-
Did you used Res2Net pretrained on ImageNet as a backbone ? But the input size of the Res2Net which pretrained on ImageNet is 224x224. The input size in your code( which is also named train_size ) is…
-
**Is your feature request related to a problem? Please describe.**
The multimaterial / multicolor painting tool can be somewhat imprecise, and the paint bucket tool works well on things with hard lin…
-
Was reading this thread in dark mode when Awful got kicked out of memory. State restoration botched the dark theme? Automatic dark mode causing something weird (e.g. preserve state in dark mode, next …
-
Does ai-toolkit support setting the number of repetitions of images folder like koaya? The number of repetitions of multiple folders is different
-
Hi,
Thank you for sharing your code. I trained and tested the model using the CAMO dataset, following the exact settings mentioned in your paper (AdamW optimizer, initial learning rate of 0.001, 35…
-
> size mismatch for image_encoder.blocks.23.mlp.lin1.weight: copying a param with shape torch.Size([4096, 1024]) from checkpoint, the shape in current model is torch.Size([5120, 1280]).
size mismatc…
-
Hi, just playing with this and I'm a bit confused.
Toggling field display depending on the value of another field is working perfectly but the fields which are being conditionally displayed are not…
-
For now I am doing inference on images following the code in the inference.ipynb file. However, I have realised that it uses image paths to be able to make the inference.
However, due to limitatio…
-
Tabs. If they're in an AppBar, it all works great. If they're not, then the background is white and the default color is white and they disappear. Seems like we should somehow be making the AppBar set…
Hixie updated
8 months ago
-
i trained my model on tesla v100 for 58,000 steps & my Average loss is 23. & My MAP is 43.42%. & even my final.weight & best.weight is not saved after training is completed.
* what command do you …