pytorch
Here are 17,735 public repositories matching this topic...
-
Updated
Sep 14, 2021 - Python
-
Updated
Sep 20, 2021 - Jupyter Notebook
-
Updated
Sep 16, 2021 - Python
-
Updated
Sep 16, 2021 - Jupyter Notebook
I'm training various model heads (HTC, cascade mask-rcnn, etc.) with the CBNetV2 backbone (implementation here) on a custom coco-format dataset with only bboxes. I'm using the following training and testing pipelines:
albu_train_transforms = [
dict(
type='ShiftScaleRotate',
shift_limit=0.0625,
scale_limit=0.0,
-
Updated
Sep 21, 2021 - JavaScript
-
Updated
Sep 20, 2021 - Python
-
Updated
Sep 20, 2021 - Python
🚀 Feature
lr_find need unique temporary checkpoint filenames.
Motivation
I'm running a number of experiment in parallel that are saving to the same folder. Thus, they have the same trainer.default_root_dir
. However, since they all have the same directory and filename, they are overwriting each other.
Pitch
lr_find temporary checkpoint should have unique filenames.
-
Updated
Sep 10, 2021 - Jupyter Notebook
-
Updated
Sep 21, 2021 - Python
Change tensor.data
to tensor.detach()
due to
pytorch/pytorch#6990 (comment)
tensor.detach()
is more robust than tensor.data
.
-
Updated
Sep 21, 2021 - Python
-
Updated
Sep 20, 2021 - C
-
Updated
Sep 15, 2021 - Python
-
Updated
Sep 21, 2021 - Python
-
Updated
May 2, 2021
-
Updated
Aug 30, 2021 - Jupyter Notebook
-
Updated
Sep 17, 2021 - Python
Bug Report
Is the issue related to model conversion?
If the ONNX checker reports issues with this model then this is most probably related to the converter used to convert the original framework model to ONNX. Please create this bug in the appropriate converter's GitHub repo (pytorch, tensorflow-onnx, sklearn-onnx, keras-onnx, onnxmltools) to get the best help.
Describe the bug
T
-
Updated
Sep 17, 2021 - Python
-
Updated
Sep 21, 2021 - Python
Is your feature request related to a problem? Please describe.
I typically used compressed datasets (e.g. gzipped) to save disk space. This works fine with AllenNLP during training because I can write my dataset reader to load the compressed data. However, the predict
command opens the file and reads lines for the Predictor
. This fails when it tries to load data from my compressed files.
-
Updated
Sep 18, 2021 - Python
-
Updated
Sep 21, 2021 - Python
-
Updated
Sep 21, 2021 - Python
-
Updated
Jul 25, 2021 - Jupyter Notebook
-
Updated
Mar 14, 2021 - Jupyter Notebook
-
Updated
Sep 20, 2021 - Python
Improve this page
Add a description, image, and links to the pytorch topic page so that developers can more easily learn about it.
Add this topic to your repo
To associate your repository with the pytorch topic, visit your repo's landing page and select "manage topics."
https://github.com/huggingface/transformers/blob/546dc24e0883e5e9f5eb06ec8060e3e6ccc5f6d7/src/transformers/models/gpt2/modeling_gpt2.py#L698
Assertions can't be relied upon for control flow because they can be disabled, as per the following: