I transfered detectnetV2 etlt model to trt engine int8 using cal.bin,can get similar detections but low scores(like 0.9(fp32) to 0.39(int8)),how to get the confidence value in tlt3.0?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Trained tlt model works with more than 95% accuracy but exported etlt performs poorly | 8 | 676 | April 19, 2023 | |
| Detectnet_v2 print confidence score in infrence image | 3 | 972 | December 31, 2020 | |
| TLT YOLOv3 Int8 can not detect anything | 16 | 1947 | December 10, 2020 | |
| Poor Result After INT8 Optimization (TLT Getting Started Guide) | 31 | 1905 | November 11, 2020 | |
| Error in export a DetectNet_v2 model in INT8 mode | 8 | 1172 | September 23, 2020 | |
| TLT trained model accuracy worse after deployment | 10 | 991 | April 29, 2021 | |
| Issue running tlt trained SSD-resnet18 on Xavier with deepstream-app | 14 | 873 | September 21, 2020 | |
| Why is tlt-infer’s label confidence threshold high above 1? | 8 | 674 | May 19, 2020 | |
| Transfer learning tool detectnet_v2.ipynp Visualize Inferences problem | 2 | 577 | August 9, 2020 | |
| Inferring Yolo_v3.trt model in python | 37 | 3963 | March 17, 2021 |