tacotron2 | Tacotron 2 - PyTorch implementation | Speech library

As pointed out elsewhere the basic problem is that QProcess goes to some trouble to avoid going through any shell. Having said that you should be able to achieve what you want by invoking a shell explicitly.

Let's say the command you would usually run under bash is ls -l | grep '\.' ...

AWS Lambdas local storage in /tmp is only 512MB. You are apparently exceeding this limit.

There are five solutions I can think of:

1. Mount a EFS volume (which already contains your trained model) to the Lambda.
2. Reduce the size of your model.
3. Stream the model in chunks to your Lambda (might be hard).
4. Not use Lambda (maybe just a plain EC2 or EKS).
5. Use a Docker container that already contains your model as Lambda.

It is hard to tell what the best solution for you is, since so much information is missing. But those solutions should give you a good starting point.

Cloud Run fully managed don't support the GPU. I would like to say not, except if the model can work (slowly) in a non GPU environment.

The alternative is to use Cloud Run for Anthos, on your own GKE cluster. In this case, you can choose the node pool configuration that you prefer, with GPU and you can. But it's not serverless, you have to manage yourselves the cluster and you have to pay it full time (don't scale to 0 like Cloud Run fully managed)

As mentioned here, the error might have nothing to do with "symlinks", and everything with the lack of Dockerfile, which should be in the Tacotron-2/docker folder.

docker build does mention:

The docker build command builds Docker images from a Dockerfile and a “context”.
A build’s context is the set of files located in the specified PATH or URL.

In your case, docker build -t tacotron-2_image docker/ is supposed to be executed in the path you have cloned the Rayhane-mamah/Tacotron-2 repository.

To be sure, you could specify said Dockerfile, but that should not be needed: