[KWS]Add kws example on HeySnips dataset. #1558
Merged
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
KPatr1ck
force-pushed
the
kws
branch
2 times, most recently
from
a35e99a to
fe7c4e5
Compare
zh794390558
reviewed
Apr 19, 2022
| @@ -0,0 +1,39 @@ | |||
| data: | |||
There was a problem hiding this comment.
配置文件参考其他egs，展开吧。不建议用层级的了。
| @@ -0,0 +1,5 @@ | |||
| #!/bin/bash | |||
|
|
|||
There was a problem hiding this comment.
这个脚本不推荐用户单独使用，正常都是从run.sh调用，我在run里加检查
| if self.feat_type in ['kaldi_fbank', 'kaldi_mfcc']: | ||
| waveform = paddle.to_tensor(waveform).unsqueeze(0) # (C, T) | ||
| record['feat'] = feat_func( | ||
| waveform=waveform, sr=self.sample_rate, **self.feat_config) |
There was a problem hiding this comment.
feature api的参数太多，且大部分为默认，我加个说明让用户根据api中的kwargs去配置。
|
|
||
| # yapf: disable | ||
| parser = argparse.ArgumentParser(__doc__) | ||
| parser.add_argument("--cfg_path", type=str, required=True) |
|
|
||
| mask = padding_mask(lengths) | ||
| num_utts = logits.shape[0] | ||
| num_keywords = logits.shape[2] |
There was a problem hiding this comment.
num_keywords 好奇怪。
B，T，D
| num_keywords = logits.shape[2] | ||
|
|
||
| loss = 0.0 | ||
| for i in range(num_utts): |
There was a problem hiding this comment.
不可以直接用CrossEntropy，需要在有效的音频长度内取max作为得分再算CE。
KPatr1ck
changed the title
|
This pull request is now in conflict :( |
在README中的模型处给出了，https://arxiv.org/pdf/2102.13552.pdf |
zh794390558
reviewed
Apr 25, 2022
| while i < len(score_list): | ||
| if score_list[i] >= threshold: | ||
| num_false_alarm += 1 | ||
| i += args.window_shift |
| for j in range(num_keywords): | ||
| # Add entropy loss CE = -(t * log(p) + (1 - t) * log(1 - p)) | ||
| if target[i] == j: | ||
| # For the keyword, do max-polling |
There was a problem hiding this comment.
看着是在T上做了max_pool后算Cross Entropy
| super(KWSModel, self).__init__() | ||
| self.backbone = backbone | ||
| self.linear = nn.Linear(self.backbone.hidden_dim, num_keywords) | ||
| self.activation = nn.Sigmoid() |
| self.kernel_size = kernel_size | ||
| self.dilation = dilation | ||
| self.causal = causal | ||
| self.receptive_fields = dilation * (kernel_size - 1) |
There was a problem hiding this comment.
kernel_size - 1 if no_dilation else dilation * (kernel_size -1)
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
PR types
New features
PR changes
Models
Describe
Add mdtc model for kws.