ecm200 caltech_birds issues

ecm200 / caltech_birds

A set of notebooks as a guide to the process of fine-grained image classification of birds species, using PyTorch based deep neural networks.

41 stars 7 forks source link

issues

Newest

Newest Most commented Recently updated Oldest Least commented Least recently updated

EARLY_STOPPING_PATIENCE config keyword sitting in wrong place in config.py

#36 ecm200 opened 3 years ago
1
setup.py dependencies need updating

#35 ecm200 opened 3 years ago
0
get_parser() overrides config file to a default file if none is given on execution

#34 ecm200 closed 3 years ago
1
Make command line argument configuration overrides work by using ARGS module in CUB_TOOLS package

#33 ecm200 opened 3 years ago
0
Testing of end points in notebook

#32 ecm200 closed 3 years ago
0
Fixing the refactor of Torchscript model generation

#31 ecm200 closed 3 years ago
0
Refactor Torchscript conversion

#30 ecm200 closed 3 years ago
1
Move Torchscript conversion function into the IgniteTrainer class and reference from the ClearMLIgniteTrainer class to keep ClearML and general useful utilities separate.

#29 ecm200 closed 3 years ago
3
Modify setup.py so that it has entry points for CLI application added.

#28 ecm200 opened 3 years ago
1
Merging clearml_dev branch into the master

#27 ecm200 closed 3 years ago
0
Models dev incorporate into master branch

#26 ecm200 closed 3 years ago
0
Write config YAMLs for all models trained so far.

#25 ecm200 closed 3 years ago
1
Dev

#24 ecm200 opened 3 years ago
0
Clearml integrations

#23 ecm200 closed 3 years ago
0
Investigate deploying training of models via ClearML platform for model and data versioning.

#22 ecm200 closed 3 years ago
1
Add support for all Torchvision models

#21 ecm200 closed 3 years ago
1
Ignite events, logging and callbacks controlled via configuration file

#20 ecm200 opened 3 years ago
0
Add Trainer class for Ignite models.

#19 ecm200 closed 3 years ago
5
Integrate torchmetrics package for all metrics used.

#18 ecm200 opened 3 years ago
1
Take Torchvision model selection code out of Trainer init method and write separate private method which is called from init method.

#17 ecm200 closed 3 years ago
2
Use config file to control model parameters so that single training scripts can be used, with new Trainer class.

#16 ecm200 closed 3 years ago
1
Refactor main script to use Trainer class which takes care of all model related objects and commence training.

#15 ecm200 closed 3 years ago
1
Investigate Lightning training scripts and train a few models.

#14 ecm200 opened 3 years ago
0
Create Trainer class for Ignite framework.

#13 ecm200 closed 3 years ago
2
Investigate Ignite training scripts and successfully train a few models.

#12 ecm200 closed 3 years ago
1
Add inference methods to Trainer class.

#11 ecm200 opened 3 years ago
1
Add support for more loss criterion to Trainer class.

#10 ecm200 opened 3 years ago
0
Add support for more learning rate schedulers to Trainer class.

#9 ecm200 closed 3 years ago
1
Add support for more optimiser functions in Trainer class.

#8 ecm200 opened 3 years ago
0
Add pytorchcv support to Trainer class to allow those models to be trained.

#7 ecm200 closed 3 years ago
3
Memory Usage

#6 DJT777 opened 3 years ago
2
Dependencies

#5 DJT777 closed 3 years ago
1
Is it fine if i use pyTorch = 1.6 because torch-lucent require pyTorch >= 1.5

#4 superpieflavour opened 4 years ago
0
ModuleNotFoundError: No module named 'pytorchcv.utils'

#3 AlphaWu opened 4 years ago
1
"pip install" doesn't work on cub_tools

#2 ecm200 opened 4 years ago
0
No dependencies listed for cub_tools package installer or for example notebooks

#1 ecm200 closed 4 years ago
1