Raise a ValueError if using the weights with the bugged plddt

This commit is contained in:
Tuscan Thompson
2025-02-26 15:54:26 -08:00
committed by Rohith Krishna
parent a13984aa67
commit 80fbf29175

View File

@@ -237,6 +237,12 @@ class EvaluateAF3:
# Load the config
self.config = OmegaConf.create(checkpoint["training_config"])
# Make sure we aren't using the version with a bug in plddt
if self.config.experiment.name == "rf2aa-af3-repro-rollout_nmw_from_scratch_af3_style_no_cb_normal_crop_cont_3":
raise ValueError(
"These weights are outdated and the plddt metric may be inaccurate. Please update to the latest available weights."
)
# Sampler sets diffusion batch size based on the following, not strictly on batch size in vaildation transform
self.config.dataset_params["diffusion_batch_size_valid"] = diffusion_batch_size
self.config.af3_inference["num_steps"] = num_steps