open-mmlab / mmaction2

OpenMMLab's Next Generation Video Understanding Toolbox and Benchmark

https://mmaction2.readthedocs.io

open-mmlab/mmaction2 Issues

KeyError: 'ActionVisualizer is not in the mmengine::visualizer registry.[Bug]
Updated 3 days ago
训练时空动作检测，如何处理类别不均衡的问题
Updated 11 days ago1
vis_cam.py not work in timesformer
Updated 12 days ago1
[Bug] Training MVit with RawFrameDataset type
Updated 14 days ago
[Bug] long-video-demo raised AssertionError on assert crop_h == img_h or crop_w == img_w
Updated 16 days ago5
[Bug] batch_size problem
Updated 17 days ago
[Feature] Export the tsm_mobilnetv2 model to onnx format
Updated 18 days ago
top1_acc is allways times 0f 0.5or0.00[Bug]
Updated 18 days ago
train[Bug]
Closed 19 days ago
train time[Bug]
Closed 19 days ago
[Bug] much longer data_time when training PoseC3d
Updated 20 days ago2
[Bug] File 'visualize_heatmap_volume' does not exist
Updated 22 days ago3
[Bug] VisualizationHook is not set in the `default_hooks` field of config. Please set `visualization=dict(type="VisualizationHook")`
Updated 23 days ago
Do we now support the method of Skeleton-based Spatio-Temporal Action Detection and Action Recognition
Updated 24 days ago2
[Bug] It is not possible to use "analysis_tools\get_flops.py" to get the number of parameters and calculations
Updated 25 days ago3
[Bug]
Updated a month ago
[Docs]
Updated a month ago
[Bug] Fail to use confusion_matrix to RGBPoseC3D
Updated a month ago3
[Bug] resize_videos.py : get invalid literal for int() with base 10: ''
Closed a month ago1
[Docs] Custom Models?
Updated a month ago1
[Docs] How to train a model with different inputs from different model feature and videos
Updated a month ago1
[Docs]
Updated a month ago1
model zoo lost
Updated 2 months ago
[Docs] How to fine tune BMN when using ActionNet data prep method 2
Updated 2 months ago1
[Bug] Use mmaction2 - main/configs/skeleton/posec3d/rgbpose_conv3d/rgbpose_conv3d py error for the demo
Updated 2 months ago
[Docs] Custom Dataset Training with PoseC3D
Closed 2 months ago2
[Bug] different performance about batch size on val pipeline and test pipeline
Updated 2 months ago
[Docs]How to add YOLO target detection frame to frame people when reasoning in TimeSformer
Updated 2 months ago
[Feature] How to get flops of the MMRecognizer3D?
Updated 2 months ago
CUDA error: CUBLAS_STATUS_EXECUTION_FAILED when calling `cublasSgemm
Updated 2 months ago
Fast inference option for MMAction2 models, like SlowFast
Updated 2 months ago1
[Feature]
Updated 3 months ago
[Docs] Custom action localization dataset. anet_anno_train.json missing
Closed 3 months ago
[Bug] 'ActionVisualizer is not in the mmengine::visualizer registry.
Updated 3 months ago1
[Docs] ACRN training on AVA dataset at 15 fps
Updated 3 months ago
Installation fails on macbook m2 [Docs]
Updated 3 months ago
[Docs] 20分钟了解 MMAction2 框架设计，步骤2导入问题
Updated 4 months ago
[Docs] Fine-tuning ActionClip
Updated 4 months ago
[Bug] How can I use clip_feature_extraction.py to input the rawframes , and then output the RGB features of each frame?
Updated 4 months ago
找不到rtmpose的ckpt[Bug]
Updated 4 months ago
[Feature] Rename the confusing argument 'format_shape' in ActionDataPreprocessor
Updated 4 months ago
get_flops.py[Bug]
Updated 4 months ago
[Bug] A IndexError is encountered when training a portion of the ntu60 dataset
Updated 5 months ago
[Docs] Running into problems training UCF101 with Slowfast
Updated 5 months ago
[Feature] How to run tracking config
Updated 5 months ago
[Bug]
Updated 5 months ago2
[Docs] https://mmaction2.readthedocs.io/ does not respond
Closed 5 months ago1
[Bug] ntu_pose_extraction in tools/data/skeleton gives keypoints in COCO format instead of NTU RGBD
Updated 5 months ago
[Docs] YOLOv8
Updated 5 months ago3
ImportError: Please install mmdet to use init_detector. For OpenMMLAB codebases, you may need to install mmcv-full first before you install the particular codebase.
Updated 5 months ago