diff --git a/examples/object_detection_compression/README.md b/examples/object_detection_compression/README.md index 9ac368a5ea0ae4f0bbaf0c4eda4224fa21af6aaa..654e2a70896711b62a7b613fea2ca580c792a29a 100644 --- a/examples/object_detection_compression/README.md +++ b/examples/object_detection_compression/README.md @@ -54,4 +54,38 @@ memory errors. The default model is `torchvision.models.detection.maskrcnn_resnet50_fpn`, you can specify any model that is part of `torchvision.models.detection` using - the `--model` argument, e.g. `--model maskrcnn_resnet50_fpn`. \ No newline at end of file + the `--model` argument, e.g. `--model maskrcnn_resnet50_fpn`. + +## Results +We present the results of pruning up to 85% sparsity using distiller. +**NOTE**- AP = Average Precision, AR = Average Recall +### Bounding Box: +| Metric | IoU Thresholds | Area | Max Detections | Baseline | 85% Sparsity | +| ------ | --- | ---- | -------------- | -------- | ------------ | +| (AP) | 0.50:0.95 | all | 100 | 0.379 | 0.352 | +| (AP) | 0.50 | all | 100 | 0.592 | 0.558 | +| (AP) | 0.75 | all | 100 | 0.410 | 0.376 | +| (AP) | 0.50:0.95 | small | 100 | 0.215 | 0.188 | +| (AP) | 0.50:0.95 | medium | 100 | 0.414 | 0.389 | +| (AP) | 0.50:0.95 | large | 100 | 0.495 | 0.460 | +| (AR) | 0.50:0.95 | all | 1 | 0.312 | 0.302 | +| (AR) | 0.50:0.95 | all | 10 | 0.494 | 0.481 | +| (AR) | 0.50:0.95 | all | 100 | 0.518 | 0.504 | +| (AR) | 0.50:0.95 | small | 100 | 0.321 | 0.300 | +| (AR) | 0.50:0.95 | medium | 100 | 0.559 | 0.546 | +| (AR) | 0.50:0.95 | large | 100 | 0.660 | 0.646 | +### Segmentation: +| Metric | IoU Thresholds| Area | Max Detections | Baseline | 85% Sparsity | +| ------ | --- | ---- | -------------- | -------- | ------------ | +| (AP) | 0.50:0.95 | all | 100 | 0.346 | 0.320 | +| (AP) | 0.50 | all | 100 | 0.561 | 0.525 | +| (AP) | 0.75 | all | 100 | 0.367 | 0.338 | +| (AP) | 0.50:0.95 | small | 100 | 0.156 | 0.134 | +| (AP) | 0.50:0.95 | medium | 100 | 0.373 | 0.346 | +| (AP) | 0.50:0.95 | large | 100 | 0.509 | 0.482 | +| (AR) | 0.50:0.95 | all | 1 | 0.294 | 0.284 | +| (AR) | 0.50:0.95 | all | 10 | 0.454 | 0.439 | +| (AR) | 0.50:0.95 | all | 100 | 0.474 | 0.459 | +| (AR) | 0.50:0.95 | small | 100 | 0.269 | 0.250 | +| (AR) | 0.50:0.95 | medium | 100 | 0.515 | 0.499 | +| (AR) | 0.50:0.95 | large | 100 | 0.631 | 0.618 |