Making large AI models cheaper, faster and more accessible
[autochunk] refactor chunk memory estimation (#2762)
* refact memory code * dont log free var memory * add memory align * update chunk target * update setting for new memory * finish test * update tracer * update typo * update test
X
Xuanlei Zhao committed
2ca9728cbb0863e27d74c7318358e339f9450d3e
Parent: b51bfec
Committed by GitHub <noreply@github.com>
on 3/8/2023, 8:22:30 AM