Peak memory (Embedding extraction; peak GPU memory allocated) of GenomeOcean-500M, FP8 (E4M3) static weights, PyTorch with FBGEMM (Yu et al. 2026) on GTDB genome fragments for metagenomic binning, 20 of a pre-sampled pool of 50 bacterial genomes per trial (Yu et al. 2026)
46.7 Peak memory (Embedding extraction; peak GPU memory allocated)
Methods
- Tested configuration
- GenomeOcean-500M, FP8 (E4M3) static weights, PyTorch with FBGEMM (Yu et al. 2026)
- Protocol
- GenomeOcean embedding extraction throughput, peak VRAM and binning fidelity, BF16 versus FP8 weights (Yu et al. 2026 Table 3)
- Dataset
- GTDB genome fragments for metagenomic binning, 20 of a pre-sampled pool of 50 bacterial genomes per trial (Yu et al. 2026)
- Procedure
- GenomeOcean embedding extraction throughput, peak VRAM and binning fidelity, BF16 versus FP8 weights (Yu et al. 2026 Table 3)
- Evaluation
- GenomeOcean-500M FP8 (E4M3) embedding extraction (Yu et al. 2026)
- Coverage
- Not reported scored / Not reported eligible
- Uncertainty
- Not reported by the source
- Evidence
- Author-reported evaluation · source checkedFP8 Inference in Genomic Foundation Models: Theoretical vs. Realized Speedups on GenomeOcean · Table 3, row Model '500M', Precision 'FP8 (E4M3)', column 'Peak VRAM (GB)'
A source-checked result verifies the numerical transcription, not every model or protocol detail. Evaluation metadata: source checked. Source checked does not mean independently reproduced.
Reproduction
- Split
- Five trials, seeds 42-46
- Adaptation
- None (pretrained checkpoint, post-training quantization only)
- Scoring implementation
- Not reported
No execution recipe has been verified for this exact configuration and evaluation. A benchmark's general instructions may use different inputs, splits or model settings.
Reproducing this published result requires matching its model configuration, data, split and scorer. Source checking or a successful smoke test does not establish score reproduction.
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.
1 evidence row matching the loaded filters
| Property and statement | Original source and location | Review and provenance |
|---|---|---|
| Reported result 46.66 Individual claims | FP8 Inference in Genomic Foundation Models: Theoretical vs. Realized Speedups on GenomeOcean Table 3, row Model '500M', Precision 'FP8 (E4M3)', column 'Peak VRAM (GB)' Version: bioRxiv preprint version 1, posted 2026-08-14 (PDF footer); bioRxiv API date 2026-08-18 | source checked ["source-hash-verification","pdf-text-parse","rendered-source-visual-check","independent-cell-check"] · 2026-10-12T07:50:32Z author reported Audit detailsIndependent review 2026-10-12: printed value, numeric value, metric, qualifier, unit, direction, uncertainty, locator and configuration and protocol identity match the source. Transcribed, not reproduced. Field: Source artifact SHA-256: Hash scope: bioRxiv stamps each PDF download, so artifact_sha256 describes one retrieval (a second download differed in 40 bytes). The pdftotext -layout text layer (poppler) of both downloads has SHA-256 bec6e695ce46c4ebd6692e01a1a93167e212dffbb97b6558f1da88b1eb453501. Extraction artifact SHA-256: |
Sources and history
Release 2026-10-22-1b7cf027e4f6 · Record review: source checked
1 source record and release history
- FP8 Inference in Genomic Foundation Models: Theoretical vs. Realized Speedups on GenomeOcean · Original source · bioRxiv preprint version 1, posted 2026-08-14 (PDF footer); bioRxiv API date 2026-08-18
Technical metadata and extraction receipts
Stable ID: model-execution-20261011-result-yu2026-genomeocean-500m-fp8-weights-embedding-peak-vram
- metric
- peak-memory
- metric direction
- lower
- unit
- gigabyte
- metric qualifier
- Embedding extraction; peak GPU memory allocated
- unit detail
- GB as printed; torch.cuda.max_memory_allocated()
- printed value
- 46.66
- numeric value
- 46.66
- source locator
- Table 3, row Model '500M', Precision 'FP8 (E4M3)', column 'Peak VRAM (GB)'
- missing metadata
- uncertainty: reason: unreported
- review
- method: source-hash-verification; pdf-text-parse; rendered-source-visual-check; independent-cell-check; method note: Archived PDF matched the pinned SHA-256; a fresh bioRxiv download had the same text-layer SHA-256. Tables 3 and 4 parsed from pdftotext -layout with a separate review script, including every ± value, and every cell read against a 170 dpi render of page 8.; reviewer: claude; reviewer note: Claude (Opus 5.5) review agent reviewer-exec, independent of the extractor uc-execution; no human review claimed; reviewed at: 2026-10-12T07:50:32Z; artifact sha256: ca54817d1f1a308f09bd3022e56cf6f8b86a875b110da0a38fe31236e39ca076; retrieval url: https://www.biorxiv.org/content/10.64898/2026.08.09.743676v1.full.pdf; note: Independent review 2026-10-12: printed value, numeric value, metric, qualifier, unit, direction, uncertainty, locator and configuration and protocol identity match the source. Transcribed, not reproduced.