Combine two ggml_tensor like torch.cat

[sankalpdayal]

I want to combine two embedding tensors where one of the embedding tensor is pre-stored which I plan to read from file. Is there a function

1. Creates a ggm_tensor from two ggm_tensors of same dimensionality.
2. Loads a ggml tensors from lets say numpy object. Even csv should do for now.

[ggerganov]

This is how you can implement torch.cat with existing ggml operators:

https://github.com/ggerganov/ggml/blob/80f21b878163608209604a11083557a791342cf0/examples/sam/main.cpp#L1066-L1071