-
Hello, I read from your paper that you reproduced the TVMM code. Can you share the code in this regard? Thank you very much.
-
https://github.com/lightly-ai/lightly/blob/66ad1b40ebf3b53512703c774988211ce283211f/lightly/loss/vicreg_loss.py#L128-L129
I think the VICReg loss removes the mean, then calls `.var()` which also de…
-
Hydra futtatható környezethez szükséges a Dataseteket, a DataLoadereket és a LightningModulokat src fájlokként hozzáadni
#8 alapján.
Módosítandó fájlok:
- [birds_datamodule.py](https://github.com…
-
前辈 你好 想问一下这个模型是否能用于图像分类呢,我也是有多个分类数据集用于不同分类器
-
#11 -es második fele.
Train futtatása classifier-rel Dockerben, tesztelve a megírt `config/` fájlokat.
Eval futtatása Dockerben.
GPU-n való tesztelés
**Kérdések:**
Lehessen külön futattni a V…
-
- [x] ONNX export of the model from pytorch: now works for `attention` (also `gnnlsh`, but less important)
- [x] check compatibility of the results between ONNX and pytorch inference outputs
- [x] r…
jpata updated
3 months ago
-
#1351 added a `VICRegProjectionHead`, let's use it also in the vicreg example code:
* examples/pytorch/vicreg.py
* examples/pytorch_lightning/vicreg.py
* examples/pytorch_lightning_distributed/vicr…
-
Hi. Thanks U for this awesome repo. I have followed the code found (`https://github.com/lightly-ai/lightly/blob/master/examples/pytorch/vicreg.py`) and change the backbone to ResNet152 and the input_s…
-
Hi,
Congratulations for such a wonderful work in SSL domain.
My question is concerned with the experiment configurations. It seems that fixed patch sizes were used in comparison and ablation exp…
-
Using the current base model (cnn-encoder + slot-attention), we have several options for decoder
1. Recon head with l2 loss (slot-attention paper)
2. Recon head with pixel BCE/CE loss,
3. Recon h…