You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I want to measure the throughput of a file storage system when saving checkpoints during fine-tuning using LitGPT. The goal is to evaluate the performance of the storage system under real-world conditions, particularly during the intensive I/O operations involved in saving large checkpoints.
While I am able to run the experiment, the saving or reading throughput would not pass 700 MB. Is there any way to increase this throughput?
The text was updated successfully, but these errors were encountered:
I want to measure the throughput of a file storage system when saving checkpoints during fine-tuning using LitGPT. The goal is to evaluate the performance of the storage system under real-world conditions, particularly during the intensive I/O operations involved in saving large checkpoints.
While I am able to run the experiment, the saving or reading throughput would not pass 700 MB. Is there any way to increase this throughput?
The text was updated successfully, but these errors were encountered: