to post here for others to find.

[Purfview]

--help

[blainek-ew]

It does not show the options. I can guess small medium large and large-v2 (v3). A list somewhere would be helpful.
--model MODEL, -m MODEL
name of the Whisper model to use (default: medium)

[Purfview]

Oh yeah, I forgot it, just use any random string as model then in error you'll get all official models.

[blainek-ew]

nice trick!

Try these known models:
['tiny', 'tiny.en', 'base', 'base.en', 'small', 'small.en', 'medium', 'medium.en', 'large-v1', 'large-v2', 'large-v3', '
large', 'large-v3-turbo', 'turbo', 'distil-large-v2', 'distil-medium.en', 'distil-small.en', 'distil-large-v3']

[Rei-0011]

it's pro version

usage: faster-whisper-xxl.exe [-h] [--model MODEL] [--model_dir MODEL_DIR] [--device DEVICE] [--output_dir OUTPUT_DIR]
[--output_format [{json,lrc,txt,text,vtt,srt,tsv,all} ...]]
[--compute_type {default,auto,int8,int8_float16,int8_float32,int8_bfloat16,int16,float16,float32,bfloat16}] [--verbose VERBOSE]
[--task {transcribe,translate}]
[--language {af,am,ar,as,az,ba,be,bg,bn,bo,br,bs,ca,cs,cy,da,de,el,en,es,et,eu,fa,fi,fo,fr,gl,gu,ha,haw,he,hi,hr,ht,hu,hy,id,is,it,ja,jw,ka,kk,km,kn,ko,la,lb,ln,lo,lt,lv,mg,mi,mk,ml,mn,mr,ms,mt,my,ne,nl,nn,no,oc,pa,pl,ps,pt,ro,ru,sa,sd,si,sk,sl,sn,so,sq,sr,su,sv,sw,ta,te,tg,th,tk,tl,tr,tt,uk,ur,uz,vi,yi,yo,yue,zh,Afrikaans,Albanian,Amharic,Arabic,Armenian,Assamese,Azerbaijani,Bashkir,Basque,Belarusian,Bengali,Bosnian,Breton,Bulgarian,Burmese,Cantonese,Castilian,Catalan,Chinese,Croatian,Czech,Danish,Dutch,English,Estonian,Faroese,Finnish,Flemish,French,Galician,Georgian,German,Greek,Gujarati,Haitian,Haitian Creole,Hausa,Hawaiian,Hebrew,Hindi,Hungarian,Icelandic,Indonesian,Italian,Japanese,Javanese,Kannada,Kazakh,Khmer,Korean,Lao,Latin,Latvian,Letzeburgesch,Lingala,Lithuanian,Luxembourgish,Macedonian,Malagasy,Malay,Malayalam,Maltese,Mandarin,Maori,Marathi,Moldavian,Moldovan,Mongolian,Myanmar,Nepali,Norwegian,Nynorsk,Occitan,Panjabi,Pashto,Persian,Polish,Portuguese,Punjabi,Pushto,Romanian,Russian,Sanskrit,Serbian,Shona,Sindhi,Sinhala,Sinhalese,Slovak,Slovenian,Somali,Spanish,Sundanese,Swahili,Swedish,Tagalog,Tajik,Tamil,Tatar,Telugu,Thai,Tibetan,Turkish,Turkmen,Ukrainian,Urdu,Uzbek,Valencian,Vietnamese,Welsh,Yiddish,Yoruba}]
[--language_detection_threshold LANGUAGE_DETECTION_THRESHOLD] [--language_detection_segments LANGUAGE_DETECTION_SEGMENTS]
[--temperature TEMPERATURE] [--best_of BEST_OF] [--beam_size BEAM_SIZE] [--patience PATIENCE] [--length_penalty LENGTH_PENALTY]
[--repetition_penalty REPETITION_PENALTY] [--no_repeat_ngram_size NO_REPEAT_NGRAM_SIZE] [--suppress_blank SUPPRESS_BLANK]
[--suppress_tokens SUPPRESS_TOKENS] [--initial_prompt INITIAL_PROMPT] [--prefix PREFIX]
[--condition_on_previous_text CONDITION_ON_PREVIOUS_TEXT] [--prompt_reset_on_temperature PROMPT_RESET_ON_TEMPERATURE]
[--without_timestamps WITHOUT_TIMESTAMPS] [--max_initial_timestamp MAX_INITIAL_TIMESTAMP]
[--temperature_increment_on_fallback TEMPERATURE_INCREMENT_ON_FALLBACK] [--compression_ratio_threshold COMPRESSION_RATIO_THRESHOLD]
[--logprob_threshold LOGPROB_THRESHOLD] [--no_speech_threshold NO_SPEECH_THRESHOLD] [--word_timestamps WORD_TIMESTAMPS]
[--highlight_words HIGHLIGHT_WORDS] [--prepend_punctuations PREPEND_PUNCTUATIONS] [--append_punctuations APPEND_PUNCTUATIONS]
[--threads THREADS] [--version] [--vad_filter VAD_FILTER] [--vad_threshold VAD_THRESHOLD]
[--vad_min_speech_duration_ms VAD_MIN_SPEECH_DURATION_MS] [--vad_max_speech_duration_s VAD_MAX_SPEECH_DURATION_S]
[--vad_min_silence_duration_ms VAD_MIN_SILENCE_DURATION_MS] [--vad_speech_pad_ms VAD_SPEECH_PAD_MS]
[--vad_window_size_samples VAD_WINDOW_SIZE_SAMPLES]
[--vad_method {silero_v4_fw,silero_v5_fw,silero_v3,silero_v4,silero_v5,pyannote_v3,pyannote_onnx_v3,auditok,webrtc}] [--vad_dump]
[--vad_dump_aud] [--vad_device VAD_DEVICE] [--hallucination_silence_threshold HALLUCINATION_SILENCE_THRESHOLD]
[--hallucination_silence_th_temp {0.0,0.2,0.5,0.8,1.0}] [--clip_timestamps CLIP_TIMESTAMPS] [--max_new_tokens MAX_NEW_TOKENS]
[--chunk_length CHUNK_LENGTH] [--hotwords HOTWORDS] [--batched] [--batch_size BATCH_SIZE] [--multilingual MULTILINGUAL]
[--model_preload MODEL_PRELOAD] [--batch_recursive] [--beep_off] [--skip] [--checkcuda] [--print_progress] [--postfix]
[--check_files] [--alt_writer_off] [--ignore_dupe_prompt IGNORE_DUPE_PROMPT] [--hallucinations_list_off] [--v3_offsets_off]
[--reprompt REPROMPT] [--prompt_reset_on_no_end {0,1,2}] [--rehot] [--unmerged] [--japanese {blend,kanji,hiragana,katakana}]
[--one_word {0,1,2}] [--sentence] [--standard] [--standard_asia] [--max_comma MAX_COMMA]
[--max_comma_cent {20,30,40,50,60,70,80,90,100}] [--max_gap MAX_GAP] [--max_line_width MAX_LINE_WIDTH]
[--max_line_count MAX_LINE_COUNT] [--min_dist_to_end {0,4,5,6,7,8,9,10,11,12}] [--ff_dump] [--ff_track {1,2,3,4,5,6}] [--ff_fc]
[--ff_lc] [--ff_invert] [--ff_mp3] [--ff_sync] [--ff_rnndn_sh] [--ff_rnndn_xiph] [--ff_fftdn [0 - 97]] [--ff_tempo [0.5 - 2.0]]
[--ff_gate] [--ff_speechnorm] [--ff_loudnorm] [--ff_silence_suppress noise duration] [--ff_lowhighpass]
[--ff_vocal_extract {mdx1_kim2,mdx2_kim2,mb-roformer}] [--roformer_overlap [0.9 - 0.125]] [--roformer_vram {4,6,8,12,16}]
[--mdx_chunk MDX_CHUNK] [--voc_device VOC_DEVICE] [--realign] [--realign_device REALIGN_DEVICE]
[--diarize {pyannote_v3.0,pyannote_v3.1,reverb_v1,reverb_v2}] [--diarize_device DIARIZE_DEVICE] [--diarize_threads DIARIZE_THREADS]
[--diarize_dump] [--speaker SPEAKER] [--num_speakers NUM_SPEAKERS] [--min_speakers MIN_SPEAKERS] [--max_speakers MAX_SPEAKERS]
[--diarize_ff DIARIZE_FF] [--return_embeddings] [--diarize_only]
audio [audio ...]

positional arguments:
audio audio file(s). You can enter a file wildcard, filelist (txt. m3u, m3u8, lst) or directory to do batch processing. Note: non-media files in
list or directory are filtered out by extension.

options:
-h, --help show this help message and exit
--model MODEL, -m MODEL
name of the Whisper model to use (default: medium)
--model_dir MODEL_DIR
the path to save model files; uses C:\Faster-Whisper-XXL_models by default (default: None)
--device DEVICE, -d DEVICE
Device to use. Default is 'cuda' if CUDA device is detected, else is 'cpu'. If CUDA GPU is a second device then set 'cuda:1'. (default:
cuda)
--output_dir OUTPUT_DIR, -o OUTPUT_DIR
directory to save the outputs. By default the same folder where the executable file is or where media file is if --batch_recursive=True.
'.'- sets to the current folder. 'source' - sets to where media file is. (default: default)
--output_format [{json,lrc,txt,text,vtt,srt,tsv,all} ...], -f [{json,lrc,txt,text,vtt,srt,tsv,all} ...]
Format(s) of the output file; specify multiple formats separated by spaces. Note: dont add audio after this arg. (default: ['srt'])
--compute_type {default,auto,int8,int8_float16,int8_float32,int8_bfloat16,int16,float16,float32,bfloat16}, -ct {default,auto,int8,int8_float16,int8_float32,int8_bfloat16,int16,float16,float32,bfloat16}
Type of quantization to use (see https://opennmt.net/CTranslate2/quantization.html). (default: auto)
--verbose VERBOSE, -v VERBOSE
whether to print out debug messages (default: False)
--task {transcribe,translate}
whether to perform X->X speech recognition ('transcribe') or X->English translation ('translate') (default: transcribe)
--language {af,am,ar,as,az,ba,be,bg,bn,bo,br,bs,ca,cs,cy,da,de,el,en,es,et,eu,fa,fi,fo,fr,gl,gu,ha,haw,he,hi,hr,ht,hu,hy,id,is,it,ja,jw,ka,kk,km,kn,ko,la,lb,ln,lo,lt,lv,mg,mi,mk,ml,mn,mr,ms,mt,my,ne,nl,nn,no,oc,pa,pl,ps,pt,ro,ru,sa,sd,si,sk,sl,sn,so,sq,sr,su,sv,sw,ta,te,tg,th,tk,tl,tr,tt,uk,ur,uz,vi,yi,yo,yue,zh,Afrikaans,Albanian,Amharic,Arabic,Armenian,Assamese,Azerbaijani,Bashkir,Basque,Belarusian,Bengali,Bosnian,Breton,Bulgarian,Burmese,Cantonese,Castilian,Catalan,Chinese,Croatian,Czech,Danish,Dutch,English,Estonian,Faroese,Finnish,Flemish,French,Galician,Georgian,German,Greek,Gujarati,Haitian,Haitian Creole,Hausa,Hawaiian,Hebrew,Hindi,Hungarian,Icelandic,Indonesian,Italian,Japanese,Javanese,Kannada,Kazakh,Khmer,Korean,Lao,Latin,Latvian,Letzeburgesch,Lingala,Lithuanian,Luxembourgish,Macedonian,Malagasy,Malay,Malayalam,Maltese,Mandarin,Maori,Marathi,Moldavian,Moldovan,Mongolian,Myanmar,Nepali,Norwegian,Nynorsk,Occitan,Panjabi,Pashto,Persian,Polish,Portuguese,Punjabi,Pushto,Romanian,Russian,Sanskrit,Serbian,Shona,Sindhi,Sinhala,Sinhalese,Slovak,Slovenian,Somali,Spanish,Sundanese,Swahili,Swedish,Tagalog,Tajik,Tamil,Tatar,Telugu,Thai,Tibetan,Turkish,Turkmen,Ukrainian,Urdu,Uzbek,Valencian,Vietnamese,Welsh,Yiddish,Yoruba}, -l {af,am,ar,as,az,ba,be,bg,bn,bo,br,bs,ca,cs,cy,da,de,el,en,es,et,eu,fa,fi,fo,fr,gl,gu,ha,haw,he,hi,hr,ht,hu,hy,id,is,it,ja,jw,ka,kk,km,kn,ko,la,lb,ln,lo,lt,lv,mg,mi,mk,ml,mn,mr,ms,mt,my,ne,nl,nn,no,oc,pa,pl,ps,pt,ro,ru,sa,sd,si,sk,sl,sn,so,sq,sr,su,sv,sw,ta,te,tg,th,tk,tl,tr,tt,uk,ur,uz,vi,yi,yo,yue,zh,Afrikaans,Albanian,Amharic,Arabic,Armenian,Assamese,Azerbaijani,Bashkir,Basque,Belarusian,Bengali,Bosnian,Breton,Bulgarian,Burmese,Cantonese,Castilian,Catalan,Chinese,Croatian,Czech,Danish,Dutch,English,Estonian,Faroese,Finnish,Flemish,French,Galician,Georgian,German,Greek,Gujarati,Haitian,Haitian Creole,Hausa,Hawaiian,Hebrew,Hindi,Hungarian,Icelandic,Indonesian,Italian,Japanese,Javanese,Kannada,Kazakh,Khmer,Korean,Lao,Latin,Latvian,Letzeburgesch,Lingala,Lithuanian,Luxembourgish,Macedonian,Malagasy,Malay,Malayalam,Maltese,Mandarin,Maori,Marathi,Moldavian,Moldovan,Mongolian,Myanmar,Nepali,Norwegian,Nynorsk,Occitan,Panjabi,Pashto,Persian,Polish,Portuguese,Punjabi,Pushto,Romanian,Russian,Sanskrit,Serbian,Shona,Sindhi,Sinhala,Sinhalese,Slovak,Slovenian,Somali,Spanish,Sundanese,Swahili,Swedish,Tagalog,Tajik,Tamil,Tatar,Telugu,Thai,Tibetan,Turkish,Turkmen,Ukrainian,Urdu,Uzbek,Valencian,Vietnamese,Welsh,Yiddish,Yoruba}
language spoken in the audio, specify None to perform language detection (default: None)
--language_detection_threshold LANGUAGE_DETECTION_THRESHOLD, -ldt LANGUAGE_DETECTION_THRESHOLD
If the maximum probability of the language tokens is higher than this value, the language is detected. (default: 0.5)
--language_detection_segments LANGUAGE_DETECTION_SEGMENTS, -lds LANGUAGE_DETECTION_SEGMENTS
Number of segments/chunks to consider for the language detection. (default: 1)
--temperature TEMPERATURE
temperature to use for sampling (default: 0)
--best_of BEST_OF, -bo BEST_OF
number of candidates when sampling with non-zero temperature (default: 5)
--beam_size BEAM_SIZE, -bs BEAM_SIZE
number of beams in beam search, only applicable when temperature is zero (default: 5)
--patience PATIENCE, -p PATIENCE
optional patience value to use in beam decoding, as in https://arxiv.org/abs/2204.05424, the default (1.0) is equivalent to conventional
beam search (default: 2.0)
--length_penalty LENGTH_PENALTY
optional token length penalty coefficient (alpha) as in https://arxiv.org/abs/1609.08144, uses simple length normalization by default
(default: 1.0)
--repetition_penalty REPETITION_PENALTY
Penalty applied to the score of previously generated tokens (set > 1.0 to penalize). (default: 1.0)
--no_repeat_ngram_size NO_REPEAT_NGRAM_SIZE
Prevent repetitions of ngrams with this size (set 0 to disable). (default: 0)
--suppress_blank SUPPRESS_BLANK
Suppress blank outputs at the beginning of the sampling. (default: True)
--suppress_tokens SUPPRESS_TOKENS
comma-separated list of token ids to suppress during sampling; '-1' will suppress most special characters except common punctuations.
None - will pass empty list. (default: -1)
--initial_prompt INITIAL_PROMPT, -prompt INITIAL_PROMPT
optional text to provide context as a prompt for the first window. Use 'None' to disable it. Note: 'auto' and 'default' are experimental
~universal prompt presets, if task=transcribe then they work only if --language is set. (default: auto)
--prefix PREFIX Optional text to provide as a prefix for the first window. A partial transcription for the current audio input. (default: None)
--condition_on_previous_text CONDITION_ON_PREVIOUS_TEXT, -condition CONDITION_ON_PREVIOUS_TEXT
if True, provide the previous output of the model as a prompt for the next window; disabling may make the text inconsistent across
windows, but the model becomes less prone to getting stuck in a failure loop. If disabled then you may want to disable --reprompt too.
(default: True)
--prompt_reset_on_temperature PROMPT_RESET_ON_TEMPERATURE
Resets prompt if temperature is above this value. Arg has effect only if condition_on_previous_text is True. (default: 0.5)
--without_timestamps WITHOUT_TIMESTAMPS
Use <|notimestamps|> to sample text tokens only. For audios shorter than 30s. (default: None)
--max_initial_timestamp MAX_INITIAL_TIMESTAMP
The initial timestamp cannot be later than this. (default: 1.0)
--temperature_increment_on_fallback TEMPERATURE_INCREMENT_ON_FALLBACK, -fallback TEMPERATURE_INCREMENT_ON_FALLBACK
temperature to increase when falling back when the decoding fails to meet either of the thresholds below. To disable fallback set it to
'None'. (default: 0.2)
--compression_ratio_threshold COMPRESSION_RATIO_THRESHOLD
if the gzip compression ratio is higher than this value, treat the decoding as failed (default: 2.4)
--logprob_threshold LOGPROB_THRESHOLD
if the average log probability is lower than this value, treat the decoding as failed (default: -1.0)
--no_speech_threshold NO_SPEECH_THRESHOLD
if the probability of the <|nospeech|> token is higher than this value AND the decoding has failed due to 'logprob_threshold', consider
the segment as silence (default: 0.6)
--word_timestamps WORD_TIMESTAMPS, -wt WORD_TIMESTAMPS
Extract word-level timestamps and refine the results based on them (default: True)
--highlight_words HIGHLIGHT_WORDS, -hw HIGHLIGHT_WORDS
underline each word as it is spoken AKA karaoke in srt and vtt output formats. Note: it has no effect with --sentence (default: False)
--prepend_punctuations PREPEND_PUNCTUATIONS
if word_timestamps is True, merge these punctuation symbols with the next word (default: "'“¿([{-)
--append_punctuations APPEND_PUNCTUATIONS
if word_timestamps is True, merge these punctuation symbols with the previous word (default: "'.。,，!！?？:：”)]}、)
--threads THREADS Number of threads used for CPU inference; By default number of the real cores but no more that 4 (default: 0)
--version Show Faster-Whisper's version number
--vad_filter VAD_FILTER, -vad VAD_FILTER
Enable the voice activity detection (VAD) to filter out parts of the audio without speech. (default: True)
--vad_threshold VAD_THRESHOLD
Probabilities above this value are considered as speech. (default: 0.45)
--vad_min_speech_duration_ms VAD_MIN_SPEECH_DURATION_MS
Final speech chunks shorter min_speech_duration_ms are thrown out. (default: 250)
--vad_max_speech_duration_s VAD_MAX_SPEECH_DURATION_S
Maximum duration of speech chunks in seconds. Longer will be split at the timestamp of the last silence. (default: None)
--vad_min_silence_duration_ms VAD_MIN_SILENCE_DURATION_MS
In the end of each speech chunk time to wait before separating it. (default: 3000)
--vad_speech_pad_ms VAD_SPEECH_PAD_MS
Final speech chunks are padded by speech_pad_ms each side. (default: 900)
--vad_window_size_samples VAD_WINDOW_SIZE_SAMPLES
Size of audio chunks fed to the silero VAD model. Values other than 512, 1024, 1536 may affect model perfomance!!! (default: 1536)
--vad_method {silero_v4_fw,silero_v5_fw,silero_v3,silero_v4,silero_v5,pyannote_v3,pyannote_onnx_v3,auditok,webrtc}
Read there - '#231'. (default: silero_v4_fw)
--vad_dump Dumps VAD timings to a subtitle file for inspection. Additionaly dumps various intermediate audio files for debug. (default: False)
--vad_dump_aud Additional arg for --vad_dump, collected chunks as audio. (default: False)
--vad_device VAD_DEVICE
Device to use for pyannote_v3/pyannote_onnx_v3 VAD. Default is 'cuda' if CUDA device is detected, else is 'cpu'. If CUDA GPU is a second
device then set 'cuda:1'. (default: cuda)
--hallucination_silence_threshold HALLUCINATION_SILENCE_THRESHOLD, -hst HALLUCINATION_SILENCE_THRESHOLD
(Experimental) When word_timestamps is True, skip silent periods longer than this threshold (in seconds) when a possible hallucination is
detected. Optimal value is somewhere between 2 - 8 seconds. Inactive if None. (default: None)
--hallucination_silence_th_temp {0.0,0.2,0.5,0.8,1.0}, -hst_temp {0.0,0.2,0.5,0.8,1.0}
(Experimental) Additional heuristic for '--hallucination_silence_threshold'. If temperature is higher that this threshold then consider
segment as possible hallucination ignoring the hst score. Inactive if 1.0. (default: 1.0)
--clip_timestamps CLIP_TIMESTAMPS, -clip CLIP_TIMESTAMPS
Comma-separated list start,end,start,end,... timestamps (in seconds) of clips to process. The last end timestamp defaults to the end of
the file. VAD is auto-disabled. (default: 0)
--max_new_tokens MAX_NEW_TOKENS
Maximum number of new tokens to generate per-chunk. (default: None)
--chunk_length CHUNK_LENGTH
The length of audio segments. If it is not None, it will overwrite the default chunk_length of the FeatureExtractor. (default: None)
--hotwords HOTWORDS Pass hint words/phrases to the model. Similar as --reprompt except that it unsafely cuts into the new tokens space when reprompt safely
cuts into the context tokens space. Has no effect if --prefix is not None. (default: None)
--batched Enables batched inference mode - transcribes audio in batched VAD's chunks. Can increase speed ~2x-8x but will decrease the long form
transcription quality a bit. Many options don't work in this mode. (default: False)
--batch_size BATCH_SIZE
When using batched inference it sets the maximum number of parallel requests to model for decoding. (default: 8)
--multilingual MULTILINGUAL
Perform language detection on every segment. (default: False)
--model_preload MODEL_PRELOAD
True: faster-whisper model is preloaded before iteration on audio, but interferes with '--ff_vocal_extract' models. False: Runs inside
iteration after '--ff...' filters. None: automatic, it's default, False if 'ff_vocal_extract' or 'diarize_only' is enabled. (default:
None)
--batch_recursive, -br
Enables recursive batch files processing. Note: If set then it changes defaults of --output_dir. (default: False)
--beep_off Disables the beep sound when operation is finished. (default: False)
--skip Skips media file if subtitle exists. Works if input is wildcard or directory. Note: Checks for the first format in the list of
--output_format. (default: False)
--checkcuda, -cc Returns CUDA device count. (for Subtitle Edit's internal use)
--print_progress, -pp
Prints progress bar instead of transcription. (default: False)
--postfix Adds language as a postfix to subtitle's filename. (default: False)
--check_files Checks input files for errors before passing all them for transcription. Works if input is wildcard or directory. (default: False)
--alt_writer_off (For dev experiments) Forcefully disables alternative subs writer func. (default: False)
--ignore_dupe_prompt IGNORE_DUPE_PROMPT, -idp IGNORE_DUPE_PROMPT
PR163 aka don't add dupes to prompt. Reduces possibility of the hallucination loops. Sets how deep to look for dupes in the segments. 0 -
Disabled. (default: 2)
--hallucinations_list_off
Disables hallucinations_list, allows known hallucinations to be added to prompt. (default: False)
--v3_offsets_off Disables custom offsets to the defaults of pseudo-vad thresholds when 'large-v3' models are in use. Note: Offsets made to combat
'large-v3' hallucinations. (default: False)
--reprompt REPROMPT, --carry_initial_prompt REPROMPT
Ensures that initial_prompt is present in prompt for all windows/chunks. Has no effect if initial_prompt is None. (default: True)
--prompt_reset_on_no_end {0,1,2}
Resets prompt if there is no end of sentence in window/chunk. 0 - disabled, 1 - looks for period, 2 - looks for period or comma. Note:
it's auto-disabled if reprompt is False and if hotwords is None. (default: 2)
--rehot For Dev experiments. Just to transfer the auto prompt preset to the hotwords routines. (default: False)
--unmerged Experimental arg for batched inference. Runs unmerged segments instead of merged, very slow. Possible cons: further decrease of batched
transcription quality. Possible pros: more accurate timestamps, less hallucinations, not missing sentences, positive effect with
--multilingual. (default: False)
--japanese {blend,kanji,hiragana,katakana}, -ja {blend,kanji,hiragana,katakana}
This option may steer Japanese output to have a more preferred writing style. (default: blend)
--one_word {0,1,2} 0) Disabled. 1) Outputs srt and vtt subtitles with one word per line. 2) As '1', plus removes whitespace and ensures >= 50ms for sub
lines. Note: VAD may slightly reduce the accuracy of timestamps on some lines. (default: 0)
--sentence Enables splitting lines to sentences. Every sentence starts in the new segment. By default meant to output whole sentence per line for
better translations, but not limited to, read about '--max_...' parameters. Doesn't work well when --language is not specified. (default:
False)
--standard Quick hardcoded preset to split lines in standard way: --max_line_width=42, --max_line_count=2, --max_comma_cent=70 and --sentence are
activated automatically. (default: False)
--standard_asia Quick hardcoded preset to split lines in standard way for some Asian languages. --max_line_width=16, --max_line_count=2,
--max_comma_cent=80 and --sentence are activated automatically. (default: False)
--max_comma MAX_COMMA
(requires --sentence) After this line length a comma is treated as the end of sentence. Note: disabled if it's over or equal to
--max_line_width. (default: 250)
--max_comma_cent {20,30,40,50,60,70,80,90,100}
(requires --sentence) Percentage of --max_line_width when it starts breaking the line after comma. Note: 100 = disabled. (default: 100)
--max_gap MAX_GAP (requires --sentence) Threshold for a gap length in seconds, longer gaps are treated as dots. (default: 3.0)
--max_line_width MAX_LINE_WIDTH
The maximum number of characters in a line before breaking the line. Note: It works with --sentence too. (default: 1000)
--max_line_count MAX_LINE_COUNT
The maximum number of lines in one sub segment. Note: It works with --sentence too. (default: 1)
--min_dist_to_end {0,4,5,6,7,8,9,10,11,12}
(requires --sentence) If from words like 'the', 'Mr.' and ect. to the end of line distance is less than set then it starts in a new line.
Note: 0 = disabled. (default: 0)
--ff_dump Dumps pre-processed audio by the filters to the 16000Hz file and prevents deletion of some intermediate audio files. (default: False)
--ff_track {1,2,3,4,5,6}
Audio track selector. 1 - selects the first audio track. (default: 1)
--ff_fc Selects only front-center channel (FC) to process. (default: False)
--ff_lc Selects only a left channel to process. Useful if audio contains inverted polarity content, for example a narration. (default: False)
--ff_invert Inverts polarity a left channel and mix audio to mono, to keep inverted polarity content and delete same signal in the channels. Useful if
audio contains inverted polarity content. (default: False)
--ff_mp3 Audio filter: Conversion to MP3 and back. (default: False)
--ff_sync Audio filter: Stretch/squeeze samples to the given timestamps, with a maximum of 3600 samples per second compensation. Input file must be
container that support storing PTS like mp4, mkv... (default: False)
--ff_rnndn_sh Audio filter: Suppress non-speech with GregorR's SH model using Recurrent Neural Networks. Notes: It's more aggressive than Xiph, discards
singing. (default: False)
--ff_rnndn_xiph Audio filter: Suppress non-speech with Xiph's original model using Recurrent Neural Networks. (default: False)
--ff_fftdn [0 - 97] Audio filter: General denoise with Fast Fourier Transform. Notes: 12 - normal strength, 0 - disabled. (default: 0)
--ff_tempo [0.5 - 2.0]
Audio filter: Adjust audio tempo. Values below 1.0 slows down audio, above - speeds up. 1.0 = disabled. (default: 1.0)
--ff_gate Audio filter: Reduce lower parts of a signal. (default: False)
--ff_speechnorm Audio filter: Extreme and fast speech amplification. (default: False)
--ff_loudnorm Audio filter: EBU R128 loudness normalization. (default: False)
--ff_silence_suppress noise duration
Audio filter: Suppress quiet parts of audio. Takes two values. First value - noise tolerance in decibels [-70 - 0] (0=disabled), second
value - minimum silence duration in seconds [0.1 - 10]. (default: [0, 3.0])
--ff_lowhighpass Audio filter: Pass 50Hz - 7800 band. sinc + afir. (default: False)
--ff_vocal_extract {mdx1_kim2,mdx2_kim2,mb-roformer}, -ff_voc {mdx1_kim2,mdx2_kim2,mb-roformer}
Audio filter: Vocals extraction. 'mdx2_kim2' is different implementation and quality is a bit higher than 'mdx1_kim2'. The best SOTA model
is - 'mb-roformer'. (default: None)
--roformer_overlap [0.9 - 0.125]
Audio overlap for RoFormer. Lower value - bigger overlap. Bigger overlap - a bit higher quality, but a lot slower. (default: 0.75)
--roformer_vram {4,6,8,12,16}
VRAM size for RoFormer. Set to your GPU VRAM if you have problems with long audios. Useful if you have low RAM too. (default: None)
--mdx_chunk MDX_CHUNK
Chunk size in seconds for mdx_kim2/mdx1_kim2 filter. Notes: Smaller is using less VRAM, but can be slower and produce a bit lower quality.
(default: 15)
--voc_device VOC_DEVICE
Device to use for vocals extractin filter. Default is 'cuda' if CUDA device is detected, else is 'cpu'. If CUDA GPU is a second device
then set 'cuda:1'. (default: cuda)
--realign Enables timestamps realignment. Note: Works only on srt files. Auto-enables --sentence, auto-disables --one_word & --highlight_words.
(default: False)
--realign_device REALIGN_DEVICE
Device to use for realignment. Default is 'cuda' if CUDA device is detected, else is 'cpu'. If CUDA GPU is a second device then set
'cuda:1'. (default: cuda)
--diarize {pyannote_v3.0,pyannote_v3.1,reverb_v1,reverb_v2}
Enables diarization. Notes: It will activate --sentence and all output formats will be affected, exept json. Read more there -
'#322'. (default: None)
--diarize_device DIARIZE_DEVICE
Device to use for '--diarize'. Default is 'cuda' if CUDA device is detected, else is 'cpu'. If CUDA GPU is a second device then set
'cuda:1'. (default: cuda)
--diarize_threads DIARIZE_THREADS
Automatic by default. (default: 0)
--diarize_dump Dump diarization output to a file. Note: No need to set when'--diarize_only'. (default: False)
--speaker SPEAKER Diarization: To replace 'SPEAKER' string with your own word. (default: SPEAKER)
--num_speakers NUM_SPEAKERS
Diarization: Number of speakers, when known. (default: None)
--min_speakers MIN_SPEAKERS
Diarization: Minimum number of speakers. Has no effect when num_speakers is provided. (default: None)
--max_speakers MAX_SPEAKERS
Diarization: Maximum number of speakers. Has no effect when num_speakers is provided. (default: None)
--diarize_ff DIARIZE_FF
True: diarize after '--ff_...' audio filters, but runs inside iteration and interferes with FW model if '--model_preload' is True. False:
diarize all files before anything else. None: automatic, it's default, True if any 'ff' filter is enabled. (default: None)
--return_embeddings, -embeddings
Dump diarization embeddings to the files. Get a one embedding per speaker. (default: False)
--diarize_only Use if you want to do diarizations without transcription. Note: No need to set '--diarize_dump'. (default: False)