Files
foundry/configs/trainer/af3.yaml
Nathaniel Corley 5a492032d5 refactor: new modelhub (#109)
* Initial commit of chiral changes

Initial checkin of chiral feature code

Add chiral metric

* Update the way chiral features are incorporated into the model

Move initialization to new func

use default pytorch reset parameters

fix initialization for chirals

config

rename argument of confidence head

fix initialization for chirals

* refactor: src nest, rename rf2aa to modelhub

* refactor: initial commit without projects

* Initial commit of chiral changes

* Initial checkin of chiral feature code

* Add chiral metric

* Remove option for double residual connection.  Add kq_norm oiptions to base (20250125) config.

* Restoring flag

* config

* rename argument of confidence head

* Update the way chiral features are incorporated into the model

* config

* rename argument of confidence head

* Update the way chiral features are incorporated into the model

* Initial commit of chiral changes

Initial checkin of chiral feature code

Add chiral metric

* Update the way chiral features are incorporated into the model

Move initialization to new func

use default pytorch reset parameters

fix initialization for chirals

config

rename argument of confidence head

fix initialization for chirals

* refactor: new modelhub

---------

Co-authored-by: fdimaio <dimaio@uw.edu>
Co-authored-by: HaotianZhangAI4Science <haotianzhang@zju.edu.cn>
2025-04-08 13:33:17 -07:00

21 lines
574 B
YAML

defaults:
- ddp
- loss: structure_prediction
- metrics: structure_prediction
_target_: modelhub.trainers.af3.AF3Trainer
validate_every_n_epochs: 1
max_epochs: 10_000
n_examples_per_epoch: 24000
prevalidate: True
# We must pre-specify the number of recycles during training so we can pre-sample recycles per batch consistently for each GPU
n_recycles_train: ${datasets.n_recycles_train}
clip_grad_max_norm: 10.0
output_dir: ${paths.output_dir}
checkpoint_every_n_epochs: 1
# precision: bf16-mixed # Mixed precision training with bfloat16 (currently does not work)