 Artem Chumachenko
|
43e029e248
test&style
|
3 lat temu |
|
Artem Chumachenko
|
bc295e6278
add load_state in power_sgd
|
3 lat temu |
|
Artem Chumachenko
|
9b070e0a5a
Merge branch 'nasty_bug_with_extra_tensors' of github.com:learning-at-home/hivemind into power_ef_new
|
3 lat temu |
 justheuristic
|
53cf4b6b8b
black
|
3 lat temu |
|
Artem Chumachenko
|
5244d8a96f
add a unit test that fails (but shouldn't)
|
3 lat temu |
|
Artem Chumachenko
|
4ddc635245
isort
|
3 lat temu |
|
Artem Chumachenko
|
54223ccf93
fix tests
|
3 lat temu |
|
Artem Chumachenko
|
68a7d022f0
black style
|
3 lat temu |
|
Artem Chumachenko
|
2e763be1de
first try
|
3 lat temu |
 Max Ryabinin
|
b1507687ae
Use hivemind.Optimizer in examples/albert (#426)
|
3 lat temu |
|
justheuristic
|
5d31c3ba9d
Implement core functionality of hivemind.Optimizer (#403)
|
3 lat temu |
|
justheuristic
|
d883387e37
Add ProgressTracker (#408)
|
3 lat temu |
|
justheuristic
|
02eee9292d
[hivemind.Optimizer] TrainingStateAverager (#407)
|
3 lat temu |
|
justheuristic
|
8fa0a8e6ae
Add GradientAverager with support for delayed averaging (#404)
|
3 lat temu |
