Poor metric results after retraining maskrcnn using TLT notebook

Accelerated Computing Intelligent Video Analytics TAO Toolkit

Morganh September 2, 2020, 8:57am 13

Update: After finetune the blog’s warmup_steps from 0 to 1000, the AP is 0.31344375, which is closed to blog’s AP( 0.334154785 ).
Note that, I am using 8gpus(v100) training.

Upload spec file for reference.
maskrcnn_blog_finetune.txt (2.1 KB)

Training doesn't converge for Mapillary Vistas Dataset training with MaskRCNN

Topic		Replies	Views
MaskRCNN Input to reshape is a tensor with 3135248 values, but the requested shape has 2691200 TAO Toolkit	38	1120	May 9, 2023
Training doesn't converge for Mapillary Vistas Dataset training with MaskRCNN TAO Toolkit	47	1663	June 16, 2022
Maskrcnn.ipynb - followed notebook and ended up with poor (almost untrained) network from instructions TAO Toolkit	13	755	October 12, 2021
Training Custom FasterRCNN resnet50 Object detection issue TAO Toolkit	9	1112	October 12, 2021
Training Instance Segmentation Models Using Mask R-CNN on the NVIDIA Transfer Learning Toolkit Technical Blog	3	1007	August 18, 2021
Error while re-training with custom dataset using tlt file- FasterRCNN TAO Toolkit	5	352	June 26, 2023
Mask rcnn poor results TAO Toolkit	4	1035	October 12, 2021
Faster RCNN on TLT 3.0 not learning the same as TLT 2.0 TAO Toolkit	15	1005	October 12, 2021
Mask_rcnn shows training logs loss 0.00000 fast_rcnn class loss: 0.00000 fast_rcnn box loss: 0.00000 TAO Toolkit	22	799	August 16, 2022
UffParser: Validator error: block_4c_bn_3/cond/Switch: Unsupported operation _Switch TAO Toolkit tensorrt	38	1371	January 11, 2022

Poor metric results after retraining maskrcnn using TLT notebook

Related topics