PR #686 Compute Loss inside the training step.

Compute Loss inside the training step. #686

JingyaHuang merged 18 commits into huggingface:main from AdamLouly:adamlouly/compute_loss_training

improved solutio

69c6f111

compute loss fix

c9180089

AdamLouly changed the title ~~Adamlouly/compute loss training~~ Compute Loss inside the training step. 3 years ago

resolve conflict

40fd135c

pengwa commented on 2023-02-13

pengwa commented on 2023-02-14

JingyaHuang added training

JingyaHuang assigned

JingyaHuang 3 years ago

JingyaHuang added onnxruntime

esolved comments

7e968107

removed duplicated code .. used main trainer compute loss

4732f2c2

added --loss_in_train flag

c47fa80f

resolve conflict

040d77ea

JingyaHuang added gpu-test

JingyaHuang removed gpu-test

pengwa commented on 2023-02-23

resolved comments

c45bd537

pengwa approved these changes on 2023-02-24

pengwa commented on 2023-02-26

JingyaHuang removed training

JingyaHuang commented on 2023-02-28

resolved comments

31178c8e

formatter usng latest black

19cfe045

add import for code quality

f268040d

formatter usng latest black

4d8624a8

pengwa commented on 2023-03-03

readding super loss compute

ee6ef104

pengwa commented on 2023-03-06

resolv comments

dc8de715

fix typo

55ad1d27

solve not exporting onnx models

432efe54

dictionary casting , bind method

2b5e57b4

trainer fix with ruff

b6ccb53a

JingyaHuang approved these changes on 2023-03-24

JingyaHuang merged 4f79ebda into main 3 years ago

Reviewers

JingyaHuang

pengwa

askhade

Assignees

JingyaHuang

Labels

onnxruntime

Milestone

No milestone

optimum Compute Loss inside the training step. #686 Merged

Compute Loss inside the training step. #686

optimum
Compute Loss inside the training step.
#686

Merged