Dataset Viewer
The dataset could not be loaded because the splits use different data file formats, which is not supported. Read more about the splits configuration. Click for more details.
Couldn't infer the same data file format for all splits. Got {NamedSplit('train'): (None, {}), NamedSplit('test'): ('hdf5', {})}
Error code:   FileFormatMismatchBetweenSplitsError

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.

Nirvana42 Virtual Cell Challenge — Checkpoints v1

Persistent storage for trained model checkpoints and pipeline artifacts. Avoids retraining when working with fresh ephemeral pods.

Contents

Models

  • models/baseline_st_10k/best.ckpt — Baseline ST (ESM2-only, 10K steps, 8xH200)
  • models/baseline_st_10k/final.ckpt — Baseline ST final checkpoint
  • models/baseline_st_10k/config.yaml — Training config

Perturbation Table Artifacts

MPEN = Manifold Perterbation Encoding Network

  • artifacts/perturbation_tables/fused_pert_table_esm2_mpen_5120d.pt — Fused ESM2+MPEN (5120D)
  • artifacts/perturbation_tables/concat_esm2_mpen_10240d.pt — Concatenated ESM2||MPEN (10240D, for DualPath)
  • artifacts/mpen/candidate_pert_features_mpen.pt — MPEN regulatory features (5120D, Evo2-40B L17)
  • artifacts/esm2/ESM2_pert_features_with_control.pt — ESM2 protein features with non-targeting control

Source

  • Base model: ST-SE-Tahoe (Arc Institute)
  • Training: 8x NVIDIA H200, DDP, batch_size=4/GPU

Evaluation

Downloads last month
19