[Cherry-Pick] Doc fixes (#1583)

* YOLOv5 doc fixes (#1574)

* IC doc fixes (#1577)

* Add explixit open-cv dep (#1575)

* Transformers doc fixes (#1578)

* Fix yolov8 export alias (#1579)

integrations/huggingface-transformers/README.md

@@ -29,11 +29,9 @@ Once trained, SparseML enables you to export models to the ONNX format, such tha
 Install with `pip`:
 
 ```bash
-pip install sparseml[torch]
+pip install sparseml[transformers]
 ```
 
-**Note**: Transformers will not immediately install with this command. Instead, a sparsification-compatible version of Transformers will install on the first invocation of the Transformers code in SparseML.
-
 ## **Tutorials**
 
 - [Sparse Transfer Learning with the Python API](tutorials/sparse-transfer-learning-bert-python.md) [**RECOMMENDED**]
@@ -50,13 +48,13 @@ pip install sparseml[torch]
 ### **Use Case Examples - Python**
 
 - [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 Custom Datasets (RottenTomatoes) and Custom Teacher from HF Hub for sentiment analysis](tutorials/sentiment-analysis/docs-sentiment-analysis-python-custom-teacher-rottentomatoes.ipynb)
 - [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 Custom Datasets (WNUT) and Custom Teacher for named entity recognition](tutorials/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!)
 
@@ -66,7 +64,7 @@ pip install sparseml[torch]
 
 SparseZoo is an open-source repository of pre-sparsified models, including BERT-base, BERT-large, RoBERTa-base, RoBERTa-large, and DistillBERT. With SparseML, you can fine-tune these pre-sparsified checkpoints onto custom datasets (while maintaining sparsity) via sparse transfer learning. This makes training inference-optimized sparse models almost identical to your typical training workflows!
 
-[Check out the available models](https://sparsezoo.neuralmagic.com/?repo=huggingface&page=1)
+[Check out the available models](https://sparsezoo.neuralmagic.com/?repos=huggingface)
 
 ### **Recipes**
 
@@ -140,15 +138,15 @@ Currently supported tasks include:
 
 Sparse Transfer is very similiar to the typical transfer learing process used to train NLP models, where we fine-tune a checkpoint pretrained on a large upstream dataset using masked language modeling onto a smaller downstream dataset. With Sparse Transfer Learning, however, we simply start the fine-tuning process from a pre-sparsified checkpoint and maintain sparsity while the training process occurs.
 
-Here, we will fine-tune a [90% pruned version of BERT](https://sparsezoo.neuralmagic.com/models/nlp%2Fmasked_language_modeling%2Fobert-base%2Fpytorch%2Fhuggingface%2Fwikipedia_bookcorpus%2Fpruned90-none) from the SparseZoo onto SST2.
+Here, we will fine-tune a [90% pruned version of BERT](https://sparsezoo.neuralmagic.com/models/obert-base-wikipedia_bookcorpus-pruned90?comparison=obert-base-wikipedia_bookcorpus-base) from the SparseZoo onto SST2.
 
 ### **Kick off Training**
 
 We will use SparseML's `sparseml.transformers.text_classification` training script.
 
 To run sparse transfer learning, we first need to create/select a sparsification recipe. For sparse transfer, we need a recipe that instructs SparseML to maintain sparsity during training and to quantize the model. 
 
-For the SST2 dataset, there is a [transfer learning recipe available in SparseZoo](https://sparsezoo.neuralmagic.com/models/nlp%2Fsentiment_analysis%2Fobert-base%2Fpytorch%2Fhuggingface%2Fsst2%2Fpruned90_quant-none), identified by the following SparseZoo stub:
+For the SST2 dataset, there is a [transfer learning recipe available in SparseZoo](https://sparsezoo.neuralmagic.com/models/obert-base-sst2_wikipedia_bookcorpus-pruned90_quantized?comparison=obert-base-sst2_wikipedia_bookcorpus-base&tab=0), identified by the following SparseZoo stub:
 ```
 zoo:nlp/sentiment_analysis/obert-base/pytorch/huggingface/sst2/pruned90_quant-none
 ```

integrations/huggingface-transformers/tutorials/question-answering/question-answering-cli.md

@@ -41,7 +41,7 @@ In this tutorial, you will learn how to:
 Install SparseML via `pip`:
 
 ```bash
-pip install sparseml[torch]
+pip install sparseml[transformers]
 ```
 
 ## **Sparse Transfer Learning onto SQuAD**

integrations/huggingface-transformers/tutorials/sentiment-analysis/docs-sentiment-analysis-python-custom-teacher-rottentomatoes.ipynb

@@ -33,7 +33,7 @@
    },
    "outputs": [],
    "source": [
-    "!pip install sparseml[torch]"
+    "!pip install sparseml[transformers]"
    ]
   },
   {

integrations/huggingface-transformers/tutorials/sentiment-analysis/docs-sentiment-analysis-python-sst2.ipynb

@@ -45,7 +45,7 @@
    },
    "outputs": [],
    "source": [
-    "!pip install sparseml[torch]"
+    "!pip install sparseml[transformers]"
    ]
   },
   {

integrations/huggingface-transformers/tutorials/sentiment-analysis/sentiment-analysis-cli.md

@@ -41,7 +41,7 @@ In this tutorial, you will learn how to:
 Install SparseML via `pip`:
 
 ```bash
-pip install sparseml[torch]
+pip install sparseml[transformers]
 ```
 
 ## Sparse Transfer Learning onto SST2 (GLUE Task)

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

@@ -31,7 +31,7 @@ SparseZoo contains pre-sparsified checkpoints of common NLP models like BERT and
 Install via `pip`:
 
 ```
-pip install sparseml[torch]
+pip install sparseml[transformers]
 ```
 
 ## **Sparse Transfer Learning onto SST2**

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

@@ -31,7 +31,7 @@ SparseZoo contains pre-sparsified checkpoints of common NLP models like BERT and
 Install via `pip`:
 
 ```
-pip install sparseml[torch]
+pip install sparseml[transformers]
 ```
 
 ## **Example: Sparse Transfer Learning onto SST2**

integrations/huggingface-transformers/tutorials/text-classification/docs-text-classification-python-custom-teacher-tweeteval.ipynb

@@ -45,7 +45,7 @@
    },
    "outputs": [],
    "source": [
-    "!pip install sparseml[torch]"
+    "!pip install sparseml[transformers]"
    ]
   },
   {

integrations/huggingface-transformers/tutorials/text-classification/docs-text-classification-python-multi-label-go_emotions.ipynb

@@ -54,7 +54,7 @@
    },
    "outputs": [],
    "source": [
-    "!pip install sparseml[torch]"
+    "!pip install sparseml[transformers]"
    ]
   },
   {

integrations/huggingface-transformers/tutorials/text-classification/docs-text-classification-python-qqp.ipynb

@@ -45,7 +45,7 @@
    },
    "outputs": [],
    "source": [
-    "!pip install sparseml[torch]"
+    "!pip install sparseml[transformers]"
    ]
   },
   {

integrations/huggingface-transformers/tutorials/text-classification/docs-text-classification-python-sick.ipynb

@@ -44,7 +44,7 @@
    },
    "outputs": [],
    "source": [
-    "!pip install sparseml[torch]"
+    "!pip install sparseml[transformers]"
    ]
   },
   {

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

@@ -44,7 +44,7 @@ In this tutorial, you will learn how to:
 Install SparseML via `pip`:
 
 ```bash
-pip install sparseml[torch]
+pip install sparseml[transformers]
 ```
 
 ## SparseML CLI

integrations/huggingface-transformers/tutorials/token-classification/docs-token-classification-python-custom-teacher-wnut.ipynb

@@ -45,7 +45,7 @@
    },
    "outputs": [],
    "source": [
-    "!pip install sparseml[torch]"
+    "!pip install sparseml[transformers]"
    ]
   },
   {

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

@@ -41,7 +41,7 @@ In this tutorial, you will learn how to:
 Install SparseML via `pip`:
 
 ```bash
-pip install sparseml[torch]
+pip install sparseml[transformers]
 ```
 
 ## Sparse Transfer Learning onto Conll2003

integrations/torchvision/README.md

@@ -46,7 +46,7 @@ pip install sparseml[torchvision]
 
 Neural Magic has pre-sparsified versions of common Torchvision models such as ResNet-50. These models can be deployed directly or can be fine-tuned onto custom dataset via sparse transfer learning. This makes it easy to create a sparse image classification model trained on your dataset.
 
-[Check out the available models](https://sparsezoo.neuralmagic.com/?domain=cv&sub_domain=classification&page=1)
+[Check out the available models](https://sparsezoo.neuralmagic.com/?useCase=classification)
 
 ### Recipes
 
@@ -104,7 +104,7 @@ sparseml.image_classification.train \
 
 For full usage, run:
 ```bash
-sparseml.image_classification --help
+sparseml.image_classification.train --help
 ```
 
 ## Quick Start: Sparse Transfer Learning with the CLI
@@ -113,7 +113,7 @@ sparseml.image_classification --help
 
 Sparse Transfer is quite similiar to the typical transfer learning process used to train image classification models, where we fine-tune a checkpoint pretrained on ImageNet onto a smaller downstream dataset. With Sparse Transfer Learning, we simply start the fine-tuning process from a pre-sparsified checkpoint and maintain sparsity while the training process occurs.
 
-In this example, we will fine-tune a 95% pruned version of ResNet-50 ([available in SparseZoo](https://sparsezoo.neuralmagic.com/models/cv%2Fclassification%2Fresnet_v1-50%2Fpytorch%2Fsparseml%2Fimagenet%2Fpruned95_quant-none)) onto ImageNette.
+In this example, we will fine-tune a 95% pruned version of ResNet-50 ([available in SparseZoo](https://sparsezoo.neuralmagic.com/models/resnet_v1-50-imagenet-pruned95_quantized?comparison=resnet_v1-50-imagenet-base)) onto ImageNette.
 
 ### Kick off Training
 

integrations/ultralytics-yolov5/README.md

@@ -29,11 +29,9 @@ Once trained, SparseML enables you to export models to the ONNX format, such tha
 Install with `pip`:
 
 ```bash
-pip install sparseml[torchvision]
+pip install sparseml[yolov5]
 ```
 
-**Note**: YOLOv5 will not immediately install with this command. Instead, a sparsification-compatible version of YOLOv5 will install on the first invocation of the YOLOv5 code in SparseML.
-
 ## Tutorials
 
 - [Sparse Transfer Learning with the CLI](tutorials/sparse-transfer-learning.md) **[HIGHLY RECOMMENDED]**
@@ -91,15 +89,15 @@ SparseML inherits most arguments from the Ultralytics repository. [Check out the
 
 Sparse Transfer is very similiar to the typical transfer learing process used to train YOLOv5 models, where we fine-tune a checkpoint pretrained on COCO onto a smaller downstream dataset. With Sparse Transfer Learning, however, we simply start the fine-tuning process from a pre-sparsified checkpoint and maintain sparsity while the training process occurs.
 
-Here, we will fine-tune a [75% pruned-quantized version of YOLOv5s](https://sparsezoo.neuralmagic.com/models/cv%2Fdetection%2Fyolov5-s%2Fpytorch%2Fultralytics%2Fcoco%2Fpruned75_quant-none) onto VOC. 
+Here, we will fine-tune a [75% pruned-quantized version of YOLOv5s](https://sparsezoo.neuralmagic.com/models/yolov5-s-coco-pruned75_quantized?comparison=yolov5-s-coco-base&tab=0) onto VOC. 
 
 ### Kick off Training
 
 We will use SparseML's `sparseml.yolov5.train` training script.
 
 To run sparse transfer learning, we first need to create/select a sparsification recipe. For sparse transfer, we need a recipe that instructs SparseML to maintain sparsity during training and to quantize the model over the final epochs.
 
-For the VOC dataset, there is a [transfer learning recipe available in SparseZoo](https://sparsezoo.neuralmagic.com/models/cv%2Fdetection%2Fyolov5-s%2Fpytorch%2Fultralytics%2Fcoco%2Fpruned75_quant-none), identified by the following SparseZoo stub:
+For the VOC dataset, there is a [transfer learning recipe available in SparseZoo](https://sparsezoo.neuralmagic.com/models/yolov5-s-coco-pruned75_quantized?comparison=yolov5-s-coco-base&tab=0), found under the recipes tab and identified by the following SparseZoo stub:
 ```bash
 zoo:cv/detection/yolov5-s/pytorch/ultralytics/coco/pruned75_quant-none?recipe_type=transfer_learn
 ```

setup.py

@@ -70,7 +70,10 @@
     "torchvision>=0.3.0,<0.15",
     "torchaudio<=0.13",
 ]
-_pytorch_vision_deps = _pytorch_deps + ["torchvision>=0.3.0,<0.15"]
+_pytorch_vision_deps = _pytorch_deps + [
+    "torchvision>=0.3.0,<0.15",
+    "opencv-python<=4.6.0.66",
+]
 _transformers_deps = _pytorch_deps + [
     f"{'nm-transformers' if is_release else 'nm-transformers-nightly'}"
     f"~={version_nm_deps}",
@@ -251,7 +254,7 @@ def _setup_entry_points() -> Dict:
         [
             "sparseml.ultralytics.train=sparseml.yolov8.train:main",
             "sparseml.ultralytics.val=sparseml.yolov8.val:main",
-            "sparseml.ultralytics.export=sparseml.yolov8.export:main",
+            "sparseml.ultralytics.export_onnx=sparseml.yolov8.export:main",
         ]
     )