MCPcopy Create free account
hub / github.com/HA-Transformer/MAT / lr_step_update

Method lr_step_update

fairseq/trainer.py:468–470  ·  view source on GitHub ↗

Update the learning rate after each update.

(self)

Source from the content-addressed store, hash-verified

source not stored for this graph (policy: none)

Callers 2

lr_stepMethod · 0.95
set_num_updatesMethod · 0.95

Calls 2

get_num_updatesMethod · 0.95
step_updateMethod · 0.45

Tested by

no test coverage detected