Override inference parameters (model architecture, num bands, num classes) from checkpoint #298
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…handling for checkpoints from different gdl versions fixes NRCan#183
…handling for checkpoints from different gdl versions fixes NRCan#183
…to 183-inference-model-override # Conflicts: # inference_segmentation.py
# Conflicts: # config/model/smp_unet.yaml # inference_segmentation.py # models/model_choice.py # train_segmentation.py
remtav
changed the title
CharlesAuthier
requested changes
Mar 24, 2022
GDL.py
Outdated
| @@ -32,6 +36,11 @@ def run_gdl(cfg: DictConfig) -> None: | |||
| """ | |||
| cfg = OmegaConf.create(cfg) | |||
|
|
|||
| # Couldn't find a better way to do this. Ideally a return_hydra_config could be set in @hydra.main decorator... | |||
There was a problem hiding this comment.
they are already in the yaml under general.config_name and general.config_path
ymoisan
reviewed
Mar 24, 2022
|
|
||
| if model_ckpt != params.model or classes_ckpt != classes or modalities_ckpt != modalities: | ||
| logging.warning(f"\nParameters from checkpoint will override inputted parameters." | ||
| f"\n\t\t\t Inputted | Overriden" |
There was a problem hiding this comment.
If a parameter read from the hydra config is overridden on the commandline, that will still be overridden by the checkpoint right ?
CharlesAuthier
approved these changes
Mar 28, 2022
mpelchat04
approved these changes
Mar 29, 2022
remtav
added a commit
to remtav/geo-deep-learning
that referenced
this pull request
Jul 5, 2022
…sses) from checkpoint (NRCan#298) * fixes NRCan#293 NRCan#246 add tests for optimizer instantiation in test_optimizers.py adapt our unet models (models/unet.py) to expect same parameter names as smp models * minor typo fixes * implement overriding model params from checkpoint with minimal error handling for checkpoints from different gdl versions fixes NRCan#183 * name model yamls as close as possible to upcoming naming convention * fix model name * implement overriding model params from checkpoint with minimal error handling for checkpoints from different gdl versions fixes NRCan#183 * small bugfix for pointing to parameters inside checkpoint * model_choice.py: add update checkpoint utility * fixes NRCan#293 NRCan#246 add tests for optimizer instantiation in test_optimizers.py adapt our unet models (models/unet.py) to expect same parameter names as smp models * minor typo fixes * name model yamls as close as possible to upcoming naming convention * small bugfix for pointing to parameters inside checkpoint * remove deeplabv3 dualhead warning and add link for deeplabv3_dualhead.py * fixes NRCan#293 NRCan#246 add tests for optimizer instantiation in test_optimizers.py adapt our unet models (models/unet.py) to expect same parameter names as smp models * name model yamls as close as possible to upcoming naming convention * minor typo fixes * update to PR 295 * GDL.py: restore to previous commit based on cauthier's comment
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
fixes #183
This feature will make it possible once again to launch an inference from 2 simple arguments:
path/url to image
path/url to model checkpoint