-
Hi,
Just FYI, in the official MelGan repo, the authors used Hinge losses. However, in the paper, the author described with L2 loss. This repo is consistent with the paper! I am setting up some experi…
-
Is this repository the official implementation?
-
The `downsample` part in each block/layer (not the skip connection part), the PyTorch do it in conv3x3 using stride=2, but official caffe version in conv1x1 with stride=2
```
conv1x1 -> caffe do …
-
I wonder why the Retrieval accuracy is almost 20% higher than the official JAX/FLAX implementation.
As the paper says, "While we achieve consistent results reported in (Tay et al. 2020) for most task…
-
**Is your feature request related to a problem? Please describe.**
Currently there is no implementation in qc-opendrive. But this rule is a basic one. Without being compliant to the schema, many ot…
-
We need to extend schema validation with custom keywords. We may describe them with additional vocabularies, but we also need to make an implementation for them to be validatable. How may it be implem…
-
https://github.com/google/automl/blob/master/efficientdet/efficientdet_arch.py
-
May I ask what is your team's HOTA score for this work? And how is it calculated? I noticed that the organizer stated, "In this challenge track, teams will be ranked on the leaderboard based on their …
-
### Feature Description
We can add an argument (for example, `--context-shift`, `--no-context-shift`) to enable/disable context shift.
If disabled:
- Requests bigger than context window will re…
-
Dear authors,
Could you share your implementation about the attacks you used to generate the dataset **[SafeDecoding-Attackers](https://huggingface.co/datasets/flydust/SafeDecoding-Attackers?row=1…