Register
Login
Resources
Docs Blog Datasets Glossary Case Studies Tutorials & Webinars
Product
Data Engine LLMs Platform Enterprise
Pricing Explore
Connect to our Discord channel
Resources  
Docs Blog Datasets Glossary Case Studies Tutorials & Webinars
Explore Pricing
Product  
Data Engine LLMs Platform Enterprise
Connect to Dagshub Discord channel
Register
Login

Deci-AI
/
super-gradients
connected to https://github.com/Deci-AI/super-gradients.git

Watch 4
Issues 62 Pull Requests 44 Discussions 0

Makefile 7.3 KB
Permalink History Raw

You have to be logged in to leave a comment. Sign In
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87 
  1. unit_tests:
    2. 

  2. 	python -m unittest tests/deci_core_unit_test_suite_runner.py
    3. 

  3. 

  4. integration_tests:
    5. 

  5. 	python -m unittest tests/deci_core_integration_test_suite_runner.py
    6. 

  6. 

  7. yolo_nas_integration_tests:
    8. 

  8. 	python -m unittest tests/integration_tests/yolo_nas_integration_test.py
    9. 

  9. 

  10. recipe_accuracy_tests:
    11. 

  11. 	python src/super_gradients/train_from_recipe.py --config-name=coco2017_pose_dekr_w32_no_dc experiment_name=shortened_coco2017_pose_dekr_w32_ap_test epochs=1 batch_size=4 val_batch_size=8 training_hyperparams.lr_warmup_steps=0 training_hyperparams.average_best_models=False training_hyperparams.max_train_batches=1000 training_hyperparams.max_valid_batches=100 multi_gpu=DDP num_gpus=4
    12. 

  12. 	python src/super_gradients/train_from_recipe.py --config-name=cifar10_resnet               experiment_name=shortened_cifar10_resnet_accuracy_test   epochs=100 training_hyperparams.average_best_models=False multi_gpu=DDP num_gpus=4
    13. 

  13. 	python src/super_gradients/train_from_recipe.py --config-name=coco2017_yolox               experiment_name=shortened_coco2017_yolox_n_map_test      epochs=10  architecture=yolox_n training_hyperparams.loss=YoloXFastDetectionLoss training_hyperparams.average_best_models=False multi_gpu=DDP num_gpus=4
    14. 

  14. 	python src/super_gradients/train_from_recipe.py --config-name=cityscapes_regseg48          experiment_name=shortened_cityscapes_regseg48_iou_test   epochs=10 training_hyperparams.average_best_models=False multi_gpu=DDP num_gpus=4
    15. 

  15. 	python src/super_gradients/examples/convert_recipe_example/convert_recipe_example.py --config-name=cifar10_conversion_params experiment_name=shortened_cifar10_resnet_accuracy_test
    16. 

  16. 	coverage run --source=super_gradients -m unittest tests/deci_core_recipe_test_suite_runner.py
    17. 

  17. 

  18. 

  19. sweeper_test:
    20. 

  20. 	python -m super_gradients.train_from_recipe -m --config-name=cifar10_resnet \
    21. 

  21. 	  ckpt_root_dir=$$PWD \
    22. 

  22. 	  experiment_name=sweep_cifar10 \
    23. 

  23. 	  training_hyperparams.max_epochs=1 \
    24. 

  24. 	  training_hyperparams.initial_lr=0.001,0.01
    25. 

  25. 

  26. 	# Make sure that experiment_dir includes $$expected_num_dir subdirectories. If not, fail
    27. 

  27. 	subdir_count=$$(find "$$PWD/sweep_cifar10" -mindepth 1 -maxdepth 1 -type d | wc -l); \
    28. 

  28. 	if [ "$$subdir_count" -ne 2 ]; then \
    29. 

  29. 	  echo "Error: $$PWD/sweep_cifar10 should include 2 subdirectories but includes $$subdir_count."; \
    30. 

  30. 	  exit 1; \
    31. 

  31. 	fi
    32. 

  32. 

  33. # Here you define a list of notebooks we want to execute and convert to markdown files
    34. 

  34. # NOTEBOOKS = hellomake.ipynb hellofunc.ipynb helloclass.ipynb
    35. 

  35. NOTEBOOKS = src/super_gradients/examples/model_export/models_export.ipynb
    36. 

  36. 

  37. # This Makefile target runs notebooks listed below and converts them to markdown files in documentation/source/
    38. 

  38. run_and_convert_notebooks_to_docs: $(NOTEBOOKS)
    39. 

  39. 	jupyter nbconvert --to markdown --output-dir="documentation/source/" --execute $@
    40. 

  40. 

  41. # This Makefile target runs notebooks listed below and converts them to markdown files in documentation/source/
    42. 

  42. check_notebooks_version_match: $(NOTEBOOKS)
    43. 

  43. 	python tests/verify_notebook_version.py $@
    44. 

  44. 

  45. WANDB_PARAMS = training_hyperparams.sg_logger=wandb_sg_logger +training_hyperparams.sg_logger_params.api_server=https://wandb.research.deci.ai +training_hyperparams.sg_logger_params.entity=super-gradients training_hyperparams.sg_logger_params.launch_tensorboard=false training_hyperparams.sg_logger_params.monitor_system=true +training_hyperparams.sg_logger_params.project_name=PoseEstimation training_hyperparams.sg_logger_params.save_checkpoints_remote=true training_hyperparams.sg_logger_params.save_logs_remote=true training_hyperparams.sg_logger_params.save_tensorboard_remote=false training_hyperparams.sg_logger_params.tb_files_user_prompt=false
    46. 

  46. 

  47. coco2017_yolo_nas_pose_s:
    48. 

  48. 	python src/super_gradients/train_from_recipe.py --config-name=coco2017_yolo_nas_pose_s $(WANDB_PARAMS)
    49. 

  49. 

  50. coco2017_yolo_nas_pose_m:
    51. 

  51. 	python src/super_gradients/train_from_recipe.py --config-name=coco2017_yolo_nas_pose_m $(WANDB_PARAMS)
    52. 

  52. 

  53. coco2017_yolo_nas_pose_m_resume:
    54. 

  54. 	python src/super_gradients/train_from_recipe.py --config-name=coco2017_yolo_nas_pose_m $(WANDB_PARAMS) resume=True
    55. 

  55. 

  56. coco2017_yolo_nas_pose_l:
    57. 

  57. 	python src/super_gradients/train_from_recipe.py --config-name=coco2017_yolo_nas_pose_l $(WANDB_PARAMS)
    58. 

  58. 

  59. coco2017_yolo_nas_pose_l_no_ema:
    60. 

  60. 	python src/super_gradients/train_from_recipe.py --config-name=coco2017_yolo_nas_pose_l $(WANDB_PARAMS) training_hyperparams.ema=False
    61. 

  61. 

  62. coco2017_yolo_nas_pose_n:
    63. 

  63. 	python src/super_gradients/train_from_recipe.py --config-name=coco2017_yolo_nas_pose_n $(WANDB_PARAMS)
    64. 

  64. 

  65. coco2017_yolo_nas_pose_n_resume:
    66. 

  66. 	python src/super_gradients/train_from_recipe.py --config-name=coco2017_yolo_nas_pose_n $(WANDB_PARAMS) resume=True
    67. 

  67. 

  68. coco2017_yolo_nas_pose_l_resume:
    69. 

  69. 	python src/super_gradients/train_from_recipe.py --config-name=coco2017_yolo_nas_pose_l $(WANDB_PARAMS) resume=True
    70. 

  70. 

  71. 

  72. coco2017_pose_former_b2:
    73. 

  73. 	python src/super_gradients/train_from_recipe.py --config-name=coco2017_former_pose_b2 $(WANDB_PARAMS)
    74. 

  74. 

  75. LOCAL_COCO_TRAINING_PARAMS = dataset_params.train_dataset_params.data_dir=/home/bloodaxe/data/coco2017 dataset_params.val_dataset_params.data_dir=/home/bloodaxe/data/coco2017 num_gpus=4 multi_gpu=DDP
    76. 

  76. LOCAL_TRAINING_PARAMS = dataset_params.train_dataset_params.data_dir=/home/bloodaxe/data/crowdpose dataset_params.val_dataset_params.data_dir=/home/bloodaxe/data/crowdpose num_gpus=1 multi_gpu=Off
    77. 

  77. 

  78. crowdpose_yolo_nas_pose_s:
    79. 

  79. 	#	CUDA_VISIBLE_DEVICES=0 python src/super_gradients/train_from_recipe.py --config-name=crowdpose_yolo_nas_pose_s $(LOCAL_TRAINING_PARAMS) &
    80. 

  80. 	#	CUDA_VISIBLE_DEVICES=1 python src/super_gradients/train_from_recipe.py --config-name=crowdpose_yolo_nas_pose_s $(LOCAL_TRAINING_PARAMS) training_hyperparams.criterion_params.classification_loss_type=bce &
    81. 

  81. 	#	CUDA_VISIBLE_DEVICES=2 python src/super_gradients/train_from_recipe.py --config-name=crowdpose_yolo_nas_pose_s $(LOCAL_TRAINING_PARAMS) training_hyperparams.criterion_params.dfl_loss_weight=1.0 &
    82. 

  82. 	#	CUDA_VISIBLE_DEVICES=3 python src/super_gradients/train_from_recipe.py --config-name=crowdpose_yolo_nas_pose_s $(LOCAL_TRAINING_PARAMS) training_hyperparams.criterion_params.pose_cls_loss_weight=0.1 training_hyperparams.criterion_params.pose_reg_loss_weight=5.0 &
    83. 

  83. 

  84. 	CUDA_VISIBLE_DEVICES=0 python src/super_gradients/train_from_recipe.py --config-name=crowdpose_yolo_nas_pose_s $(LOCAL_TRAINING_PARAMS) training_hyperparams.criterion_params.pose_cls_loss_weight=0.10 training_hyperparams.criterion_params.pose_reg_loss_weight=5.0 training_hyperparams.criterion_params.dfl_loss_weight=1.0 &
    85. 

  85. 	CUDA_VISIBLE_DEVICES=1 python src/super_gradients/train_from_recipe.py --config-name=crowdpose_yolo_nas_pose_s $(LOCAL_TRAINING_PARAMS) training_hyperparams.criterion_params.pose_cls_loss_weight=0.01 training_hyperparams.criterion_params.pose_reg_loss_weight=5.0 training_hyperparams.criterion_params.dfl_loss_weight=1.0 &
    86. 

  86. 	CUDA_VISIBLE_DEVICES=2 python src/super_gradients/train_from_recipe.py --config-name=crowdpose_yolo_nas_pose_s $(LOCAL_TRAINING_PARAMS) training_hyperparams.criterion_params.pose_cls_loss_weight=0.01 training_hyperparams.criterion_params.pose_reg_loss_weight=1.0 training_hyperparams.criterion_params.dfl_loss_weight=2.5 &
    87. 

  87. 	CUDA_VISIBLE_DEVICES=3 python src/super_gradients/train_from_recipe.py --config-name=crowdpose_yolo_nas_pose_s $(LOCAL_TRAINING_PARAMS) training_hyperparams.criterion_params.pose_cls_loss_weight=0.10 training_hyperparams.criterion_params.pose_reg_loss_weight=5.0 training_hyperparams.criterion_params.dfl_loss_weight=1.0 training_hyperparams.criterion_params.classification_loss_type=bce &
    88.
Prev
Next
Tip!

Press p or to see the previous file or, n or to see the next file

Comments

Write Preview
Loading...
Integrate AWS S3
Use S3 remote
Select bucket
Access key
Finish

Use AWS S3 as storage!

Browsing data directories saved to S3 is possible with DAGsHub. Let's configure your repository to easily display your data in the context of any commit!

Specify your S3 bucket

Access key (If needed)

Congratulations!

super-gradients is now integrated with AWS S3!

Delete Storage Key

Are you sure you want to delete this access key?

No
Yes
Integrate Google Cloud Storage
Use Google Storage
Select bucket
Upload key
Finish

Use Google Cloud Storage!

Browsing data directories saved to Google Cloud Storage is possible with DAGsHub. Let's configure your repository to easily display your data in the context of any commit!

Specify your Google Storage bucket

Service Account Key

Congratulations!

super-gradients is now integrated with Google Cloud Storage!

Delete Storage Key

Are you sure you want to delete this access key?

No
Yes
Integrate Azure Cloud Storage
Use Azure Storage
Select bucket
Set key
Finish

Use Azure Cloud Storage!

Browsing data directories saved to Azure Cloud Storage is possible with DAGsHub. Let's configure your repository to easily display your data in the context of any commit!

Specify your Azure Storage bucket

Access key (If needed)

Congratulations!

super-gradients is now integrated with Azure Cloud Storage!

Delete Storage Key

Are you sure you want to delete this access key?

No
Yes
Integrate S3 compatible storage
Use S3 like remote
Select bucket
Access key
Finish

Use any S3 compatible storage!

Browsing data directories saved to S3 compatible storage is possible with DAGsHub. Let's configure your repository to easily display your data in the context of any commit!

Specify your S3 bucket

Bucket name cannot be the same as the repository name. Please change one of them.

Access key (If needed)

Congratulations!

super-gradients is now integrated with your S3 compatible storage!

Delete Storage Key

Are you sure you want to delete this access key?

No
Yes