tinygrad

geohot/tinygrad

Fork 0

forked from tinygrad/tinygrad

Files

T

History

chenyuandGitHub 4bd5535d72 update mlperf resnet default hparams (#3758 )

we might be able to have higher lr given smaller BS, but this is good.

Trained to 75.9%
https://wandb.ai/chenyuxyz/tinygrad-examples_mlperf/runs/xi2f48se/overview

2024-03-15 12:09:26 -04:00

dataloader.py

MLPerf Resnet (cleaned up) (#3573 )

2024-03-14 00:53:41 -04:00

helpers.py

MLPerf Resnet (cleaned up) (#3573 )

2024-03-14 00:53:41 -04:00

initializers.py

MLPerf Resnet (cleaned up) (#3573 )

2024-03-14 00:53:41 -04:00

lr_schedulers.py

PolynomialDecayWithWarmup + tests (#3649 )

2024-03-07 18:53:36 -05:00

metrics.py

Add MLPerf UNet3D model (#775 )

2023-05-28 20:38:19 -07:00

model_eval.py

MLPerf Resnet (cleaned up) (#3573 )

2024-03-14 00:53:41 -04:00

model_spec.py

move globalcounters to ops (#2960 )

2024-01-01 14:21:02 -08:00

model_train.py

update mlperf resnet default hparams (#3758 )

2024-03-15 12:09:26 -04:00

optimizers.py

lars optimizer + tests (#3631 )

2024-03-06 18:11:01 -05:00

README

start on mlperf models

2023-05-10 16:30:49 -07:00

README

Each model should be a clean single file.
They are imported from the top level `models` directory

It should be capable of loading weights from the reference imp.

We will focus on these 5 models:

# Resnet50-v1.5 (classic) -- 8.2 GOPS/input
# Retinanet
# 3D UNET (upconvs)
# RNNT
# BERT-large (transformer)

They are used in both the training and inference benchmark:
https://mlcommons.org/en/training-normal-21/
https://mlcommons.org/en/inference-edge-30/
And we will submit to both.

NOTE: we are Edge since we don't have ECC RAM