rewirebio.iobenchmarks
Evidence claim

Relevance of Pretraining GPUs and runtime for short-range models (HyenaDNA Table A.2) to "Select an execution workflow for large-scale diagnostic genomics"

Use case
Select an execution workflow for large-scale diagnostic genomics
Assessed by
Pretraining GPUs and runtime for short-range models (HyenaDNA Table A.2)
Relevance
outside_scope
Endpoint
GPUs, wall-clock time and GPU-hours of pretraining for DNABERT 110M, Nucleotide Transformer 2.5B and HyenaDNA 436K and 1.6M (not shown as evidence)

Outside scope. Pretraining from scratch is spent once by a model's developers and does not measure running a model in a team's workflow. Recorded so these figures are not read as execution cost. The model card's suggested GPUs for inference are kept as a claim (gfm-guide-20261014-claim-hyenadna-suggested-gpus), not a result.

Evidence

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

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.

15 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-10-22-1b7cf027e4f6
Property and statementOriginal source and locationReview and provenance
Citation locations
1 values
  • source id
    gfm-guide-20261014-source-hyenadna-arxiv-v2
    locator
    Table A.2 (page 17)
Context-only references
HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution, arXiv v2

Original source ↗

gfm-guide-20261014-source-hyenadna-arxiv-v2: Table A.2 (page 17)

Version: arXiv 2306.15794v2
Retrieved: 2026-10-14T12:48:26Z

not individually reviewed

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.citation_locators

Source artifact SHA-256: 2f471549352ff571ac55faa80c8da453f9bf1ff5b01b6b87fa893457828e798c

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

Inspected artifact

Constraints
1 values
  • Do not cite these pretraining figures as the cost of running any of these models in a workflow.
Context-only references
HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution, arXiv v2

Original source ↗

gfm-guide-20261014-source-hyenadna-arxiv-v2: Table A.2 (page 17)

Version: arXiv 2306.15794v2
Retrieved: 2026-10-14T12:48:26Z

not individually reviewed

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.constraints

Source artifact SHA-256: 2f471549352ff571ac55faa80c8da453f9bf1ff5b01b6b87fa893457828e798c

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

Inspected artifact

Endpoint
GPUs, wall-clock time and GPU-hours of pretraining for DNABERT 110M, Nucleotide Transformer 2.5B and HyenaDNA 436K and 1.6M (not shown as evidence)
Context-only references
HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution, arXiv v2

Original source ↗

gfm-guide-20261014-source-hyenadna-arxiv-v2: Table A.2 (page 17)

Version: arXiv 2306.15794v2
Retrieved: 2026-10-14T12:48:26Z

not individually reviewed

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.endpoint

Source artifact SHA-256: 2f471549352ff571ac55faa80c8da453f9bf1ff5b01b6b87fa893457828e798c

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

Inspected artifact

Claimed field
links:assessed_by:gfm-guide-20261014-protocol-hyenadna-pretraining-compute
Context-only references
HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution, arXiv v2

Original source ↗

gfm-guide-20261014-source-hyenadna-arxiv-v2: Table A.2 (page 17)

Version: arXiv 2306.15794v2
Retrieved: 2026-10-14T12:48:26Z

not individually reviewed

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.field

Source artifact SHA-256: 2f471549352ff571ac55faa80c8da453f9bf1ff5b01b6b87fa893457828e798c

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

Inspected artifact

Limitations
4 values
  • The DNABERT and Nucleotide Transformer GPU-hours are about 2.5 times the GPUs times wall clock printed in the same table (source evidence concern).
  • The DNABERT and Nucleotide Transformer columns are compiled by the HyenaDNA authors without a cited source.
  • Mixed GPU types (RTX 2080 Ti, A100 80GB, A100 40GB), and the HyenaDNA models are far smaller and pretrain on different data and sequence lengths.
  • Preprint (arXiv v2).
Context-only references
HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution, arXiv v2

Original source ↗

gfm-guide-20261014-source-hyenadna-arxiv-v2: Table A.2 (page 17)

Version: arXiv 2306.15794v2
Retrieved: 2026-10-14T12:48:26Z

not individually reviewed

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.limitations

Source artifact SHA-256: 2f471549352ff571ac55faa80c8da453f9bf1ff5b01b6b87fa893457828e798c

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

Inspected artifact

Pinned records
3 values
  • record id
    gfm-guide-20261014-protocol-hyenadna-pretraining-compute
    fields
    • status
    • name
    • description
    • facets
    • source_ids
    • links
    • attributes
    sha256
    2c3d15f510ddcd6fe719052afebf8fb438bfd6fbba682e5439f7fedf1194a1e1
  • record id
    gfm-guide-20261014-source-hyenadna-arxiv-v2
    fields
    • status
    • attributes.artifact_sha256
    sha256
    6bebd2a77e3d25b4187d8da8524ff0d3938162be25ef131dad9a502a5d7c4b5e
  • record id
    use-case-diagnostic-genomics-model-execution
    fields
    • status
    • name
    • description
    • facets
    • attributes.question
    • attributes.decision
    • attributes.inputs
    • attributes.output
    • attributes.exclusions
    • attributes.setting
    • attributes.clinical_scope
    • attributes.intended_users
    sha256
    1c2a12431556cc6581f5aab019cf91576d2393387b88a8a2bea35ea2a3ee4a70
Context-only references
HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution, arXiv v2

Original source ↗

gfm-guide-20261014-source-hyenadna-arxiv-v2: Table A.2 (page 17)

Version: arXiv 2306.15794v2
Retrieved: 2026-10-14T12:48:26Z

not individually reviewed

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.pins

Source artifact SHA-256: 2f471549352ff571ac55faa80c8da453f9bf1ff5b01b6b87fa893457828e798c

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

Inspected artifact

Rationale
Outside scope. Pretraining from scratch is spent once by a model's developers and does not measure running a model in a team's workflow. Recorded so these figures are not read as execution cost. The model card's suggested GPUs for inference are kept as a claim (gfm-guide-20261014-claim-hyenadna-suggested-gpus), not a result.
Context-only references
HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution, arXiv v2

Original source ↗

gfm-guide-20261014-source-hyenadna-arxiv-v2: Table A.2 (page 17)

Version: arXiv 2306.15794v2
Retrieved: 2026-10-14T12:48:26Z

not individually reviewed

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.rationale

Source artifact SHA-256: 2f471549352ff571ac55faa80c8da453f9bf1ff5b01b6b87fa893457828e798c

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

Inspected artifact

Reason
Add relevance judgements for the execution and compute protocols of the genomic foundation model guide sources batch (data/omics/gfm-guide-sources-20261014, issue #162), independently reviewed in PR #169.
Context-only references
HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution, arXiv v2

Original source ↗

gfm-guide-20261014-source-hyenadna-arxiv-v2: Table A.2 (page 17)

Version: arXiv 2306.15794v2
Retrieved: 2026-10-14T12:48:26Z

not individually reviewed

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.reason

Source artifact SHA-256: 2f471549352ff571ac55faa80c8da453f9bf1ff5b01b6b87fa893457828e798c

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

Inspected artifact

Relevance
outside_scope
Context-only references
HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution, arXiv v2

Original source ↗

gfm-guide-20261014-source-hyenadna-arxiv-v2: Table A.2 (page 17)

Version: arXiv 2306.15794v2
Retrieved: 2026-10-14T12:48:26Z

not individually reviewed

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.relevance

Source artifact SHA-256: 2f471549352ff571ac55faa80c8da453f9bf1ff5b01b6b87fa893457828e798c

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

Inspected artifact

Revision
1
Context-only references
HyenaDNA: Long-Range Genomic Sequence Modeling at Single Nucleotide Resolution, arXiv v2

Original source ↗

gfm-guide-20261014-source-hyenadna-arxiv-v2: Table A.2 (page 17)

Version: arXiv 2306.15794v2
Retrieved: 2026-10-14T12:48:26Z

not individually reviewed

No individual claim review recorded

Source has a recorded evidence concern. Consult its source page before using the claim.

Audit details

Field: attributes.revision

Source artifact SHA-256: 2f471549352ff571ac55faa80c8da453f9bf1ff5b01b6b87fa893457828e798c

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

Inspected artifact

How to read the evidence

Sources and history

Release 2026-10-22-1b7cf027e4f6 · Record review: source checked

1 source record and release historyDownload this release (gzip)
Technical metadata and extraction receipts

Stable ID: use-case-mapping-gfm-guide-20261014-hyenadna-pretraining-compute

field
links:assessed_by:gfm-guide-20261014-protocol-hyenadna-pretraining-compute
value
gfm-guide-20261014-protocol-hyenadna-pretraining-compute
relevance
outside_scope
endpoint
GPUs, wall-clock time and GPU-hours of pretraining for DNABERT 110M, Nucleotide Transformer 2.5B and HyenaDNA 436K and 1.6M (not shown as evidence)
rationale
Outside scope. Pretraining from scratch is spent once by a model's developers and does not measure running a model in a team's workflow. Recorded so these figures are not read as execution cost. The model card's suggested GPUs for inference are kept as a claim (gfm-guide-20261014-claim-hyenadna-suggested-gpus), not a result.
constraints
Do not cite these pretraining figures as the cost of running any of these models in a workflow.
limitations
The DNABERT and Nucleotide Transformer GPU-hours are about 2.5 times the GPUs times wall clock printed in the same table (source evidence concern).; The DNABERT and Nucleotide Transformer columns are compiled by the HyenaDNA authors without a cited source.; Mixed GPU types (RTX 2080 Ti, A100 80GB, A100 40GB), and the HyenaDNA models are far smaller and pretrain on different data and sequence lengths.; Preprint (arXiv v2).
revision
1
reason
Add relevance judgements for the execution and compute protocols of the genomic foundation model guide sources batch (data/omics/gfm-guide-sources-20261014, issue #162), independently reviewed in PR #169.
citation locators
source id: gfm-guide-20261014-source-hyenadna-arxiv-v2; locator: Table A.2 (page 17)
source locator
gfm-guide-20261014-source-hyenadna-arxiv-v2: Table A.2 (page 17)
review
method: source-hash-verification; independent-cell-check; scope-review; reviewer: claude; reviewer note: Claude (Opus 5.5) review agent review-169, independent of the agent that drafted these judgements for issue #162; no human review claimed; reviewed at: 2026-10-14T19:00:00Z; note: Independent relevance review of an outside-scope judgement: pretraining compute does not measure running a model in a team's workflow. Table A.2 re-read in the pinned arXiv v2 PDF (hash matches); the 2.5-fold GPU-hours discrepancy in the limitations is confirmed. No measured evidence is shown. See docs/reviews/use-cases/diagnostic-genomics-model-execution-judgements-2026-10-14.md.
pins
record id: gfm-guide-20261014-protocol-hyenadna-pretraining-compute; fields: status; name; description; facets; source_ids; links; attributes; sha256: 2c3d15f510ddcd6fe719052afebf8fb438bfd6fbba682e5439f7fedf1194a1e1; record id: gfm-guide-20261014-source-hyenadna-arxiv-v2; fields: status; attributes.artifact_sha256; sha256: 6bebd2a77e3d25b4187d8da8524ff0d3938162be25ef131dad9a502a5d7c4b5e; record id: use-case-diagnostic-genomics-model-execution; fields: status; name; description; facets; attributes.question; attributes.decision; attributes.inputs; attributes.output; attributes.exclusions; attributes.setting; attributes.clinical_scope; attributes.intended_users; sha256: 1c2a12431556cc6581f5aab019cf91576d2393387b88a8a2bea35ea2a3ee4a70
Related records

Suggest a correction