loginethu/whisper-a100 🖼️❓🔢✓📝 → ❓

▶️ 54 runs 📅 May 2024 ⚙️ Cog 0.9.8

About

openai whisper model on A100 hardware

Example Output

Output

{"segments":[{"id":0,"end":6.8,"seek":0,"text":" Die litte verhaal die hulle vertel is vals.","start":0,"tokens":[50364,3229,287,9786,1306,1641,304,978,32335,68,6509,338,307,371,1124,13,50704],"avg_logprob":-0.9236625600083966,"temperature":0,"no_speech_prob":0.12543471157550812,"compression_ratio":1.5411764705882354},{"id":1,"end":11.24,"seek":0,"text":" Die deur was gebied, gesloten en gebolt ook.","start":6.8,"tokens":[50704,3229,368,374,390,21125,1091,11,5019,75,21990,465,1519,1763,2282,7839,13,50926],"avg_logprob":-0.9236625600083966,"temperature":0,"no_speech_prob":0.12543471157550812,"compression_ratio":1.5411764705882354},{"id":2,"end":15.16,"seek":0,"text":" Ripe paars is pas voor een koninklijke tafel.","start":11.24,"tokens":[50926,497,6527,2502,685,307,1736,7358,3881,5897,475,75,33611,1846,29874,13,51122],"avg_logprob":-0.9236625600083966,"temperature":0,"no_speech_prob":0.12543471157550812,"compression_ratio":1.5411764705882354},{"id":3,"end":19.56,"seek":0,"text":" Een dikke witte steen was op die ronde kaarpot.","start":15.16,"tokens":[51122,25374,1026,47764,261,9786,2126,268,390,999,978,367,7259,6799,6529,310,13,51342],"avg_logprob":-0.9236625600083966,"temperature":0,"no_speech_prob":0.12543471157550812,"compression_ratio":1.5411764705882354},{"id":4,"end":24,"seek":0,"text":" Die kite diep en gewees, maar bleef alweer.","start":19.56,"tokens":[51342,3229,350,642,274,414,79,465,1519,826,279,11,10314,5408,5666,419,826,260,13,51564],"avg_logprob":-0.9236625600083966,"temperature":0,"no_speech_prob":0.12543471157550812,"compression_ratio":1.5411764705882354},{"id":5,"end":28.84,"seek":0,"text":" Die pleasende uur vlieg al te snel.","start":24,"tokens":[51564,3229,3362,296,5445,344,374,371,6302,70,419,535,42582,13,51806],"avg_logprob":-0.9236625600083966,"temperature":0,"no_speech_prob":0.12543471157550812,"compression_ratio":1.5411764705882354},{"id":6,"end":33.04,"seek":2884,"text":" Die kam was vermoeid met een milde wob.","start":28.84,"tokens":[50364,3229,9727,390,1306,3280,68,327,1131,3881,15154,68,261,996,13,50574],"avg_logprob":-0.8795321328299386,"temperature":0,"no_speech_prob":0.023886293172836304,"compression_ratio":1.548148148148148},{"id":7,"end":37.2,"seek":2884,"text":" Die kam was vermoeid met een wilde mob.","start":33.04,"tokens":[50574,3229,9727,390,1306,3280,68,327,1131,3881,4868,68,4298,13,50782],"avg_logprob":-0.8795321328299386,"temperature":0,"no_speech_prob":0.023886293172836304,"compression_ratio":1.548148148148148},{"id":8,"end":41.84,"seek":2884,"text":" Dit sterke arm zal jou honder scheldig.","start":37.2,"tokens":[50782,25270,18924,330,3726,29599,11110,2157,1068,956,5957,328,13,51014],"avg_logprob":-0.8795321328299386,"temperature":0,"no_speech_prob":0.023886293172836304,"compression_ratio":1.548148148148148},{"id":9,"end":46.36,"seek":2884,"text":" Toen hy haar een wit orkiet gegeven, bluste die.","start":41.84,"tokens":[51014,1407,268,2477,39371,3881,32161,420,74,1684,1519,432,553,11,888,301,975,978,13,51240],"avg_logprob":-0.8795321328299386,"temperature":0,"no_speech_prob":0.023886293172836304,"compression_ratio":1.548148148148148},{"id":10,"end":48.56,"seek":2884,"text":" Die bietel droned in die warme june zon.","start":46.36,"tokens":[51240,3229,272,1684,338,1224,266,292,294,978,1516,1398,8156,68,710,266,13,51350],"avg_logprob":-0.8795321328299386,"temperature":0,"no_speech_prob":0.023886293172836304,"compression_ratio":1.548148148148148}],"translation":null,"transcription":" Die litte verhaal die hulle vertel is vals. Die deur was gebied, gesloten en gebolt ook. Ripe paars is pas voor een koninklijke tafel. Een dikke witte steen was op die ronde kaarpot. Die kite diep en gewees, maar bleef alweer. Die pleasende uur vlieg al te snel. Die kam was vermoeid met een milde wob. Die kam was vermoeid met een wilde mob. Dit sterke arm zal jou honder scheldig. Toen hy haar een wit orkiet gegeven, bluste die. Die bietel droned in die warme june zon.","detected_language":"afrikaans"}

Performance Metrics

9.37s Prediction Time

9.39s Total Time

All Input Parameters

{
  "audio": "https://replicate.delivery/mgxm/e5159b1b-508a-4be4-b892-e1eb47850bdc/OSR_uk_000_0050_8k.wav",
  "model": "large-v2",
  "language": "af",
  "translate": false,
  "temperature": 0,
  "transcription": "plain text",
  "suppress_tokens": "-1",
  "logprob_threshold": -1,
  "no_speech_threshold": 0.6,
  "condition_on_previous_text": true,
  "compression_ratio_threshold": 2.4,
  "temperature_increment_on_fallback": 0.2
}

Input Parameters

audio (required) Type: string: Audio file
model Default: large-v2: Choose a Whisper model.
language: language spoken in the audio, specify None to perform language detection
patience Type: number: 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
translate Type: booleanDefault: false: Translate the text to English when set to True
temperature Type: numberDefault: 0: temperature to use for sampling
transcription Default: plain text: Choose the format for the transcription
initial_prompt Type: string: optional text to provide as a prompt for the first window.
suppress_tokens Type: stringDefault: -1: comma-separated list of token ids to suppress during sampling; '-1' will suppress most special characters except common punctuations
logprob_threshold Type: numberDefault: -1: if the average log probability is lower than this value, treat the decoding as failed
no_speech_threshold Type: numberDefault: 0.6: 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
condition_on_previous_text Type: booleanDefault: true: 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
compression_ratio_threshold Type: numberDefault: 2.4: if the gzip compression ratio is higher than this value, treat the decoding as failed
temperature_increment_on_fallback Type: numberDefault: 0.2: temperature to increase when falling back when the decoding fails to meet either of the thresholds below

Output Schema

Example Execution Logs

Transcribe with large-v2 model
  0%|          | 0/5241 [00:00<?, ?frames/s]
 55%|█████▌    | 2884/5241 [00:05<00:04, 574.96frames/s]
100%|██████████| 5241/5241 [00:08<00:00, 588.55frames/s]
100%|██████████| 5241/5241 [00:08<00:00, 586.24frames/s]

Version Details

Version ID: a1083f40a9e6721ff24710e7b96f85f4876de6e6574de142ba641ac0333b6dee
Version Created: May 22, 2024

Run on Replicate →