rewire.itbenchmarks
← Back to results
Benchmark

PerturBench

PerturBench evaluates predicted single-cell perturbation responses with explicit aggregation and metric choices.

Sourcesaltoslabs/perturbench official source · Pinned README: Data; Model evaluation; custom dataset configuration

76 evaluations · 76 results

Overview

Datasets

Processed AnnData datasets with perturbation/covariate metadata and configurable feature selections.

Metrics

Expression/change aggregation precedes metrics such as cosine, Pearson, RMSE, MSE, MAE and R-squared; optional rank metrics form another view.

Allowed inputs

Predicted/observed expression and perturbation/covariate metadata in AnnData.

Sourcesaltoslabs/perturbench official source · Pinned README: Data; Model evaluation; custom dataset configuration
Evaluation procedure diagram
How it worksEvaluation procedure
Evaluation procedure1. Allowed inputs: Predicted/observed expression and perturbation/covariate metadata in AnnData.. Then: 2. Splits: Cross-cell-type and combination-prediction splits are supported, along with explicit custom split files.. Then: 3. Metrics: Expression/change aggregation precedes metrics such as cosine, Pearson, RMSE, MSE, MAE and R-squared; optional rank metrics form another view.Evaluation procedure1. Allowed inputs: Predicted/observed expression and perturbation/covariate metadata in AnnData.. Then: 2. Splits: Cross-cell-type and combination-prediction splits are supported, along with explicit custom split files.. Then: 3. Metrics: Expression/change aggregation precedes metrics such as cosine, Pearson, RMSE, MSE, MAE and R-squared; optional rank metrics form another view.Evaluation procedure1. Allowed inputs: Predicted/observed expression and perturbation/covariate metadata in AnnData.. Then: 2. Splits: Cross-cell-type and combination-prediction splits are supported, along with explicit custom split files.. Then: 3. Metrics: Expression/change aggregation precedes metrics such as cosine, Pearson, RMSE, MSE, MAE and R-squared; optional rank metrics form another view.

Conceptual procedure. Task variants and protocol versions retain their separate scoring conditions.

Sourcesaltoslabs/perturbench official source · Pinned README: Data; Model evaluation; custom dataset configuration

Source reviewed · Automated source review, 2026-09-16. All specifications and missing details

Results

Each comparison retains its reviewed evaluation scope, dataset and metric. Results are shown without a pooled ranking.

PerturBench CB-COSINE: combination prediction on Norman19, Cosine similarity of log fold change

cosine_logfc (fraction) · Higher values are better.

PerturBench CB-COSINE: combination prediction on Norman19, Cosine similarity of log fold change · Norman19 (PerturBench split)

Evidence origin: Author-reported evaluation.

PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation Analysis · Table 3, column(Cosine, log fold change (LogFC))
  • RMSE and both rank metrics are better when lower. The rank metrics measure how often another perturbation's prediction is closer than the right one's.
  • The two experiments use different datasets and splits, so their figures are not comparable to each other.
Comparison details and limitations

Every method PerturBench reports on combination prediction on Norman19, Cosine similarity of log fold change, scored with Cosine similarity of log fold change on Norman19.

  • Author-reported numbers, source checked but not independently reproduced.

Automated source review: 2026-09-18. Numerical source review does not establish independent reproduction.

Dots show point estimates. Whiskers show only explicitly defined uncertainty (standard deviation, standard error or a labelled interval); their definitions remain in Table. Unresolved uncertainty is not plotted. Differences do not establish statistical significance.

Showing 9 of 9 matching rows.

Tested configuration
00.250.50.751
Reported score
  1. LA0.79 ± 1 × 10 − 2
  2. LA (scGPT)0.77 ± 4 × 10 − 3
  3. Decoder0.73 ± 2 × 10 − 2
  4. CPA* (scGPT)0.70 ± 1 × 10 − 2
  5. Linear0.60 ± 2 × 10 − 2
  6. CPA* (noAdv)0.55 ± 9 × 10 − 2
  7. CPA*0.52 ± 6 × 10 − 2
  8. Biolord*0.41 ± 2 × 10 − 2
  9. SAMS-VAE*0.17 ± 2 × 10 − 3

Methods and evaluation design

Procedure, tasks and evaluated configurations

How it works

Evaluation methodology

PerturBench predicts single-cell responses in held-out perturbation–context combinations. It implements cross-covariate, combinatorial and inverse-combinatorial partitions, comparing learned models with simple controls. Rank-based metrics complement expression-error metrics to reveal models that fail to distinguish perturbations.

Sourcesperturbench primary benchmark evidence · Experimental setup; Appendix datasets and data splitting

Evaluation design

Benchmarks bring together tasks and protocols. A task describes the biological question; a protocol defines a particular test.

Tasks

These source-backed links do not make different protocols or scores interchangeable.

Baseline coverage

Reference methods help show what a model adds beyond simple controls. We track a null control and a conventional method for each protocol.

No concrete protocols are explicitly linked to this suite. Protocol identification and baseline selection are outstanding.

Protocol coverage CSV · Model evaluation matrix · Source table · Release and checksums

Coverage is derived from release 2026-09-29-06401fd5b220. Source citations describe the original records; they do not validate an unreviewed baseline proposal. No results have been generated by this audit.

Run this benchmark

Train and evaluate a perturbation model

Install the environment, load a benchmark dataset with its split, and train a model through the repository's configuration system.

Generate predictions and evaluate them. This recipe does not establish reproduction of a particular published score.

Dataset access
Downloaded by the repository's dataset accessors.
Model and weights
None required; models are trained from the data.
Licences
Project licence: see repository. Upstream data licences are separate and unreported here.
Software
Python with the repository's conda environment and hydra configs.
Hardware
Not stated in the cited section. Several of these steps expect a GPU.
Required inputs and expected outputs

Inputs

  • A perturbation response model implemented against the repository's base class.

Outputs

  • Trained model checkpoints and the evaluation the pipeline runs.

Execution steps

  1. 1. Install (Command line)

    Source reviewed; these instructions have not been executed by rewire.

    conda create -n [env-name] python=3.11
conda activate [env-name]
cd [/path/to/PerturBench/]
pip3 install -e .
# or
pip3 install -e .[cli]
    PerturBench: repository README · README.md at c84038bc, Install PerturBench, lines 17-22

  2. 2. Load a dataset (Python)

    Source reviewed; these instructions have not been executed by rewire.

    from perturbench.data.accessors.srivatsan20 import Sciplex3

srivatsan20_accessor = Sciplex3()
adata = srivatsan20_accessor.get_anndata() ## Get the preprocessed anndata object
torch_dataset = srivatsan20_accessor.get_dataset() ## Get a PyTorch Dataset
    PerturBench: repository README · README.md at c84038bc, Dataset Access, lines 35-39

  3. 3. Apply the benchmark split (Python)

    Source reviewed; these instructions have not been executed by rewire.

    from perturbench.data.accessors.jiang24 import Jiang24

jiang24_accessor = Jiang24()
split = jiang24_accessor.get_split()
    PerturBench: repository README · README.md at c84038bc, Data Splitting, lines 49-52

  4. 4. Train a model (Command line)

    Source reviewed; these instructions have not been executed by rewire.

    python <path-to-repo-folder>/src/perturbench/modelcore/train.py <config-options>
    PerturBench: repository README · README.md at c84038bc, Hydra Training Script, lines 66-66

Use your own model

Run your model locally and return predictions keyed by the input IDs. The evaluator supplies biological inputs without test labels and owns scoring. This interface is not a sandbox for model code.

Pass your existing prediction function into this adapter. Its output direction must match the selected protocol.

class MyModelAdapter:
    def __init__(self, score):
        self.score = score

    def predict(self, inputs):
        return {row["id"]: float(self.score(row)) for row in inputs}

# adapter = MyModelAdapter(your_prediction_function)
# report = rewirebench.run(prepared, adapter, output="runs/my-model")

Alternatively, generate a keyed prediction file in your existing model environment and use the score-only recipe. Your model code and weights do not need to be shared.

PerturBench: repository README · README.md at c84038bc
Scope and limitations
  • Quoted from the project's README and not executed by rewire, so the commands are evidence of what the project documents rather than a verified run.
  • The project may have changed since the pinned commit.
  • RMSE and both rank metrics on this page are better when lower.
  • The two experiments use different datasets and are not comparable to each other.

Contribute a result for review. The library can submit an exported evaluation for private review when intake is open. Check the contribution page for access and sign-in.

Original repository instructions

Run this benchmark

Official installation, processed-data download, dataset accessors and Hydra/evaluator instructions are available. Local cache paths and task-specific split files must be configured; several datasets have manual split-generation notebooks. A default invocation must not be presented as reproducing every study split.

A maintained rewire runner has not been verified for this benchmark. Check data access, weights, licences, dependencies and hardware in the linked official documentation; requirements have not been fully extracted.

altoslabs/perturbench / README.md · README.md lines 15–63 (Install, datasets, splits and usage)
Strengths, limitations and unresolved questions

Strengths and limitations

Strengths supported by sources

  • Separating response aggregation from scoring makes a major source of metric disagreement explicit.
    Sourcesaltoslabs/perturbench official source · Pinned README: Data; Model evaluation; custom dataset configuration

Limitations and conditions

  • A covariate-transfer split can expose a perturbation in another cell context during training; it should not be called completely unseen-perturbation prediction. Report the exact split and perturbation inputs.
    Sourcesperturbench primary benchmark evidence · Experimental setup; Appendix datasets and data splitting
Profile review details

Primary paper and/or task implementation reviewed for the explicitly cited methodology claims. Scope-limited absence is recorded only after the documented source search; no model runs or independent reproduction.

Stable record: discovery-benchmark-perturbench

Specifications

Inputs, training, access and other details

Explanatory profile: source reviewed · Automated source review, 2026-09-16. Review applies to the cited claims; unresolved fields are listed below. Numerical results retain their own review status.

Data, procedure and scoring
PropertyDescription and evidence
DatasetsProcessed AnnData datasets with perturbation/covariate metadata and configurable feature selections.
Sourcesaltoslabs/perturbench official source · Pinned README: Data; Model evaluation; custom dataset configuration
SplitsCross-cell-type and combination-prediction splits are supported, along with explicit custom split files.
Sourcesaltoslabs/perturbench official source · Pinned README: Data; Model evaluation; custom dataset configuration
MetricsExpression/change aggregation precedes metrics such as cosine, Pearson, RMSE, MSE, MAE and R-squared; optional rank metrics form another view.
Sourcesaltoslabs/perturbench official source · Pinned README: Data; Model evaluation; custom dataset configuration
BaselinesReproduction configurations include linear reference models.
Sourcesaltoslabs/perturbench official source · Pinned README: Data; Model evaluation; custom dataset configuration
Leakage controlsSplit and covariate definitions are configuration inputs; exact evaluation files must be pinned.
Sourcesaltoslabs/perturbench official source · Pinned README: Data; Model evaluation; custom dataset configuration
UncertaintyFor the best hyperparameter configuration the authors run four additional training seeds, yielding five runs. Error bars represent standard deviation of model performance across those runs.
Sourcesperturbench primary benchmark evidence · Experimental setup; Appendix datasets and data splitting
Entity typePerturbation-response evaluation framework.
Sourcesaltoslabs/perturbench official source · Pinned README: Data; Model evaluation; custom dataset configuration
OrganismsThe evaluated datasets use human cell-line perturbation systems, including McFaline-Figueroa’s glioblastoma cell contexts and Srivatsan’s chemical perturbation cell lines. The framework itself is not restricted to those organisms.
Sourcesperturbench primary benchmark evidence · Experimental setup; Appendix datasets and data splitting
AssaysSingle-cell perturbation response measurements.
Sourcesaltoslabs/perturbench official source · Pinned README: Data; Model evaluation; custom dataset configuration
Allowed inputsPredicted/observed expression and perturbation/covariate metadata in AnnData.
Sourcesaltoslabs/perturbench official source · Pinned README: Data; Model evaluation; custom dataset configuration
AdaptationSupports supervised response prediction with explicit cell-type and combination holdouts.
Sourcesaltoslabs/perturbench official source · Pinned README: Data; Model evaluation; custom dataset configuration
Applicable tests and references

Applicability is distinct from a completed evaluation.

Evidence

Source checking verifies the cited claim or transcription. It does not establish independent reproduction.

Papers and result coverage

Last literature check: 2026-09-17. Dated primary-source discovery and protocol/table screening. Source checking does not mean experimental reproduction. Only separately extracted and independently reviewed numeric batches are publishable.

Paper or primary resourceVersionReference
PerturBench: Benchmarking Machine Learning Models for Cellular Perturbation AnalysisPrimary full-text snapshot retrieved 2026-09-17; exact bytes pinned by SHA-256Read source
Historical gaps recorded on 2026-09-17

The catalogue now holds 76 result rows for this benchmark. A note below about pending extraction describes the state on 2026-09-17 and may since have been answered by a later batch. The result rows and their sources are the current record.

  • complete comparable numeric result batch: Specific candidate tables and protocol boundaries are documented; no graph values or incomplete winner-only selection are converted into publishable rows.
Search and extraction details

primary protocol reviewed

Searches

  • PerturBench benchmarking perturbation models 2408.10609

Evidence locations

  • v1 Tables 2–4; Appendix E protocols

Evidence table

Inspect claims, sources and review details

Trace each statement to its source and review. A context-only reference supports the record generally; it does not verify an individual field. Source checking does not reproduce an experiment.

One row per statement and cited source. Multiple citations are not independent evaluations. Shared locators are labelled explicitly.

27 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-09-29-06401fd5b220
Property and statementOriginal source and locationReview and provenance
Diagram caption
Conceptual procedure. Task variants and protocol versions retain their separate scoring conditions.
Individual claims		altoslabs/perturbench official source

Original source ↗

Pinned README: Data; Model evaluation; custom dataset configuration

Version: c84038bc1ea409aa54f3832cfa6f34f5059adf0c
Retrieved: 2026-09-16T10:30:22.733246+00:00

source checked

automated source review · 2026-09-16

Audit details

Primary paper and/or task implementation reviewed for the explicitly cited methodology claims. Scope-limited absence is recorded only after the documented source search; no model runs or independent reproduction.

Field: attributes.profile.diagram.caption

Source artifact SHA-256: 4aa29fa1d4333a72013fd2c60545218f0015b85b95b76878cbdb88962ff4e55d

Hash scope: Hash scope not separately documented; inspect source record

Diagram steps
  • Allowed inputs: Predicted/observed expression and perturbation/covariate metadata in AnnData.
  • Splits: Cross-cell-type and combination-prediction splits are supported, along with explicit custom split files.
  • Metrics: Expression/change aggregation precedes metrics such as cosine, Pearson, RMSE, MSE, MAE and R-squared; optional rank metrics form another view.
Individual claims		altoslabs/perturbench official source

Original source ↗

Pinned README: Data; Model evaluation; custom dataset configuration

Version: c84038bc1ea409aa54f3832cfa6f34f5059adf0c
Retrieved: 2026-09-16T10:30:22.733246+00:00

source checked

automated source review · 2026-09-16

Audit details

Primary paper and/or task implementation reviewed for the explicitly cited methodology claims. Scope-limited absence is recorded only after the documented source search; no model runs or independent reproduction.

Field: attributes.profile.diagram.steps

Source artifact SHA-256: 4aa29fa1d4333a72013fd2c60545218f0015b85b95b76878cbdb88962ff4e55d

Hash scope: Hash scope not separately documented; inspect source record

Diagram title
Evaluation procedure
Individual claims		altoslabs/perturbench official source

Original source ↗

Pinned README: Data; Model evaluation; custom dataset configuration

Version: c84038bc1ea409aa54f3832cfa6f34f5059adf0c
Retrieved: 2026-09-16T10:30:22.733246+00:00

source checked

automated source review · 2026-09-16

Audit details

Primary paper and/or task implementation reviewed for the explicitly cited methodology claims. Scope-limited absence is recorded only after the documented source search; no model runs or independent reproduction.

Field: attributes.profile.diagram.title

Source artifact SHA-256: 4aa29fa1d4333a72013fd2c60545218f0015b85b95b76878cbdb88962ff4e55d

Hash scope: Hash scope not separately documented; inspect source record

Datasets
Processed AnnData datasets with perturbation/covariate metadata and configurable feature selections.
Individual claims		altoslabs/perturbench official source

Original source ↗

Pinned README: Data; Model evaluation; custom dataset configuration

Version: c84038bc1ea409aa54f3832cfa6f34f5059adf0c
Retrieved: 2026-09-16T10:30:22.733246+00:00

source checked

automated source review · 2026-09-16

Audit details

Primary paper and/or task implementation reviewed for the explicitly cited methodology claims. Scope-limited absence is recorded only after the documented source search; no model runs or independent reproduction.

Field: attributes.profile.facts.0.value

Source artifact SHA-256: 4aa29fa1d4333a72013fd2c60545218f0015b85b95b76878cbdb88962ff4e55d

Hash scope: Hash scope not separately documented; inspect source record

Splits
Cross-cell-type and combination-prediction splits are supported, along with explicit custom split files.
Individual claims		altoslabs/perturbench official source

Original source ↗

Pinned README: Data; Model evaluation; custom dataset configuration

Version: c84038bc1ea409aa54f3832cfa6f34f5059adf0c
Retrieved: 2026-09-16T10:30:22.733246+00:00

source checked

automated source review · 2026-09-16

Audit details

Primary paper and/or task implementation reviewed for the explicitly cited methodology claims. Scope-limited absence is recorded only after the documented source search; no model runs or independent reproduction.

Field: attributes.profile.facts.1.value

Source artifact SHA-256: 4aa29fa1d4333a72013fd2c60545218f0015b85b95b76878cbdb88962ff4e55d

Hash scope: Hash scope not separately documented; inspect source record

Adaptation
Supports supervised response prediction with explicit cell-type and combination holdouts.
Individual claims		altoslabs/perturbench official source

Original source ↗

Pinned README: Data; Model evaluation; custom dataset configuration

Version: c84038bc1ea409aa54f3832cfa6f34f5059adf0c
Retrieved: 2026-09-16T10:30:22.733246+00:00

source checked

automated source review · 2026-09-16

Audit details

Primary paper and/or task implementation reviewed for the explicitly cited methodology claims. Scope-limited absence is recorded only after the documented source search; no model runs or independent reproduction.

Field: attributes.profile.facts.10.value

Source artifact SHA-256: 4aa29fa1d4333a72013fd2c60545218f0015b85b95b76878cbdb88962ff4e55d

Hash scope: Hash scope not separately documented; inspect source record

Metrics
Expression/change aggregation precedes metrics such as cosine, Pearson, RMSE, MSE, MAE and R-squared; optional rank metrics form another view.
Individual claims		altoslabs/perturbench official source

Original source ↗

Pinned README: Data; Model evaluation; custom dataset configuration

Version: c84038bc1ea409aa54f3832cfa6f34f5059adf0c
Retrieved: 2026-09-16T10:30:22.733246+00:00

source checked

automated source review · 2026-09-16

Audit details

Primary paper and/or task implementation reviewed for the explicitly cited methodology claims. Scope-limited absence is recorded only after the documented source search; no model runs or independent reproduction.

Field: attributes.profile.facts.2.value

Source artifact SHA-256: 4aa29fa1d4333a72013fd2c60545218f0015b85b95b76878cbdb88962ff4e55d

Hash scope: Hash scope not separately documented; inspect source record

Baselines
Reproduction configurations include linear reference models.
Individual claims		altoslabs/perturbench official source

Original source ↗

Pinned README: Data; Model evaluation; custom dataset configuration

Version: c84038bc1ea409aa54f3832cfa6f34f5059adf0c
Retrieved: 2026-09-16T10:30:22.733246+00:00

source checked

automated source review · 2026-09-16

Audit details

Primary paper and/or task implementation reviewed for the explicitly cited methodology claims. Scope-limited absence is recorded only after the documented source search; no model runs or independent reproduction.

Field: attributes.profile.facts.3.value

Source artifact SHA-256: 4aa29fa1d4333a72013fd2c60545218f0015b85b95b76878cbdb88962ff4e55d

Hash scope: Hash scope not separately documented; inspect source record

Leakage controls
Split and covariate definitions are configuration inputs; exact evaluation files must be pinned.
Individual claims		altoslabs/perturbench official source

Original source ↗

Pinned README: Data; Model evaluation; custom dataset configuration

Version: c84038bc1ea409aa54f3832cfa6f34f5059adf0c
Retrieved: 2026-09-16T10:30:22.733246+00:00

source checked

automated source review · 2026-09-16

Audit details

Primary paper and/or task implementation reviewed for the explicitly cited methodology claims. Scope-limited absence is recorded only after the documented source search; no model runs or independent reproduction.

Field: attributes.profile.facts.4.value

Source artifact SHA-256: 4aa29fa1d4333a72013fd2c60545218f0015b85b95b76878cbdb88962ff4e55d

Hash scope: Hash scope not separately documented; inspect source record

Uncertainty
For the best hyperparameter configuration the authors run four additional training seeds, yielding five runs. Error bars represent standard deviation of model performance across those runs.
Individual claims		perturbench primary benchmark evidence

Original source ↗

Experimental setup; Appendix datasets and data splitting

Version: 2408.10609v1
Retrieved: 2026-09-16T21:07:13.159027+00:00

source checked

automated source review · 2026-09-16

Audit details

Primary paper and/or task implementation reviewed for the explicitly cited methodology claims. Scope-limited absence is recorded only after the documented source search; no model runs or independent reproduction.

Field: attributes.profile.facts.5.value

Source artifact SHA-256: 5c4804565dd9faa17a11853a79e9847dcb6da73715b4c91f62e5b274cc79f186

Hash scope: Hash scope not separately documented; inspect source record

Inspected artifact

All evidence (CSV)All evidence (JSONL)How to read the evidence

Sources and history

View linked audit checks and correction history

Release 2026-09-29-06401fd5b220 · Record review: discovered

5 source records and release historyDownload this release
Technical metadata and extraction receipts

Stable ID: discovery-benchmark-perturbench

areas
single-cell
entity level
suite
scope note
Specialist molecular or omics evaluation; protocol details require review before numerical comparison.
task
Predicting cellular perturbation responses
version
Not reported
benchmark research
review date: 2026-09-17; status: primary_protocol_reviewed; primary sources: evidence-expansion-perturbench-5c480456; inspected locators: v1 Tables 2–4; Appendix E protocols; searched queries: PerturBench benchmarking perturbation models 2408.10609; gaps: complete comparable numeric result batch: Specific candidate tables and protocol boundaries are documented; no graph values or incomplete winner-only selection are converted into publishable rows.; claim scope: Dated primary-source discovery and protocol/table screening. Source checking does not mean experimental reproduction. Only separately extracted and independently reviewed numeric batches are publishable.
historical missing metadata
dataset release: unextracted; metric implementation: unextracted; split manifest: unextracted; version: unextracted
metadata review scope
historical_missing_metadata preserves the original discovery state. Current descriptive evidence and missingness are recorded in profile.facts; numerical-result review is separate.
entity classification
review date: 2026-09-17; rationale: The cited profile describes a collection of evaluation tasks or protocols; retain it as the top-level benchmark suite. Its datasets and individual protocols remain separate records.; source ids: src-discovery-altoslabs-perturbench; source locator: Pinned README: Data; Model evaluation; custom dataset configuration; ambiguities: None recorded
run documentation
record id: discovery-benchmark-perturbench; source ids: run-doc-perturbench-readme-md-c84038bc; status: official_documentation_linked; summary: Official installation, processed-data download, dataset accessors and Hydra/evaluator instructions are available. Local cache paths and task-specific split files must be configured; several datasets have manual split-generation notebooks. A default invocation must not be presented as reproducing every study split.; source locator: README.md lines 15–63 (Install, datasets, splits and usage)
run recipes
id: perturbench-official; protocol id: discovery-benchmark-perturbench; version: c84038bc1ea409aa54f3832cfa6f34f5059adf0c; title: Train and evaluate a perturbation model; purpose: generate_and_evaluate; summary: Install the environment, load a benchmark dataset with its split, and train a model through the repository's configuration system.; inputs: A perturbation response model implemented against the repository's base class.; outputs: Trained model checkpoints and the evaluation the pipeline runs.; requirements: data: Downloaded by the repository's dataset accessors.; weights: None required; models are trained from the data.; licence: Project licence: see repository. Upstream data licences are separate and unreported here.; software: Python with the repository's conda environment and hydra configs.; hardware: Not stated in the cited section. Several of these steps expect a GPU.; instructions: runtime: command_line; title: Install; code: conda create -n [env-name] python=3.11 conda activate [env-name] cd [/path/to/PerturBench/] pip3 install -e . # or pip3 install -e .[cli]; status: source_reviewed_not_executed; source ids: project-recipe-perturbench-c84038bc; source locator: README.md at c84038bc, Install PerturBench, lines 17-22; runtime: python; title: Load a dataset; code: from perturbench.data.accessors.srivatsan20 import Sciplex3 srivatsan20_accessor = Sciplex3() adata = srivatsan20_accessor.get_anndata() ## Get the preprocessed anndata object torch_dataset = srivatsan20_accessor.get_dataset() ## Get a PyTorch Dataset; status: source_reviewed_not_executed; source ids: project-recipe-perturbench-c84038bc; source locator: README.md at c84038bc, Dataset Access, lines 35-39; runtime: python; title: Apply the benchmark split; code: from perturbench.data.accessors.jiang24 import Jiang24 jiang24_accessor = Jiang24() split = jiang24_accessor.get_split(); status: source_reviewed_not_executed; source ids: project-recipe-perturbench-c84038bc; source locator: README.md at c84038bc, Data Splitting, lines 49-52; runtime: command_line; title: Train a model; code: python <path-to-repo-folder>/src/perturbench/modelcore/train.py <config-options>; status: source_reviewed_not_executed; source ids: project-recipe-perturbench-c84038bc; source locator: README.md at c84038bc, Hydra Training Script, lines 66-66; limitations: Quoted from the project's README and not executed by rewire, so the commands are evidence of what the project documents rather than a verified run.; The project may have changed since the pinned commit.; RMSE and both rank metrics on this page are better when lower.; The two experiments use different datasets and are not comparable to each other.; source ids: project-recipe-perturbench-c84038bc; source locator: README.md at c84038bc
Related records

Suggest a correction