-
Hi,
Thank you for the amazing work on this repo. I have identified what might be an issue with the implementation of Lightly's Gaussian Blur that is negatively impacting the performance of most of th…
-
Hi,
I noticed a difference between the code ([main_vicreg.py#L202](https://github.com/facebookresearch/vicreg/blob/main/main_vicreg.py#L202)) and the Algorithm 1 in the paper where the subtraction …
-
Hi Théo,
I was impressed by your sslsv project.
In the README.md, self supervised version is well introduced. (Maybe, " python train.py configs/vicreg_b256.yml ")
If i want to use the supervised…
-
This seems to be a good practice. But it is not realized in the optimizer class.
-
Hi,
When running the commands specified on the readme file to load the pretrained models on PyTorch Hub, they fail for resnet50x2 and resnet200x2. The issue is that the callable methods that load s…
-
**Describe the bug**
In addition to `solo.utils.misc.remove_bias_and_norm_from_weight_decay`, we might consider also `solo.utils.misc.ignore_dim1_parameters` to ignore weight decay *and `lars_adapt…
-
Define a parent class `Twins` for siamese networks,
subclass it for `BarlowTwins` and `VICReg` overriding loss functions
-
Hi,
The torch.cov in the covariance_loss function requires torch >= 1.10.
Replacing:
> _, D = emb.size()
> cov_emb = torch.cov(emb.T)
> cov_ref_emb = torch.cov(ref_emb.T)
to:
>…
-
Hi. Thanks for the great work!
I'm trying to reproduce the results of Table 12 (impact of expander dimensionality) in your paper.
Could you teach me what hyperparameters you used in the experiments…
-
"For VOC07 Everingham et al. (2010), we train a linear SVM with LIBLINEAR Fan et al. (2008). The images are center cropped and resized to 224 × 224, and the C values are computed with cross-validation…