Initial model weights for DiffDock-Pocket RL v1.0.0, a reinforcement learning fine-tuned version of DiffDock-Pocket trained using the RLDiff repository with PoseBusters validity as the reward signal. The model weights are attached as DD_Pocket_RL_score_model.pt.