Add part: lstm block #66
Open
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
christophmluscher
requested review from
albertz,
NeoLegends,
JackTemaki,
Gerstenberger,
michelwi and
Atticus1806
NeoLegends
approved these changes
Dec 18, 2024
i6_models/parts/lstm.py
Outdated
| enforce_sorted: bool | ||
|
|
||
| @classmethod | ||
| def from_dict(cls, model_cfg_dict: Dict): |
There was a problem hiding this comment.
Same Q as in the other PR: why is this necessary now, and hasn't been for the other assemblies?
albertz
reviewed
Dec 18, 2024
Co-authored-by: Albert Zeyer <[email protected]>
albertz
reviewed
Dec 19, 2024
albertz
reviewed
Dec 19, 2024
albertz
reviewed
Dec 19, 2024
albertz
reviewed
Dec 19, 2024
Comment on lines
+53
to
+54
| if seq_len.get_device() >= 0: | ||
| seq_len = seq_len.cpu() |
There was a problem hiding this comment.
Suggested change
| if seq_len.get_device() >= 0: | |
| seq_len = seq_len.cpu() | |
| seq_len = seq_len.cpu() |
albertz
reviewed
Dec 19, 2024
| ) | ||
|
|
||
| def forward(self, x: torch.Tensor, seq_len: torch.Tensor) -> Tuple[torch.Tensor, torch.Tensor]: | ||
| if not torch.jit.is_scripting() and not torch.jit.is_tracing(): |
There was a problem hiding this comment.
Why only when not scripting? Don't you want that seq_len is always on CPU?
There was a problem hiding this comment.
I followed the example in the blstm part.
if not torch.jit.is_scripting() and not torch.jit.is_tracing():
# during graph mode we have to assume all Tensors are on the correct device,
# otherwise move lengths to the CPU if they are on GPU
if seq_len.get_device() >= 0:
seq_len = seq_len.cpu()
I did not copy the comment over... I did not yet get to look why this is necessary
@JackTemaki you implemented the BLSTM IIRC. You remember why this was done in this way?
Co-authored-by: Albert Zeyer <[email protected]>
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.
Adds LSTM Block