rewirebio.iobenchmarks
← Browse all sources
Source

Back to BERT in 2026: ModernGENA as a Strong, Efficient Baseline for DNA Foundation Models

bioRxiv preprint version 1, posted 2026-04-23 · Preprint

DOI: 10.64898/2026.04.21.719816

What this source contributed

Records in this release that cite this source. A record can cite several sources.

Results

Values as the source printed them. Compare results only under compatible protocols, datasets and metrics.

Model or configurationMetricValueDatasetEvaluation
DNABERT-2 on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 256 tokens)68.57 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)DNABERT-2 (117M) inference throughput on one A100 (Aspidova et al. 2026)
DNABERT-2 on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 512 tokens)62.63 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)DNABERT-2 (117M) inference throughput on one A100 (Aspidova et al. 2026)
DNABERT-2 on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 1024 tokens)53.42 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)DNABERT-2 (117M) inference throughput on one A100 (Aspidova et al. 2026)
DNABERT-2 on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 2048 tokens)40.68 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)DNABERT-2 (117M) inference throughput on one A100 (Aspidova et al. 2026)
DNABERT-2 on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 4096 tokens)27.57 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)DNABERT-2 (117M) inference throughput on one A100 (Aspidova et al. 2026)
DNABERT-2 on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 8192 tokens)15.48 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)DNABERT-2 (117M) inference throughput on one A100 (Aspidova et al. 2026)
DNABERT-2 on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 256 tokens)3802 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)DNABERT-2 (117M) inference throughput on one A100 (Aspidova et al. 2026)
DNABERT-2 on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 512 tokens)1433 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)DNABERT-2 (117M) inference throughput on one A100 (Aspidova et al. 2026)
DNABERT-2 on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 1024 tokens)376 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)DNABERT-2 (117M) inference throughput on one A100 (Aspidova et al. 2026)
DNABERT-2 on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 2048 tokens)90 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)DNABERT-2 (117M) inference throughput on one A100 (Aspidova et al. 2026)
DNABERT-2 on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 4096 tokens)15 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)DNABERT-2 (117M) inference throughput on one A100 (Aspidova et al. 2026)
DNABERT-2 on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 8192 tokens)5 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)DNABERT-2 (117M) inference throughput on one A100 (Aspidova et al. 2026)
GENA base on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 256 tokens)91.78 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)GENA base (110M) inference throughput on one A100 (Aspidova et al. 2026)
GENA base on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 512 tokens)81.98 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)GENA base (110M) inference throughput on one A100 (Aspidova et al. 2026)
GENA base on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 256 tokens)6637 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)GENA base (110M) inference throughput on one A100 (Aspidova et al. 2026)
GENA base on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 512 tokens)2206 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)GENA base (110M) inference throughput on one A100 (Aspidova et al. 2026)
GENA large on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 256 tokens)27.13 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)GENA large (336M) inference throughput on one A100 (Aspidova et al. 2026)
GENA large on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 512 tokens)24.79 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)GENA large (336M) inference throughput on one A100 (Aspidova et al. 2026)
GENA large on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 256 tokens)4922 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)GENA large (336M) inference throughput on one A100 (Aspidova et al. 2026)
GENA large on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 512 tokens)1633 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)GENA large (336M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 256 tokens)78.12 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 512 tokens)77.73 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 1024 tokens)77.20 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 2048 tokens)76.38 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 4096 tokens)74.83 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 8192 tokens)71.15 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 256 tokens)3631 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 512 tokens)1811 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 1024 tokens)901 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 2048 tokens)446 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 4096 tokens)218 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 8192 tokens)92 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base (FA2 + autocast bf16) on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 256 tokens)552.44 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (FA2 + autocast bf16) (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base (FA2 + autocast bf16) on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 512 tokens)529.18 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (FA2 + autocast bf16) (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base (FA2 + autocast bf16) on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 1024 tokens)507.61 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (FA2 + autocast bf16) (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base (FA2 + autocast bf16) on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 2048 tokens)479.04 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (FA2 + autocast bf16) (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base (FA2 + autocast bf16) on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 4096 tokens)425.88 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (FA2 + autocast bf16) (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base (FA2 + autocast bf16) on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 8192 tokens)332.54 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (FA2 + autocast bf16) (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base (FA2 + autocast bf16) on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 256 tokens)3631 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (FA2 + autocast bf16) (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base (FA2 + autocast bf16) on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 512 tokens)1811 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (FA2 + autocast bf16) (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base (FA2 + autocast bf16) on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 1024 tokens)901 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (FA2 + autocast bf16) (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base (FA2 + autocast bf16) on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 2048 tokens)446 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (FA2 + autocast bf16) (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base (FA2 + autocast bf16) on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 4096 tokens)218 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (FA2 + autocast bf16) (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA base (FA2 + autocast bf16) on one A100 80 GB (Aspidova et al. 2026)Maximum batch size (Encoder forward pass on one A100 80 GB; sequence length 8192 tokens)92 countFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA base (FA2 + autocast bf16) (135M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA large on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 256 tokens)26.20 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA large (377M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA large on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 512 tokens)26.11 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA large (377M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA large on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 1024 tokens)26.01 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA large (377M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA large on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 2048 tokens)25.85 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA large (377M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA large on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 4096 tokens)25.56 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA large (377M) inference throughput on one A100 (Aspidova et al. 2026)
ModernGENA large on one A100 80 GB (Aspidova et al. 2026)Inference throughput (Encoder forward pass on one A100 80 GB; sequence length 8192 tokens)24.86 sample-per-secondFixed-length token sequences of 256 to 8,192 tokens (Aspidova et al. 2026)ModernGENA large (377M) inference throughput on one A100 (Aspidova et al. 2026)

Claims

Statements this source supports, each with where in the source it was checked.

Models, benchmarks and other records

Models (1)

Configurations (9)

Protocols (1)

Datasets (1)

Evaluations (9)

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.

19 evidence rows matching the loaded filters

Claims, original sources and review scope · Release 2026-10-23-5347695500ed
Property and statementOriginal source and locationReview and provenance
Artifact SHA-256
238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa
Source metadata		Back to BERT in 2026: ModernGENA as a Strong, Efficient Baseline for DNA Foundation Models

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-04-23
Retrieved: 2026-10-11T21:39:38Z

catalogued

No individual claim review recorded

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

Audit details

Field: attributes.artifact_sha256

Source artifact SHA-256: 238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa

Hash scope: bioRxiv stamps each PDF download, so artifact_sha256 describes one retrieval; retrieved_at is the file's modification time. The pdftotext -layout text layer has SHA-256 ebaf0541c173359940737efde2c15d4ab4aa37d7cc323b3922a818feb5be6b4b. A second download at 21:43Z was refused with HTTP 429.

Inspected artifact

Artifact url
https://www.biorxiv.org/content/10.64898/2026.04.21.719816v1.full.pdf
Source metadata		Back to BERT in 2026: ModernGENA as a Strong, Efficient Baseline for DNA Foundation Models

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-04-23
Retrieved: 2026-10-11T21:39:38Z

catalogued

No individual claim review recorded

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

Audit details

Field: attributes.artifact_url

Source artifact SHA-256: 238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa

Hash scope: bioRxiv stamps each PDF download, so artifact_sha256 describes one retrieval; retrieved_at is the file's modification time. The pdftotext -layout text layer has SHA-256 ebaf0541c173359940737efde2c15d4ab4aa37d7cc323b3922a818feb5be6b4b. A second download at 21:43Z was refused with HTTP 429.

Inspected artifact

Doi
10.64898/2026.04.21.719816
Source metadata		Back to BERT in 2026: ModernGENA as a Strong, Efficient Baseline for DNA Foundation Models

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-04-23
Retrieved: 2026-10-11T21:39:38Z

catalogued

No individual claim review recorded

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

Audit details

Field: attributes.doi

Source artifact SHA-256: 238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa

Hash scope: bioRxiv stamps each PDF download, so artifact_sha256 describes one retrieval; retrieved_at is the file's modification time. The pdftotext -layout text layer has SHA-256 ebaf0541c173359940737efde2c15d4ab4aa37d7cc323b3922a818feb5be6b4b. A second download at 21:43Z was refused with HTTP 429.

Inspected artifact

Evidence concerns
1 values
  • source id
    model-execution-20261011-source-aspidova2026-moderngena
    message
    Appendix B Table 2 and Figure 1 (axis 'Inference speed (tokens/s)') give throughput in tokens/s, but the printed values are inconsistent with the printed batch sizes: ModernGENA base (FA2 + autocast bf16) at L = 256 prints 552.44 tokens/s with a batch of 3,631 sequences (929,536 tokens), about 28 minutes per batch, and ModernGENA base without FA2 at L = 8192 prints 71.15 tokens/s with a batch of 92 (753,664 tokens), about 2.9 hours per batch, against a stated pretraining throughput of 1.33M tokens/s on eight A100 GPUs (Section 3.2.4). The unit or scale of the throughput column is unclear. Batch sizes share the table. Values are recorded as printed.
    source locator
    Appendix B Table 2 (page 9); Figure 1 (page 4); Section 3.2.4
    artifact sha256
    238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa
    reviewed at
    2026-10-12T07:50:32Z
    review method
    automated-source-review
Source metadata		Back to BERT in 2026: ModernGENA as a Strong, Efficient Baseline for DNA Foundation Models

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-04-23
Retrieved: 2026-10-11T21:39:38Z

catalogued

No individual claim review recorded

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

Audit details

Field: attributes.evidence_concerns

Source artifact SHA-256: 238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa

Hash scope: bioRxiv stamps each PDF download, so artifact_sha256 describes one retrieval; retrieved_at is the file's modification time. The pdftotext -layout text layer has SHA-256 ebaf0541c173359940737efde2c15d4ab4aa37d7cc323b3922a818feb5be6b4b. A second download at 21:43Z was refused with HTTP 429.

Inspected artifact

Hash scope
bioRxiv stamps each PDF download, so artifact_sha256 describes one retrieval; retrieved_at is the file's modification time. The pdftotext -layout text layer has SHA-256 ebaf0541c173359940737efde2c15d4ab4aa37d7cc323b3922a818feb5be6b4b. A second download at 21:43Z was refused with HTTP 429.
Source metadata		Back to BERT in 2026: ModernGENA as a Strong, Efficient Baseline for DNA Foundation Models

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-04-23
Retrieved: 2026-10-11T21:39:38Z

catalogued

No individual claim review recorded

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

Audit details

Field: attributes.hash_scope

Source artifact SHA-256: 238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa

Hash scope: bioRxiv stamps each PDF download, so artifact_sha256 describes one retrieval; retrieved_at is the file's modification time. The pdftotext -layout text layer has SHA-256 ebaf0541c173359940737efde2c15d4ab4aa37d7cc323b3922a818feb5be6b4b. A second download at 21:43Z was refused with HTTP 429.

Inspected artifact

Licence
CC-BY-4.0
Source metadata		Back to BERT in 2026: ModernGENA as a Strong, Efficient Baseline for DNA Foundation Models

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-04-23
Retrieved: 2026-10-11T21:39:38Z

catalogued

No individual claim review recorded

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

Audit details

Field: attributes.licence

Source artifact SHA-256: 238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa

Hash scope: bioRxiv stamps each PDF download, so artifact_sha256 describes one retrieval; retrieved_at is the file's modification time. The pdftotext -layout text layer has SHA-256 ebaf0541c173359940737efde2c15d4ab4aa37d7cc323b3922a818feb5be6b4b. A second download at 21:43Z was refused with HTTP 429.

Inspected artifact

Media type
application/pdf
Source metadata		Back to BERT in 2026: ModernGENA as a Strong, Efficient Baseline for DNA Foundation Models

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-04-23
Retrieved: 2026-10-11T21:39:38Z

catalogued

No individual claim review recorded

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

Audit details

Field: attributes.media_type

Source artifact SHA-256: 238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa

Hash scope: bioRxiv stamps each PDF download, so artifact_sha256 describes one retrieval; retrieved_at is the file's modification time. The pdftotext -layout text layer has SHA-256 ebaf0541c173359940737efde2c15d4ab4aa37d7cc323b3922a818feb5be6b4b. A second download at 21:43Z was refused with HTTP 429.

Inspected artifact

Publication status
preprint
Source metadata		Back to BERT in 2026: ModernGENA as a Strong, Efficient Baseline for DNA Foundation Models

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-04-23
Retrieved: 2026-10-11T21:39:38Z

catalogued

No individual claim review recorded

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

Audit details

Field: attributes.publication_status

Source artifact SHA-256: 238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa

Hash scope: bioRxiv stamps each PDF download, so artifact_sha256 describes one retrieval; retrieved_at is the file's modification time. The pdftotext -layout text layer has SHA-256 ebaf0541c173359940737efde2c15d4ab4aa37d7cc323b3922a818feb5be6b4b. A second download at 21:43Z was refused with HTTP 429.

Inspected artifact

Retrieved at
2026-10-11T21:39:38Z
Source metadata		Back to BERT in 2026: ModernGENA as a Strong, Efficient Baseline for DNA Foundation Models

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-04-23
Retrieved: 2026-10-11T21:39:38Z

catalogued

No individual claim review recorded

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

Audit details

Field: attributes.retrieved_at

Source artifact SHA-256: 238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa

Hash scope: bioRxiv stamps each PDF download, so artifact_sha256 describes one retrieval; retrieved_at is the file's modification time. The pdftotext -layout text layer has SHA-256 ebaf0541c173359940737efde2c15d4ab4aa37d7cc323b3922a818feb5be6b4b. A second download at 21:43Z was refused with HTTP 429.

Inspected artifact

Url
https://doi.org/10.64898/2026.04.21.719816
Source metadata		Back to BERT in 2026: ModernGENA as a Strong, Efficient Baseline for DNA Foundation Models

Original source ↗

No field-specific location recorded

Version: bioRxiv preprint version 1, posted 2026-04-23
Retrieved: 2026-10-11T21:39:38Z

catalogued

No individual claim review recorded

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

Audit details

Field: attributes.url

Source artifact SHA-256: 238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa

Hash scope: bioRxiv stamps each PDF download, so artifact_sha256 describes one retrieval; retrieved_at is the file's modification time. The pdftotext -layout text layer has SHA-256 ebaf0541c173359940737efde2c15d4ab4aa37d7cc323b3922a818feb5be6b4b. A second download at 21:43Z was refused with HTTP 429.

Inspected artifact

How to read the evidence

Sources and history

Release 2026-10-23-5347695500ed · Record review: source checked

Preprint retrieved and hashed for the diagnostic genomics execution literature pass (2026-10-11).

Release history

No supporting source is linked yet.

Read original source

Download this release (gzip)
Technical metadata and extraction receipts

Stable ID: model-execution-20261011-source-aspidova2026-moderngena

areas
dna-genomes
contexts
research
url
https://doi.org/10.64898/2026.04.21.719816
artifact url
https://www.biorxiv.org/content/10.64898/2026.04.21.719816v1.full.pdf
version
bioRxiv preprint version 1, posted 2026-04-23
retrieved at
2026-10-11T21:39:38Z
artifact sha256
238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa
doi
10.64898/2026.04.21.719816
publication status
preprint
licence
CC-BY-4.0
media type
application/pdf
hash scope
bioRxiv stamps each PDF download, so artifact_sha256 describes one retrieval; retrieved_at is the file's modification time. The pdftotext -layout text layer has SHA-256 ebaf0541c173359940737efde2c15d4ab4aa37d7cc323b3922a818feb5be6b4b. A second download at 21:43Z was refused with HTTP 429.
evidence concerns
source id: model-execution-20261011-source-aspidova2026-moderngena; message: Appendix B Table 2 and Figure 1 (axis 'Inference speed (tokens/s)') give throughput in tokens/s, but the printed values are inconsistent with the printed batch sizes: ModernGENA base (FA2 + autocast bf16) at L = 256 prints 552.44 tokens/s with a batch of 3,631 sequences (929,536 tokens), about 28 minutes per batch, and ModernGENA base without FA2 at L = 8192 prints 71.15 tokens/s with a batch of 92 (753,664 tokens), about 2.9 hours per batch, against a stated pretraining throughput of 1.33M tokens/s on eight A100 GPUs (Section 3.2.4). The unit or scale of the throughput column is unclear. Batch sizes share the table. Values are recorded as printed.; source locator: Appendix B Table 2 (page 9); Figure 1 (page 4); Section 3.2.4; artifact sha256: 238ab9916d54b419bf31fd79ecf146d654eb6574d3b441aa2043d39632e709fa; reviewed at: 2026-10-12T07:50:32Z; review method: automated-source-review
Related records

    Suggest a correction