Alex Cheema
0870e6bfb0
Squashed 'tinychat/' content from commit fa7e734b4
|
há 1 ano atrás | |
|---|---|---|
| .. | ||
| training_submission_v4.0 | há 1 ano atrás | |
| README | há 1 ano atrás | |
| dataloader.py | há 1 ano atrás | |
| helpers.py | há 1 ano atrás | |
| initializers.py | há 1 ano atrás | |
| losses.py | há 1 ano atrás | |
| lr_schedulers.py | há 1 ano atrás | |
| metrics.py | há 1 ano atrás | |
| model_eval.py | há 1 ano atrás | |
| model_spec.py | há 1 ano atrás | |
| model_train.py | há 1 ano atrás | |
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
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