I was able to get the calibration to succeed by increasing the number of images used from 500 to 7500. Answer found here https://devtalk.nvidia.com/default/topic/1036911/tensorrt/cannot-run-sampleuffssd-for-int8-inference-calibration-failed-/post/5281105/#5281105
Related topics
Topic | Replies | Views | Activity | |
---|---|---|---|---|
sampleUffSSD conversion fails? (KeyError: 'image_tensor') | 22 | 4142 | October 12, 2021 | |
sampleUffSSD with custom ssd_mobilenet_v1 model | 37 | 4490 | October 12, 2021 | |
sampleUffSSD with newer tensorflow models (2018) | 16 | 2515 | November 13, 2020 | |
[TensorRT] ERROR: Parameter check failed at: Utils.cpp::reshapeWeights::71, condition: input.values != nullptr | 13 | 5618 | October 12, 2021 | |
Tensorrt support for SSD_inception trained on custom dataset | 15 | 2619 | October 12, 2021 | |
Problems with SSD Mobilenet v2 UFF | 35 | 7941 | October 18, 2021 | |
Tensorrt fails for custom ssd_inception Model | 18 | 2803 | May 14, 2020 | |
ERROR: UFFParser: Parser error: BoxPredictor_0/Reshape: Reshape: -1 dimension specified more than 1 ... | 25 | 5610 | September 24, 2020 | |
How to retrain ssd_inception_v2_coco_2017_11_17 from the tensorrt - samples | 33 | 6952 | October 18, 2021 | |
run SSD_MobileNetV2 (Tensorflow object detection API) on TensorRT | 18 | 5041 | October 12, 2021 |