I am using a ssd-inception v2 pretrained object detection model from
GitHub - dusty-nv/jetson-inference: Hello AI World guide to deploying deep-learning inference networks and deep vision primitives with TensorRT and NVIDIA Jetson.
Can I have the evaluation results when the model was trained with COCO dataset?
What was the accuracy of the training, validation, and test dataset in terms of mean average precision (mAP) for different Intersections over Union (IoU)?
Quick help will be much appreciated.
@tareq.khan, the original ssd-inception-v2-coco model came from the TensorFlow Model Zoo with a mAP of 24:
This file has been truncated.
# TensorFlow 1 Detection Model Zoo
We provide a collection of detection models pre-trained on the
[COCO dataset](http://cocodataset.org), the
[Kitti dataset](http://www.cvlibs.net/datasets/kitti/), the
[Open Images dataset](https://storage.googleapis.com/openimages/web/index.html),
the [AVA v2.1 dataset](https://research.google.com/ava/) the
[iNaturalist Species Detection Dataset](https://github.com/visipedia/inat_comp/blob/master/2017/README.md#bounding-boxes)
[Snapshot Serengeti Dataset](http://lila.science/datasets/snapshot-serengeti).
These models can be useful for out-of-the-box inference if you are interested in
categories already in those datasets. They are also useful for initializing your
models when training on novel datasets.
In the table below, we list each such pre-trained model including:
* a model name that corresponds to a config file that was used to train this
Thanks for the quick reply. Really appreciate.
Three more questions.
How many images were used for training, validation, and testing?
Can I get the confusion matrix regarding predicting the object labels for the test set?
Can mAP of 24 be expressed as a normalized number from 0 to 1?
@tareq.khan, sorry I don’t know the specifics of how that model was trained, as it came pre-trained from TensorFlow. I would recommend filing a GitHub issue against the TF model zoo repo if you require more details.