Create linkcheck.yml

.github/workflows/linkcheck.yml

@@ -0,0 +1,22 @@
+name: Check Markdown links
+
+on:
+  push:
+    branches:
+      - main
+  pull_request:
+    branches:
+      - main
+
+  # Allows you to run this workflow manually from the Actions tab
+  workflow_dispatch:
+
+jobs:
+  markdown-link-check:
+    runs-on: ubuntu-latest
+    steps:
+    - uses: actions/checkout@v2
+    - uses: gaurav-nelson/github-action-markdown-link-check@v1
+      with:
+        use-quiet-mode: 'yes'
+        config-file: '.github/workflows/mlc_config.json'

.github/workflows/mlc-config.json

@@ -0,0 +1,13 @@
+{
+    "ignorePatterns": [
+        {
+            "pattern": ".*localhost.*"
+        },
+        {
+            "pattern": ".*127\\.0\\.0\\.1.*"
+        },
+        {
+            "pattern": ".*0\\.0\\.0\\.0.*"
+        }
+    ]
+}

README.md

@@ -164,7 +164,7 @@ sparseml.yolov5.train \
   --hyp hyps/hyp.finetune.yaml --cfg yolov5s.yaml --patience 0
 ```
 
-- Check out the [YOLOv5 CLI example](ultralytics-yolov5/tutorials/sparse-transfer-learning.md) for more details on the YOLOv5 training pipeline
+- Check out the [YOLOv5 CLI example](integrations/ultralytics-yolov5/tutorials/sparse-transfer-learning.md) for more details on the YOLOv5 training pipeline
 - Check out the [Hugging Face CLI example](integrations/huggingface-transformers/tutorials/sparse-transfer-learning-bert.md) for more details on the available NLP training pipelines
 - Check out the [Torchvision CLI example](integrations/torchvision/tutorials/sparse-transfer-learning.md) for more details on the image classification training pipelines
 

integrations/huggingface-transformers/tutorials/sparse-transfer-learning-bert-python.md

@@ -302,13 +302,13 @@ The command creates a `./deployment` folder in your local directory, which conta
 
 Take a look at the tutorials for more examples in other use cases:
 
-- [Sparse Transfer with GLUE Datasets (SST2) for sentiment analysis](tutorials/sentiment-analysis/docs-sentiment-analysis-python-sst2.ipynb)
-- [Sparse Transfer with Custom Datasets (RottenTomatoes) and Custom Teacher from HF Hub for sentiment analysis](tutorials/sentiment-analysis/docs-sentiment-analysis-python-custom-teacher-rottentomatoes)
-- [Sparse Transfer with GLUE Datasets (QQP) for multi-input text classification](tutorials/text-classification/docs-text-classification-python-qqp.ipynb)
-- [Sparse Transfer with Custom Datasets (SICK) for multi-input text classification](tutorials/text-classification/docs-text-classification-python-sick.ipynb)
-- [Sparse Transfer with Custom Datasets (TweetEval) and Custom Teacher for single input text classificaiton](tutorials/text-classification/docs-text-classification-python-custom-teacher-tweeteval.ipynb)
-- [Sparse Transfer with Custom Datasets (GoEmotions) for multi-label text classification](tutorials/text-classification/docs-text-classification-python-multi-label-go_emotions.ipynb)
-- [Sparse Transfer with Conll2003 for named entity recognition](tutorials/token-classification/docs-token-classification-python-conll2003.ipynb)
-- [Sparse Transfer with Custom Datasets (WNUT) and Custom Teacher for named entity recognition](tutorials/token-classification/docs-token-classification-custom-teacher-wnut.ipynb)
+- [Sparse Transfer with GLUE Datasets (SST2) for sentiment analysis](sentiment-analysis/docs-sentiment-analysis-python-sst2.ipynb)
+- [Sparse Transfer with Custom Datasets (RottenTomatoes) and Custom Teacher from HF Hub for sentiment analysis](sentiment-analysis/docs-sentiment-analysis-python-custom-teacher-rottentomatoes.ipynb)
+- [Sparse Transfer with GLUE Datasets (QQP) for multi-input text classification](text-classification/docs-text-classification-python-qqp.ipynb)
+- [Sparse Transfer with Custom Datasets (SICK) for multi-input text classification](text-classification/docs-text-classification-python-sick.ipynb)
+- [Sparse Transfer with Custom Datasets (TweetEval) and Custom Teacher for single input text classification](text-classification/docs-text-classification-python-custom-teacher-tweeteval.ipynb)
+- [Sparse Transfer with Custom Datasets (GoEmotions) for multi-label text classification](text-classification/docs-text-classification-python-multi-label-go_emotions.ipynb)
+- [Sparse Transfer with Conll2003 for named entity recognition](token-classification/docs-token-classification-python-conll2003.ipynb)
+- [Sparse Transfer with Custom Datasets (WNUT) and Custom Teacher for named entity recognition](token-classification/docs-token-classification-python-custom-teacher-wnut.ipynb)
 - Sparse Transfer with SQuAD (example coming soon!)
 - Sparse Transfer with Squadshifts Amazon (example coming soon!)

integrations/huggingface-transformers/tutorials/token-classification/token-classification-cli.md

@@ -175,7 +175,7 @@ A `deployment` folder is created in your local directory, which has all of the f
 
 ## Sparse Transfer Learning with a Custom Dataset (WNUT_17)
 
-Beyond the Conll2003 dataset, we can also use a dataset from the Hugging Face Hub or from local files. Let's try an example of each for the sentiment analysis using [WNUT 17](wnut_17), which is also a NER task.
+Beyond the Conll2003 dataset, we can also use a dataset from the Hugging Face Hub or from local files. Let's try an example of each for the sentiment analysis using WNUT_17, which is also a NER task.
 
 For simplicity, we will perform the fine-tuning without distillation. Although the transfer learning recipe contains distillation
 modifiers, by setting `--distill_teacher disable` we instruct SparseML to skip distillation.

integrations/old-examples/dbolya-yolact/README.md

@@ -106,7 +106,7 @@ The following table lays out the root-level files and folders along with a descr
 | Folder/File Name                                | Description                                                                                                           |
 |-------------------------------------------------|-----------------------------------------------------------------------------------------------------------------------|
 | [recipes](./recipes)                            | Typical recipes for sparsifying YOLACT models along with any downloaded recipes from the SparseZoo.                   |
-| [yolact](./yolact)                              | Integration repository folder used to train and sparsify YOLACT models (`setup_integration.sh` must run first).       |
+| yolact                                          | Integration repository folder used to train and sparsify YOLACT models (`setup_integration.sh` must run first).       |
 | [README.md](./README.md)                        | Readme file.                                                                                                          |
 | [tutorials](./tutorials)                        | Easy to follow sparsification tutorials for YOLACT  models.                                                            |
 

integrations/old-examples/keras/README.md

@@ -31,7 +31,7 @@ The techniques include, but are not limited to:
 
 ## Tutorials
 
-- [Classification](https://github.com/neuralmagic/sparseml/blob/main/integrations/keras/notebooks/classification.ipynb)
+- [Classification](https://github.com/neuralmagic/sparseml/blob/main/integrations/old-examples/keras/notebooks/classification.ipynb)
 
 ## Installation
 

integrations/old-examples/pytorch/tutorials/classification_sparse_transfer_learning_tutorial.md

@@ -74,7 +74,7 @@ Note: The models above were originally trained and sparsified on the [ImageNet](
 
 - After noting respective SparseZoo model stub, [train.py](../train.py) script can be used to download checkpoint and [Imagenette](https://github.com/fastai/imagenette) and kick-start transfer learning. 
    The transfer learning process itself is guided using recipes; We include example [recipes](../recipes) for classification along with others in the SparseML [GitHub repository](https://github.com/neuralmagic/sparseml). 
-[Learn more about recipes and modifiers](../../../docs/source/recipes.md).
+[Learn more about recipes and modifiers](https://github.com/neuralmagic/sparseml/tree/main/docs/source/recipes.md).
 
 - Run the following example command to kick off transfer learning for [ResNet-50](https://arxiv.org/abs/1512.03385) starting from a moderately pruned checkpoint from [SparseZoo](https://sparsezoo.neuralmagic.com/):
 ```

integrations/old-examples/pytorch/tutorials/sparsifying_pytorch_models_using_recipes.md

@@ -49,7 +49,7 @@ pip install "sparseml[torchvision, dev]"
 Recipes are YAML or Markdown files that SparseML uses to easily define and control the sparsification of a model.
 Recipes consist of a series of `Modifiers` that can influence the training process in different ways. A list of
 common modifiers and their uses is provided
-[here](../../../docs/source/recipes.md#modifiers-intro).
+[here](https://github.com/neuralmagic/sparseml/tree/main/docs/source/recipes.md#modifiers-intro).
 
 SparseML provides a recipe for sparsifying a ResNet-50 model trained on the tiny Imagenette dataset. The recipe can
 be viewed in the browser
@@ -71,7 +71,7 @@ of the parameters list for a single `GMPruningModifier`.
 
 Recipes can integrated into training flows with a couple of lines of code by using a `ScheduledModifierManager`
 that wraps the PyTorch `Optimizer` step.  An example of how this is done can be found
-[here](../../../docs/source/code.md#pytorch-sparsification).
+[here](https://github.com/neuralmagic/sparseml/tree/main/docs/source/code.md#pytorch-sparsification).
 
 For this example, we can use the `sparseml.image_classification.train` utility.  This utility runs a
 PyTorch training flow that is modified by a `ScheduledModifierManager` and takes a recipe as an input.

integrations/old-examples/rwightman-timm/README.md

@@ -81,7 +81,7 @@ python train.py \
 ```  
 
 Documentation on the original script can be found
-[here](https://rwightman.github.io/pytorch-image-models/scripts/).
+[here](https://huggingface.co/docs/timm/training_script).
 The latest commit hash that `train.py` is based on is included in the docstring.
 
 
@@ -112,5 +112,5 @@ python export.py \
     --config ./path/to/checkpoint/args.yaml
 ```
 
-The DeepSparse Engine [accepts ONNX formats](https://docs.neuralmagic.com/sparseml/source/onnx_export.html) and is engineered to significantly speed up inference on CPUs for the sparsified models from this integration.
+The DeepSparse Engine [accepts ONNX formats](https://docs.neuralmagic.com/archive/sparseml/source/onnx_export.html) and is engineered to significantly speed up inference on CPUs for the sparsified models from this integration.
 Examples for loading, benchmarking, and deploying can be found in the [DeepSparse repository here](https://github.com/neuralmagic/deepsparse).

integrations/old-examples/tensorflow_v1/README.md

@@ -31,7 +31,7 @@ The techniques include, but are not limited to:
 
 ## Tutorials
 
-- [Classification](https://github.com/neuralmagic/sparseml/blob/main/integrations/tensorflow_v1/notebooks/classification.ipynb)
+- [Classification](https://github.com/neuralmagic/sparseml/blob/main/integrations/old-examples/tensorflow_v1/notebooks/classification.ipynb)
 
 ## Installation
 

integrations/old-examples/ultralytics-yolov3/README.md

@@ -35,8 +35,8 @@ The techniques include, but are not limited to:
 
 ## Tutorials
 
-- [Sparsifying YOLOv3 Using Recipes](https://github.com/neuralmagic/sparseml/blob/main/integrations/ultralytics-yolov3/tutorials/sparsifying_yolov3_using_recipes.md)
-- [Sparse Transfer Learning With YOLOv3](https://github.com/neuralmagic/sparseml/blob/main/integrations/ultralytics-yolov3/tutorials/yolov3_sparse_transfer_learning.md)
+- [Sparsifying YOLOv3 Using Recipes](https://github.com/neuralmagic/sparseml/blob/main/integrations/old-examples/ultralytics-yolov3/tutorials/sparsifying_yolov3_using_recipes.md)
+- [Sparse Transfer Learning With YOLOv3](https://github.com/neuralmagic/sparseml/blob/main/integrations/old-examples/ultralytics-yolov3/tutorials/yolov3_sparse_transfer_learning.md)
 
 ## Installation
 

integrations/old-examples/ultralytics-yolov3/recipes/yolov3-spp.pruned.md

@@ -149,7 +149,7 @@ pruning_modifiers:
 
 This recipe creates a sparse, [YOLOv3-SPP](https://arxiv.org/abs/1804.02767) model that achieves 97% recovery of its baseline accuracy on the COCO detection dataset.
 Training was done using 4 GPUs at half precision using a total training batch size of 256 with the
-[SparseML integration with ultralytics/yolov3](https://github.com/neuralmagic/sparseml/tree/main/integrations/ultralytics-yolov3).
+[SparseML integration with ultralytics/yolov3](https://github.com/neuralmagic/sparseml/tree/main/integrations/old-examples/ultralytics-yolov3).
 
 When running, adjust hyperparameters based on training environment and dataset.
 
@@ -159,7 +159,7 @@ When running, adjust hyperparameters based on training environment and dataset.
 
 ## Training
 
-To set up the training environment, follow the instructions on the [integration README](https://github.com/neuralmagic/sparseml/blob/main/integrations/ultralytics-yolov3/README.md).
+To set up the training environment, follow the instructions on the [integration README](https://github.com/neuralmagic/sparseml/blob/main/integrations/old-examples/ultralytics-yolov3/README.md).
 Using the given training script from the `yolov3` directory the following command can be used to launch this recipe.  
 The contents of the `hyp.pruned.yaml` hyperparameters file is given below.
 Adjust the script command for your GPU device setup. 

integrations/old-examples/ultralytics-yolov3/recipes/yolov3-spp.pruned.short.md

@@ -148,7 +148,7 @@ pruning_modifiers:
 
 This recipe creates a sparse [YOLOv3-SPP](https://arxiv.org/abs/1804.02767) model in a shortened schedule as compared to the original pruned recipe.
 It will train faster, but will recover slightly worse.
-Use the following [SparseML integration with ultralytics/yolov3](https://github.com/neuralmagic/sparseml/tree/main/integrations/ultralytics-yolov3) to run.
+Use the following [SparseML integration with ultralytics/yolov3](https://github.com/neuralmagic/sparseml/tree/main/integrations/old-examples/ultralytics-yolov3) to run.
 
 When running, adjust hyperparameters based on training environment and dataset.
 
@@ -158,7 +158,7 @@ When running, adjust hyperparameters based on training environment and dataset.
 
 ## Training
 
-To set up the training environment, follow the instructions on the [integration README](https://github.com/neuralmagic/sparseml/blob/main/integrations/ultralytics-yolov3/README.md).
+To set up the training environment, follow the instructions on the [integration README](https://github.com/neuralmagic/sparseml/blob/main/integrations/old-examples/ultralytics-yolov3/README.md).
 Using the given training script from the `yolov3` directory the following command can be used to launch this recipe.  
 The contents of the `hyp.pruned.yaml` hyperparameters file is given below.
 Adjust the script command for your GPU device setup. 

integrations/old-examples/ultralytics-yolov3/recipes/yolov3-spp.pruned_quantized.md

@@ -223,7 +223,7 @@ quantization_modifiers:
 
 This recipe creates a sparse-quantized, [YOLOv3-SPP](https://arxiv.org/abs/1804.02767) model that achieves 94% recovery of its baseline accuracy on the COCO detection dataset.
 Training was done using 4 GPUs at half precision using a total training batch size of 256 with the
-[SparseML integration with ultralytics/yolov3](https://github.com/neuralmagic/sparseml/tree/main/integrations/ultralytics-yolov3).
+[SparseML integration with ultralytics/yolov3](https://github.com/neuralmagic/sparseml/tree/main/integrations/old-examples/ultralytics-yolov3).
 
 When running, adjust hyperparameters based on training environment and dataset.
 
@@ -237,7 +237,7 @@ This additionally means that the checkpoints are saved using state_dicts rather
 
 ## Training
 
-To set up the training environment, follow the instructions on the [integration README](https://github.com/neuralmagic/sparseml/blob/main/integrations/ultralytics-yolov3/README.md).
+To set up the training environment, follow the instructions on the [integration README](https://github.com/neuralmagic/sparseml/blob/main/integrations/old-examples/ultralytics-yolov3/README.md).
 Using the given training script from the `yolov3` directory the following command can be used to launch this recipe.  
 The contents of the `hyp.pruned_quantized.yaml` hyperparameters file is given below.
 Adjust the script command for your GPU device setup. 

integrations/old-examples/ultralytics-yolov3/recipes/yolov3-spp.pruned_quantized.short.md

@@ -223,7 +223,7 @@ quantization_modifiers:
 
 This recipe creates a sparse-quantized [YOLOv3-SPP](https://arxiv.org/abs/1804.02767) model in a shortened shcedule as compared to the original pruned recipe.
 It will train faster, but will recover slightly worse.
-Use the following [SparseML integration with ultralytics/yolov3](https://github.com/neuralmagic/sparseml/tree/main/integrations/ultralytics-yolov3) to run.
+Use the following [SparseML integration with ultralytics/yolov3](https://github.com/neuralmagic/sparseml/tree/main/integrations/old-examples/ultralytics-yolov3) to run.
 
 When running, adjust hyperparameters based on training environment and dataset.
 
@@ -237,7 +237,7 @@ This additionally means that the checkpoints are saved using state_dicts rather
 
 ## Training
 
-To set up the training environment, follow the instructions on the [integration README](https://github.com/neuralmagic/sparseml/blob/main/integrations/ultralytics-yolov3/README.md).
+To set up the training environment, follow the instructions on the [integration README](https://github.com/neuralmagic/sparseml/blob/main/integrations/old-examples/ultralytics-yolov3/README.md).
 Using the given training script from the `yolov3` directory the following command can be used to launch this recipe.  
 The contents of the `hyp.pruned_quantized.yaml` hyperparameters file is given below.
 Adjust the script command for your GPU device setup. 

integrations/old-examples/ultralytics-yolov3/recipes/yolov3-spp.test.md

@@ -224,7 +224,7 @@ quantization_modifiers:
 This is a test recipe useful for quickly evaluating the time and resources needed for pruning and quantizing a model.
 In addition, it offers a quick integration tests pathway.
 This recipe creates a sparse-quantized [YOLOv3-SPP](https://arxiv.org/abs/1804.02767) model that will not be accurate.
-Use the following [SparseML integration with ultralytics/yolov3](https://github.com/neuralmagic/sparseml/tree/main/integrations/ultralytics-yolov3) to run.
+Use the following [SparseML integration with ultralytics/yolov3](https://github.com/neuralmagic/sparseml/tree/main/integrations/old-examples/ultralytics-yolov3) to run.
 
 Note that half-precision, EMA, and pickling are not supported for quantization.
 Therefore, once quantization is run, all three will be disabled for the training pipeline.
@@ -236,7 +236,7 @@ This additionally means that the checkpoints are saved using state_dicts rather
 
 ## Training
 
-To set up the training environment, follow the instructions on the [integration README](https://github.com/neuralmagic/sparseml/blob/main/integrations/ultralytics-yolov3/README.md).
+To set up the training environment, follow the instructions on the [integration README](https://github.com/neuralmagic/sparseml/blob/main/integrations/old-examples/ultralytics-yolov3/README.md).
 Using the given training script from the `yolov3` directory the following command can be used to launch this recipe.  
 The contents of the `hyp.pruned_quantized.yaml` hyperparameters file is given below.
 Adjust the script command for your GPU device setup. 

integrations/old-examples/ultralytics-yolov3/recipes/yolov3-spp.transfer_learn_pruned.md

@@ -130,7 +130,7 @@ pruning_modifiers:
 This recipe transfer learns from a sparse, [YOLOv3-SPP](https://arxiv.org/abs/1804.02767) model.
 It was originally tested on the VOC dataset and achieved 0.84 mAP@0.5.
 
-Training was done using 4 GPUs at half precision with the [SparseML integration with ultralytics/yolov3](https://github.com/neuralmagic/sparseml/tree/main/integrations/ultralytics-yolov3).
+Training was done using 4 GPUs at half precision with the [SparseML integration with ultralytics/yolov3](https://github.com/neuralmagic/sparseml/tree/main/integrations/old-examples/ultralytics-yolov3).
 
 When running, adjust hyperparameters based on training environment and dataset.
 
@@ -142,7 +142,7 @@ The training results for this recipe are made available through Weights and Bias
 
 ## Training
 
-To set up the training environment, follow the instructions on the [integration README](https://github.com/neuralmagic/sparseml/blob/main/integrations/ultralytics-yolov3/README.md).
+To set up the training environment, follow the instructions on the [integration README](https://github.com/neuralmagic/sparseml/blob/main/integrations/old-examples/ultralytics-yolov3/README.md).
 Using the given training script from the `yolov3` directory the following command can be used to launch this recipe.  
 Adjust the script command for your GPU device setup. 
 Ultralytics supports both DataParallel and DDP.