Update lit_gpt commit to PyTorch 2.2

[Chris113113]

The PR's primary purpose is updating lit-gpt's commit to a PyTorch 2.2 commit. This also comes with a few other things:

Logs from new image: http://shortn/_klGl5LKuQm

• Removes building of flash_attn resulting in more reliable lit-gpt image builds.
• Removes Lightning.Trainer in favor of Lightning.Fabric (provided by the lit-gpt openwebtext sample)
• Refactors most of the nsight workflow into a separate Callback class.
• Updates the base image to 24.01, which comes bundled with PyTorch 2.2, and NCCL 2.19.
• Updates the ncclPlugin and rxdmContainer versions
• Adds volume mount for /data onto the host which prevents re-downloading the dataset on subsequent runs.

[Chris113113]

Mostly LGTM. In addition to my comments, two things:

1. From this PR it looks like we are adjusting this example to run Llama-2-13B instead of Llama-2-70B. Just want to double check this is intentional
2. Could you attach some link (using short-gen) where you are able to run this workload?

1. Good catch on 13B, I was using it to experiment.

2. http://shortn/_klGl5LKuQm, added to description.