Support, citation, and license
Get help
Search or open an issue in the PoreFM issue tracker. Include the command, error traceback, PoreFM version or Git revision, model release, sequencing chemistry, GPU, and PyTorch/CUDA/FlashAttention versions. For input problems, describe whether you used BAM + POD5 or an SG-tagged BAM, and whether the reads were aligned and contained move tables.
A small reproducible example is helpful. Review logs and run_manifest.json
before sharing them: manifests can contain local paths, hostnames, and input
configuration snapshots.
Cite PoreFM
Cite PoreFM as software using CITATION.cff. In a methods section, report:
- The software version and Git revision.
- The model repository and release tag, and checkpoint checksum when available.
- Sequencing chemistry, target selection, read filters, and relevant scoring settings.
- How read-level results were summarized, including coverage and decision thresholds.
Keep run_manifest.json with your results. Model downloads record the resolved
release and checksum there. Cite the source datasets and reference assays
separately; the HeLa example includes its
GLORI and BACS references.
License
PoreFM code is distributed under the MIT license. The vendored Oxford Nanopore k-mer resources retain their upstream license and copyright notices. Consult each model or dataset repository for the terms that apply to those separately distributed files.