00001101

Is it possible to extract frame-wise features with the parameter "options"?

4

`smile = opensmile.Smile(feature_set=opensmile.FeatureSet.eGeMAPSv01b, feature_level=opensmile.FeatureLevel.Functionals)` The above script gives me features with shape `1x88`, but I want frame-wise features with shape `Nx88`, so I changed the above statement to: `smile =...

Problem with seq2seq model for both Chinese and English

2

Hi, I'm training an ASR model for both Chinese and English based on `seq2seq` following this [librispeech](https://github.com/facebookresearch/wav2letter/tree/master/recipes/models/seq2seq_tds) recipe. My training `target` is `ltr` which looks like this: ``` | a...

I'm not using `ivector` for nnet3 model, referring to the help: ``` --online : You can set this to false to disable online iVector estimation and have all the data...

Which version of scrapy to use for this project?

2

I keep getting `import errors`, so I wonder if anyone could tell us which version of scrapy to use. So far, the `import errors` including: - `from scrapy import log`...

I found some issues about your implementation, including: 1. [this line](https://github.com/vzxxbacq/PLDA/blob/cc4c0e0d1557cc964900ae2f46d5a89fd54ec1f4/plda.py#L403): is it a bug ? 2. [this line](https://github.com/vzxxbacq/PLDA/blob/cc4c0e0d1557cc964900ae2f46d5a89fd54ec1f4/plda.py#L290): the original kaldi implementation, `mix_var` has an inverse operation which is...

why define a new graph when running test ?

1

To my knowledge, the test process may be simplified as follows: restore the model and get tensors we need, including the input utterances and the output embedding matrix, something like:...