returnn
returnn copied to clipboard
rwth-i6
The RWTH extensible training framework for universal recurrent neural networks
I think some tests fail (at least locally for me). Current version is TF 2.16, but whenever someone gets to the issue here, maybe check for the latest stable TF...
I just wanted to track this here: There seem to be some WER degradation in some setup by @Marvin84 occuring in TensorFlow 2.14 and not in earlier versions (although this...
Makes a common pattern much easier. @albertz Any comments on the added docs?
Jintao (@jiangj-dc) has a config where this is in the rec layer: ``` "end": {"class": "compare", "from": ":i", "kind": "equal", "value": 3}, ``` That fails with: ``` EXCEPTION Traceback (most...
This PR removes some some usage of `num_outputs` in the dataset implementations. This PR is a collection of works and best reviewed and merged on a commit-by-commit basis. I'm going...
Bumps [tensorflow](https://github.com/tensorflow/tensorflow) from 2.11.1 to 2.12.1. Release notes Sourced from tensorflow's releases. TensorFlow 2.12.1 Release 2.12.1 Bug Fixes and Other Changes The use of the ambe config to build and...
![dependabot[bot] avatar](https://avatars.githubusercontent.com/in/29110?v=4 "dependabot[bot] avatar"){kind=avatar}
Hi, I'm having a problem with `rf.RelPosCausalSelfAttention` when using it in a transformer decoder. It fails because it wants to remove `single_step_dim` from a tensor that does not have it...
Just like we do in TensorFlow, print some info on the sequences in the batch: Seq lens, size, etc. (Basically just copy the logic.)
I just saw this in the CI (at commit d5b954b8f6e4c84ec2c289733590e1bf4154ba8b): ``` ============================= test session starts ============================== platform linux -- Python 3.10.[14](https://github.com/rwth-i6/returnn/actions/runs/9909690500/job/27378323845#step:7:15), pytest-8.2.2, pluggy-1.5.0 rootdir: /home/runner/work/returnn/returnn configfile: pytest.ini collected 2 items...
There will only be one `saved_tensors_hooks` active, specifically for the most recent `gradient_checkpoint_scope`. So any of the earlier pack hooks will not be used, when there are multiple simultaneous `gradient_checkpoint_scope`s....