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

Method load_checkpoint

fairseq/trainer.py:134–196  ·  view source on GitHub ↗

Load all training state from a checkpoint file.

(
        self,
        filename,
        reset_optimizer=False,
        reset_lr_scheduler=False,
        optimizer_overrides=None,
        reset_meters=False,
    )

Source from the content-addressed store, hash-verified

source not stored for this graph (policy: none)

Callers 6

mainFunction · 0.80
mainFunction · 0.80
load_checkpointFunction · 0.80

Calls 10

get_modelMethod · 0.95
_build_optimizerMethod · 0.95
set_num_updatesMethod · 0.95
get_num_updatesMethod · 0.95
lr_stepMethod · 0.95
printFunction · 0.85
existsMethod · 0.45
load_state_dictMethod · 0.45
updateMethod · 0.45
resetMethod · 0.45

Tested by 3