icefall

icefall copied to clipboard

Using of double question mark in error reports

1

Could you please approve that "??" is equivalent to unk in Kaldi.

AlexandderGorodetski

Training using narrow band data (8KHz)

1

Hello guys, Could you please let me know how to work with narrow band data (8KHz). What configuration files should be updated? Thanks, AlexG.

AlexandderGorodetski

max_duration

15

Hello guys, I am working with K2-Tedlium3 experiment. In the following command what is the meaning of max-duration? I assumed that it is maximal number of frames in training utterances...

AlexandderGorodetski

MMI Training with Librispeech recipe doesn't converge with or without alignment

6

Hi, Thanks always for your research. I've been trying to train conformer_mmi with librispeech recipe at the current commit. (e334e570d838cbe15188201f6bd47c009b9292be) The only thing I changed in conformer_mmi/train.py is codes for...

dan-legit

tedlium prepare.sh file should be updated

3

Hello There is following error during running prepare.sh of tedlium recipe ValueError: It seems that 'sph2pipe' binary is not installed; did you run 'lhotse install-sph2pipe'? The command 'lhotse install-sph2pipe' can...

AlexandderGorodetski

[WIP] Fix MMI recipe

3

This PR try to fix https://github.com/k2-fsa/icefall/issues/685 and to tune a better result for MMI models. - [x] Make current model converge. - [ ] Tune a better result. - [...

pkufool

Training command: ```bash export CUDA_VISIBLE_DEVICES="0,1,2,3,4,5,6,7" ./pruned_transducer_stateless7/train.py \ --world-size 8 \ --num-epochs 20 \ --use-fp16 1 \ --max-duration 250 \ --exp-dir pruned_transducer_stateless7/exp \ --feedforward-dims "512,512,1024,1024,512" \ --master-port 12535 \ --training-subset L...

csukuangfj

Improve diagnostics for when models diverge

1

Guys, Occasionally, as I tune my latest version of the new setup, I get divergence. It looks like this: ``` 2022-11-20 01:34:30,723 INFO [train.py:916] Epoch 20, batch 850, loss[loss=0.1723, simple_loss=0.2561,...

danpovey

Phone based lang vs BPE

1

Hi guys, I need to use phones based lang because I've lots of words that I need to provide the exact pronunciations for them. I'm wondering what would be the...

userasrk2

Tensorboard data from remote computer

3

Hello guys, I've just completed TedLium3 training. But I did it on remote AWS machine. How can I see tensorboard data? Should I download this file events.out.tfevents.1668622331.bdcf5ab79459.10825.0 to my computer...

AlexandderGorodetski