jax
0cfed4ef
- [pallas:mosaic_gpu] Shrink `max_concurrent_iteration` based on the total number of steps
Go
Login via GitHub
Home
Pricing
FAQ
Install
Login
via GitHub
Commit
View On
GitHub
Commit
1 year ago
[pallas:mosaic_gpu] Shrink `max_concurrent_iteration` based on the total number of steps PiperOrigin-RevId: 680990842
References
#24003 - [pallas:mosaic_gpu] Shrink `max_concurrent_iteration` based on the total number of steps
Author
superbobry
Committer
Google-ML-Automation
Parents
a644e23a
Loading