Search results

Why is numpy 2.0 not supported?

Status: Open.

#119 In kensho-technologies/pyctcdecode;

· Zirunis opened on Jan 21, 2026

OOV words with small LM

Status: Open.

#115 In kensho-technologies/pyctcdecode;

· davidavdav opened on Nov 6, 2023

Support for Language Models other than KenLM?

Status: Open.

#110 In kensho-technologies/pyctcdecode;

· 21jun opened on Mar 16, 2023

Hotwords runtime is slow

Status: Open.

#109 In kensho-technologies/pyctcdecode;

· nirraviv opened on Feb 19, 2023

How lm_score and logit_Score is calculated in decorder.py?

Status: Open.

#108 In kensho-technologies/pyctcdecode;

· profsatwinder opened on Feb 6, 2023

pyctcdecode not working with Nemo finetuned model

Status: Open.

#93 In kensho-technologies/pyctcdecode;

· manjuke opened on Jan 6, 2023

Return the integer token ids along with text in decode_beams()

Status: Open.

#92 In kensho-technologies/pyctcdecode;

· titu1994 opened on Nov 16, 2022

How does the code handle the blank token when decoding logits from a pre-trained acoustic model?

Status: Open.

#90 In kensho-technologies/pyctcdecode;

· geliAI opened on Sep 30, 2022

Add support for LM from transformers

Status: Open.

#86 In kensho-technologies/pyctcdecode;

· Theodotus1243 opened on Sep 7, 2022

Unigrams not provided warning

Status: Open.

#85 In kensho-technologies/pyctcdecode;

· to-schi opened on Aug 31, 2022

BPE models - chunking of audio potentially "orphans" subwords of a word across chunks

Status: Open.

#84 In kensho-technologies/pyctcdecode;

· leecming82 opened on Aug 30, 2022

Recovering rescored logits

Status: Open.

#83 In kensho-technologies/pyctcdecode;

· OllieBroadhurst opened on Aug 23, 2022