Fangjun Kuang

I tried to use the pre-trained model with colab to decode test-clean. It took about 7 hours with a batch size of 1. See the log below. (Batch size 1...

> Any chance that you could also try to share some statistics with CPU only ? Here is the log for decoding with CPU using `test-clean`. ~The run for `test-other`...

The code for reproducing is available at https://github.com/csukuangfj/k2_decoding_benchmark/blob/master/librispeech/sp-main.py (With the commit fb12cee562da1f802e4c05ebfb27d4589ff88b64) ```bash python3 ./sp-main.py ./sp-params.yaml ```

> Is it with a single core? Does speechbrain use multiple cores if they are available? The machine I was using is a virtual machine with 64 cores.

> Just to understand, why you still call the graph HLG ? and not TLG ?, Please see the comment from @danpovey in https://github.com/k2-fsa/snowfall/issues/121#issuecomment-793314492 > I'd rather call it HLG...

> If you see, token 3 (which is space) is labeled ' ' instead of '' or in sentence piece '_' You can modify the script so that if a...

> So the HLG compile, but in the decoding phase, I have an error when using the HLG.fst > It seems that the intersect between H_L_G is wrong somewhere... >...

> is there a comparaison between you ctc_topo and the one proposed by Miao in EESEN? No, not yet. The one proposed by Miao in EESEN is very similar to...

> @30stomercury could you take a look at that? I don't understand why the results vary with batch size here ... I think several reasons can cause this problem: (1)...

The site is also not synchronized with this repository.