-
### Critical Point不一定是訓練network時的最大障礙
當Loss不再下降，我們認為卡在Critical Point，意味著Gradient很小。但Gradient真的很小嗎？
 with the objective of increasing accuracy. The problem is that when I use linear evaluation I get an…
-
Hi! I really like the your work and I tried that on my RTX 3060 12G, but it kept going NaN so I added lr scheduler and refactor the whole codebase. Here is the link https://github.com/author31/scm_nbd…
-
Repository cloned at 17.8.2022 22:05
The project seems to have all the building blocks in place. There is a GUI and means to train and evaluate the model. However, since there is no manual yet, I a…
-
I could be wrong in this, but looking at the code for [sgd](https://github.com/dmlc/MXNet.jl/blob/master/src/optimizers/sgd.jl) and [adam](https://github.com/dmlc/MXNet.jl/blob/master/src/optimizers/a…
-
Pytorch recently included pretrained object detection models such FastRCNN (
`torchvision.models.detection.faster_rcnn`).
Do you know how to integrate them with fastai?
-
Hey,
Basically, after I have fine-tuned a model and it runs the final performance check up, the validations metrics report something like 0.60mAP for my custom dataset, which seem reasonable.
Th…
-
Data augmentation:
- [ ] Random Resized Crop
- [x] Horizontal Flip
- [ ] Random Augment (see [code](https://github.com/tensorflow/tpu/blob/master/models/official/efficientnet/autoaugment.py#L672)…
-
### Describe the issue linked to the documentation
The [documentation for SGDClassifier](https://scikit-learn.org/stable/modules/sgd.html#sgd) says that with `learning_rate='optimal'` the initial lea…
-
Would it be helpful to add ability to manually change learning rate during training? I have naive implementation of this feature [here](https://github.com/r-zemblys/tensorflow-wavenet/tree/dev).
![lr2…