Only 10 submissions will be allowed per participant. This includes submissions for both Closed ASR and Open ASR tasks.
Please upload your submissions in text filetype(.txt) and follow the below format.
<utt_id_1> decoded_text_for_utt_1
<utt_id_2> decoded_text_for_utt_2
<utt_id_3> decoded_text_for_utt_3
Along with every submission, details of the underlying model and the tool used should be provided. For example: Kaldi HMM-GMM based system, ESPnet Transformer based E2E system, etc. If language model is used, it must be mentioned.
In case the submission is for the Open ASR Task, details of the external data/models used should be disclosed. For example: 10K sentences crawled from web was used to build a language model, Librispeech acoustic model parameters were used to initialize the network, etc.