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 found that when the context length is 512k, the training speed is too slow, which is different from your experimental results. It takes 585 seconds for training a batch of 512k , which is to 512000/585.85=873.94 tokens/s
And I used A100-80G*8 with NVLINK.
I found that when the context length is 512k, the training speed is too slow, which is different from your experimental results. It takes 585 seconds for training a batch of 512k , which is to 512000/585.85=873.94 tokens/s
And I used A100-80G*8 with NVLINK.
The text was updated successfully, but these errors were encountered: