aquariusjay

Please make sure the checkpoint is properly loaded. 1. The ckpt downloaded from projects/panoptic_deeplab.md is all trained. They should not be used as the initial_checkpoint. If you simply want to...

For reference, please see [cityscapes_test_server_evaluation.md](https://github.com/google-research/deeplab2/blob/main/g3doc/setup/cityscapes_test_server_evaluation.md) and [coco_test_server_evaluation.md](https://github.com/google-research/deeplab2/blob/main/g3doc/setup/coco_test_server_evaluation.md). Hope that helps.

Hi @prabal27, Thanks for the issue. We have not verified converting the model to TF-Lite framework, and thus we are not sure if it is fully supported. Cheers,

Hi @prabal27, Thanks for asking. We currently have no plan to add TF-Lite support yet, given the limited bandwith. Cheers,

Hi @abaybektursun, Thanks for bringing up the issue. We are working on a new DeepLab_demo.ipynb, using kMaX-DeepLab checkpoints. Please stay tuned. Cheers,

Hi @abaybektursun, To keep you updated, we have just open-sourced a new ipynb demo using kMaX-DeepLab checkpoints. You may want to take a look. Cheers,

Hi @rcruzgar, Thanks for the issue. However, it goes beyond our scope to help you debug. We would suggest you run our provided tutorials (e.g., [Cityscapes](https://github.com/google-research/deeplab2/blob/main/g3doc/setup/cityscapes.md)). Cheers,

Hello, Thanks for reporting the issue. Unfortunately, if you want to train a semantic-only model, you could not use the trained panoptic checkpoints for initialization (as shown in the error...

You should be able to use multi-gpu by specifying the [num_gpu](https://github.com/google-research/deeplab2/blob/0e51b1d43996f75d3f50a1bcccf4ea62eee1803d/g3doc/setup/getting_started.md#putting-everything-together) flag.