Q&A

Q1: Why does my model not show up in /v1/models?

Following the steps below to debug your model:

  1. Check your config.json file:

  • Check if it is valid JSON format:

    import json
    config = json.load(open('config.json', 'r'))
    

    This code should run without exception.

  • If you are using custom transforms, make sure you follow the instructions in Bring your own Transforms.

  1. Check your model file:

  1. Upload model to AIAA again:

Once you make sure all the pieces are correct, upload your model again to AIAA.

  1. Increase trtis_model_timeout:

AIAA will poll TRTIS for this amount of time before AIAA claims the model is not imported correctly. If you are using TRTIS engine (which is the default case), you can try using a larger timeout to ensure the model import success. e.g.: start_aas.sh --trtis_model_timeout 120.

  1. Check your logs:

If all the above steps do not work, start using flag --debug 1 and check log files in /workspace/logs. You can also go to Nvidia Developer Forums.

Note

Currently, AIAA requires models to have a single input and a single output. Multi-class segmentation can be achieved by having multiple channels in output.

Q2: Why are the models returning bad results?

Most of the time, this is caused by the mismatch of data. Make sure your testing data in AIAA have the same characteristics as the data that you used to train your models.

That would include the following:

  1. Resolution/Spacing

  2. Orientation

  3. Contrast/Phase

For example, the pre-trained segmentation models on NGC are using data from Medical Segmentation Decathlon.

We re-scale the image to have a spacing of [1.0, 1.0, 1.0] and make sure the affine matrix of Nifti have all positive values.

Hint

Clara Train API provides some nice transforms to tackle the resolution and orientation problems.

Tip

If you trained your model with data augmentation like RandomAxisFlip and RandomZoom then it will be insensitive to orientation.

Q3: Does AIAA support 2D models?

Yes, we do support 2D models. However, this is only supported by directly interacting with the AIAA server API via HTTP post requests. (Please refer to Tutorial: Brain Segmentation PyTorch for an example.)

We are planning to support 2D models in other clients in the future.

Q4: What if my GPU card does not have enough memory?

If your GPU card is very tight on memory, you can do some of the following points to alleviate this:

  1. Load fewer models in AIAA server

  2. Reduce roi (the size of scanning window) in config_aiaa.json

  3. Try to reduce your network size

Q5: Why can’t I start AIAA?

If you start the docker using --net=host, make sure AIAA port and TRTIS ports are not used by other processes.

If you use -p [host port]:[docker port] to run docker, then just make sure the [host port] is not used by other processes.

Q6: How can I start the AIAA server clean?

To start it all clean, remove the workspace folder and create a new one. Then start AIAA server with the new workspace.

Q7: What is the format of data that AIAA expect?

You can provide your own data loader to load data in any format you want (png, jpg, NumPy array). Please refer to Bring your own Data Loader.

Notice that AIAA currently does not support batching, it supports inference on one image/volume for each request. So you need to make sure the ShapeFormat in the end of your pre-transforms chain should not have the batch dimension (“N”). If you are writing custom transforms, make sure you take care of ShapeFormat.

Q8: Why is AIAA occupying all the GPU memories when I am not running any inference?

When AIAA runs with TRTIS backend, it will put one model instance on every GPU that is visible inside the docker.

Users can use -e NVIDIA_VISIBLE_DEVICES=[ids of the GPU you want to use] to control what GPUs are visible. For the number of model instances on each GPU, users can modify gpu_instance_count under trtis in their model configs.

When a model instance is loaded in GPU, even if it is not serving any inference requests at that moment, it will occupy some amount of GPU memory. As a result, if users want to free that GPU memory, they will have to either stop the AIAA server or unload some models (using DELETE model API).

Q9: Does AIAA use apache?

Yes. Advanced users can modify apache configs for AIAA which are normally located at /opt/nvidia/medical/nvmidl/apps/aas/www/conf/ in the docker.

By default, it runs as www-data user/group for security reasons. Hence the ownership of AIAA workspace will get modified accordingly.

Q10: Can I run multiple containers of AIAA in the same host?

Yes. But you have to make sure you are using different ports for TRTIS and they do not overlap. In such cases avoid using --net=host and use direct port mapping to make sure AIAA port and TRTIS ports are not used by other processes.

You use -p [host port]:[docker port] to run docker and make sure [host port] is not used by other processes. For example -p 9000:80 to map a different host port for HTTP access and -p 9001:443 for HTTPS.

Then you can try

  • curl http://127.0.0.1:9000/v1/models

  • curl --insecure https://127.0.0.1:9001/v1/models (if you are running AIAA in ssl mode)

Also recommended to use a different port for TRTIS server while starting AIAA. For example: start_aiaa.sh --trtis_port 8500

Note

Apache inside docker always runs at HTTP port 80 and SSL port 443

Hint

More discussions can be found in Nvidia Developer Forums