Naman Nandan

TorchServe `0.10.0` has been released which includes [worker auto recovery fix](https://github.com/pytorch/serve/pull/2746): https://github.com/pytorch/serve/releases/tag/v0.10.0

@gukwonku please set your `responseTimeout` to be greater than model inference time. Also, the worker recovery issue has been [fixed](https://github.com/pytorch/serve/pull/2746) and is included in the latest release: https://github.com/pytorch/serve/releases/tag/v0.10.0

Token authorization is now supported in TorchServe `0.10.0`: https://github.com/pytorch/serve/blob/v0.10.0/docs/token_authorization_api.md

@geraldstanje to answer your question > how is sagemaker than be able to use torchServe if they dont implement the ts.torch_handler.base_handler? The [PyTorch inference containers](https://github.com/aws/deep-learning-containers/blob/master/available_images.md#sagemaker-framework-containers-sm-support-only) that are compatible with SageMaker...

@boqigao Thank you for sharing your use-case and feature request! As you rightly pointed out, the worker initialization timeout is currently [hard coded to 2 minutes](https://github.com/pytorch/serve/blob/master/frontend/server/src/main/java/org/pytorch/serve/wlm/WorkerLifeCycle.java#L153-L159). This could potentially be...

@danial880 Once you've started torchserve and loaded the model could you please try the following API to check that the model has been successfully loaded and the workers have started...