Nikolay Karpov
Nikolay Karpov
@titu1994 I covered half of [the table](https://docs.google.com/spreadsheets/d/11feRNo9Bx0jnS947PkeGd0j8wPOgw4CIVfRnzpqZe48/edit?usp=sharing) and PR already huge. let's review it first. Then will continue with AggregateTokenizer
There is a work in progress in the PR https://github.com/NVIDIA/NeMo/pull/8428
You can try decoding without N-gram at the branch _karpnv/beamsearch_ with parameters ``` python3 ./scripts/asr_language_modeling/ngram_lm/eval_beamsearch_ngram_ctc.py \ model_path=./am_model.nemo \ dataset_manifest=./manifest.json \ preds_output_folder=/tmp \ ctc_decoding.strategy=flashlight \ ctc_decoding.beam.nemo_kenlm_path="" \ ctc_decoding.beam.beam_size=[4] \ ctc_decoding.beam.beam_beta=[0.5] ```