Heimdall
importedsoftware/heimdall
Exploring how tokenization shapes the performance and generalization of single-cell foundation models.
Machine-generated from the listed sources and not yet reviewed by a human.
- Category
- Software & Systems
- Subcategory
- unknown
- License
- MIT(osi)
- Status
- active
- Maturity
- deployed
- Organization
- ma-compbio-lab
- Country
- unknown
- Repository
- github.com/ma-compbio-lab/Heimdall
- Documentation
- unknown
- Tags
- foundation-models · foundation-models-for-biology · representation-learning · single-cell · tokenization · transcriptomics
- Regulatory
- unknown
Computed from shared tags, weighted so a rare tag counts for more than a common one. These are suggestions, not curated relationships.
- PertEvalfoundation-models · single-cell · transcriptomics
Evaluation suite for transcriptomic perturbation effect prediction models. Includes support for single-cell foundation models.
- scConceptfoundation-models · single-cell · transcriptomics
Contrastive pre-training for technology-agnostic single-cell representations beyond reconstruction
- MEMEfoundation-models-for-biology
[npj Digital Medicine 2025] Multiple Embedding Model for EHR (MEME) used for strong prediction on Emergency Department tasks
- SynthSleepNetfoundation-models-for-biology
[IEEE TCYB] Toward Foundational Model for Sleep Analysis Using a Multimodal Hybrid Self-Supervised Learning Framework
- foundation-cancer-image-biomarkerfoundation-models · representation-learning
[Nature Machine Intelligence 2024] Code and evaluation repository for the paper
- GENEBfoundation-models · representation-learning
GENEB: ICML 2026 benchmark for genomic foundation models across 100 tasks and 13 functional categories.
- api.github.com/repos/ma-compbio-lab/Heimdallretrieved 2026-08-25 · via github-api
Machine-imported from GitHub search. Last push 2026-04-16, 20 stars, license reported as MIT. Category and schematic were assigned by keyword heuristics and are unreviewed.
Not yet verified by a human. Correct this record →
/v1/entries/55.json→ .entries["heimdall"]
Entries are sharded 64 ways by a stable hash of the id, so a consumer can find any record without an index.