Intron (https://intron.io) is a health-tech startup building ASR for clinicians with African accents, helping to reduce burnout from tedious documentation for a... See more

Hi dear Zindians, If somebody who participated could be kind to share his or her baseline notebook for learning?Sincerely,Akmal

I've only worked on the competition for it's last 6-7 days only, so I'm pretty happy to have landed on the 7th place :)I used whisper's tiny, base, small and me... See more

Hi, Receive our Congratulations Top10 and everyone participated! Could you please share your winning solution notebooks with learners?

Lets Hope everybody gets the result they want !

Hi Everyone, hi @Krishna_Priya, @Mohammed_NASRI, @seif_aber, @mithilsalunkhe, @Eloff, Could you please explain how you solved this problem? Does it have anythin... See more

I encounter this error everytime I attempt training the whisper small model on the train data. The training runs till about 0.18 epochsRuntimeError: The size of... See more

# 1 The architectureFirst, the challenge is in english, as a result, you can use existing english models. Throwing in an idea from Modular Deep Learning. You ca... See more

Please can you clarify what is an acceptable use of metadata for this competition. The host has already implied that it could be used to remove problematic utte... See more

No one is able to submit solutions in the past 2 days. Even the submission files that were accepted before are not being accepted now. This is resulting in a lo... See more

Has anyone been able to unzip the files in the afrispeech-train dataset provided on this platform? I am trying to unzip and load the files onto Google drive but... See more

malformed reference csv at line 6Not able to make any submission, even the previously submitted solutions are giving the same error.@zindi can you please look i... See more

Can you explain whatthe WER score means? Does a score of 0.97 mean 97/100 or is it really 0.97/100? I would really appreciate your help with this.

Please someone should help me I can't run this cell in the starter notebook.from transformers import Seq2SeqTrainingArgumentstraining_args = Seq2SeqTrainingArgu... See more

Following the successful NVIDIA x Intron NeMo Masterclass last week, this Google Colab tutorial was made available by the awesome research team at Intron to hel... See more

I was doing fine with huggingface-hub-0.13.4/transformers-4.28.0.dev0But now I have an incompatibility problem between huggingface-hub-0.14.1/transformers-4.29.... See more

Take your next big step in Automatic Speech Recognition for African accents! Learn about NVIDIA's cutting-edge technology for African ASR at the upcoming NVIDIA... See more

Hello guys, I tried to download the entire dataset locally using load_dataset("tobiolatunji/afrispeech-200", "all", streaming=False) The downloading and extract... See more

NVIDIA is the global leader in AI computing. NVIDIA NeMo is a conversational AI toolkit built for researchers working on automatic speech recognition (ASR), tex... See more

Hello everyone, does anyone knows how I can extract only a subset of the train data ( not all the amount of data)? It is about 401G.O and it seems that Google C... See more

Does anyone know how I may use the sample submission file that was provided to me to generate predictions? When I load it up to retrieve the audio IDs and make... See more

I'm trying to make a submmit with an .csv or a .xls generated with pandas- when I submitt the submit.csv, it merges the headers cells "ID" and "transcript" toge... See more

Hi, I'm trying to submit my submission but I get the above error. I've tried index=False, concatenating the test_meta and dev_meta to form the submission file a... See more

When I loaded and worked with WhisperFeatureExtractor and WhisperTokenizer as feature extractor and tokenizer, I saved 3 seconds on my CPU TIME (4 workers) for... See more

I think we finally found the fix for this error. RuntimeError: Caught RuntimeError in replica 0 on device 0. Original Traceback (most recent call last): Fil... See more