Yangang Cao

loss increase and appear nan

9

Hi, thanks for your excellent work. I extract feature from speech(pcm, 12GB) and noise(pcm, 9GB), and set count into 10000000. Then, I run run_train.py and get the following output: Can...

gsc存在高频缺失的情况

1

首先感谢贵团队的工作，经过测试，发现gsc存在高频缺失的现象，不知道有没有改进的方法

different phones.txt and words.txt

1

Hi dear author, I have trained my own model with this doc https://github.com/alphacep/vosk-api/tree/master/training, but it's totally different from vosk-model-small-en-us-0.15 and vosk-model-en-us-0.22-lgraph. 1.words.txt and phones.txt in vosk-model-en-us-0.22-lgraph are different from my...

fstdeterminizestar nnet3 model(not chain)

1

Hi, dear author when I ran ./local/lookahead/run_lookahead_no_ivector.sh in a nnet3 model to get vosk-like model, it cost all memory(more than 450GB) in server, and killed, it stuck in fstdeterminizestar. And...

Hi, dear maintainer, I don't understand the meaning of `score` in `ps_alignment_entry_t` in `ps_alignment_internel.h`, and you even commet it with "Alignment score (fairly meaningless)". Can I think of better alignment...

where to get the dictionary of vosk-model-en-us-0.22-lgraph

11

Hi, dear author, I want to get the dictionary of vosk-model-en-us-0.22-lgraph to check every phone in a word, where can I get? Thanks very much

grammar in Kaldi

2

Hi, dear author, setting grammar in Vosk is very useful, so I copy UpdateGrammarFst() to Kaldi and test it with open source chain model(http://kaldi-asr.org/models/m13), but the result is bad( I...

I try to run aecmos.py(there are total 800 files to test) in MacBook Air, Linux server in China and Linux server in America, they are all failed with "json_body =...

TypeError: decorator() missing 1 required positional argument: 'f' on M1 Mac

1

local install on M1 Mac, appear this error: ``` Traceback (most recent call last): File "/opt/homebrew/bin/melo", line 33, in sys.exit(load_entry_point('melotts==0.1.2', 'console_scripts', 'melo')()) TypeError: decorator() missing 1 required positional argument: 'f'...

my own data length limit

4

Hi, dear author, when I infer my own data, the length exceed 50 and error occurs. I find this line in gopt.py, `self.pos_embed = nn.Parameter(torch.zeros(1, 55, self.embed_dim))` So the max...