[FEATURE REQUEST] Is it possible to train a 3B openflamingo on 4 V100 on one node with deepspeed by simplifying pre-training data used？

[Yang-bug-star]

Is your feature request related to a problem? Please describe.
A clear and concise description of what the problem is. Ex. I'm always frustrated when [...]

Describe the workflow you want to enable.
A clear and concise description of what you want to happen.

Describe your proposed solution.
How do you propose to address this? A high-level description is fine, but detailed suggestions are also welcome.

Describe alternatives you've considered
A clear and concise description of any alternative solutions or features you've considered.

Additional context
Add any other context or screenshots about the feature request here.

Are you willing to help implement this feature?
If so, please comment on how long you expect it to take or what kind of support you would require from the OpenFlamingo team!