Update CI to use pnpm v7 (#235)

* Bug Fix

* Update CI to use pnpm v7

- Fixed pnpm version to 7 so that ci is not breaking

* Fix Again

Author: theory-in-progress

.github/workflows/ci.yml

@@ -68,7 +68,7 @@ jobs:
           pip install wheel setuptools pip -Uqq
           pip install -r ./scripts/requirements.txt -f https://download.pytorch.org/whl/cpu/torch_stable.html -q
           pip uninstall -y tqdm
-          npm i -g pnpm
+          npm install pnpm@7 -g
           pnpm i --frozen-lockfile --color
 
       - name: Download datasets
@@ -104,7 +104,7 @@ jobs:
             pnpm-and-pip-cache-
 
       - run: pip install -Uq pip wheel && bash scripts/run_code_style.sh install
-      - run: npm i -g pnpm
+      - run: npm install pnpm@7 -g
       - run: pnpm i --frozen-lockfile --color
       - run: pnpm lint
       - run: python scripts/check_copies.py

CONTRIBUTING.md

@@ -78,11 +78,11 @@ To add a new template,
 
 5. Once everything looks good, put the template name with respective filenames (in array format) in the `templates.json` in `src/templates` directory. This is required to tell Code-Generator to fetch the files from this template.
 
-6. In the `README.md`, copy the training launch code from `src/template-common/README.md`.
+6. In the `README.md`, copy the training launch code from `src/templates/template-common/README.md`.
 
-7. For the `requirements.txt`, copy from the `src/template-common/requirements.txt`.
+7. For the `requirements.txt`, copy from the `src/templates/template-common/requirements.txt`.
 
-8. For the `utils.py`, copy the starter code from the `src/template-common/utils.py`.
+8. For the `utils.py`, copy the starter code from the `src/templates/template-common/utils.py`.
 
 9. You can check if the copied codes needed are up-to-date with the base codes with: `python scripts/check_copies.py`
 

src/templates/template-vision-segmentation/data.py

@@ -85,9 +85,9 @@ def setup_data(config: Namespace):
             download=False,
         )
     except RuntimeError as e:
-        raise e(
+        raise RuntimeError(
             "Dataset not found. You can use `download_datasets` from data.py function to download it."
-        )
+        ) from e
 
     dataset_eval = VOCSegmentationPIL(
         root=config.data_path, year="2012", image_set="val", download=False

src/templates/template-vision-segmentation/requirements.txt

@@ -2,7 +2,7 @@ torch>=1.10.2
 torchvision>=0.11.3
 pytorch-ignite>=0.4.8
 pyyaml
-albumentations
+albumentations>=1.3.0
 image_dataset_viz
 
 #::: if (['neptune', 'polyaxon'].includes(it.logger)) { :::#