Agent skill · nvidia
kermt-continue-pretrain
Continue KERMT pretraining on a custom SMILES corpus with a grover_base, cmim, or hybrid checkpoint. Use a local checkpoint or optionally download a pinned Hugging Face model bundle using HF_TOKEN if configured. Run containerized training and write model bundles, prepared data, logs, and checkpoints to user-selected host directories.
What it needs
About 9k tokens when loaded.
What this skill does
kermt-continue-pretrain Continue pretraining from a user-supplied KERMT checkpoint (groverbase / cmim / hybrid). The skill is the workflow orchestrator: it validates inputs, prepares the corpus, launches the runner, and returns a run directory. Skill and runtime paths Set SKILLDIR to the absolute path of this installed skill directory. Export KERMTREPO as the absolute path to the KERMT checkout used for model execution. The bundled container helper mounts that checkout at /workspace and this skill at /skill (read-only). Commands inside the container use /skill/scripts/; defaults are bundled in config/. See Released models for checkpoint bundle requirements. Downloads and local outputs The optional released-model branch reads config/releasedmodel.json for the Hugging Face repository, pinned revision, and filenames. The bundled scripts/fetchreleasedmodel.py downloads the model bundle over HTTPS into the host directory the user selects. Public models work without credentials; if HFTOKEN is set, the container helper forwards it for Hugging Face authentication. Prepared data, logs, and workflow results go into the chosen run directory. Hardware requirements GPUs: 1–N CUDA-capable NVIDIA GPUs. The runner auto-detects via torch.cuda.devicecount(); --gpus 0,2 overrides. On a single GPU the runner falls back to --batchsize 32 --saveinterval 500; on multi-GPU it uses the defaultspretrain.json values (currently batchsize 256). Note: --gpus N uses torch.cuda indexing, which can differ from nvidia-smi's display order on multi-GPU hosts (PCI bus vs. CUDA enumeration). To target a specific physical GPU, set CUDAVISIBLEDEVICES before invoking, or run python -c "import torch; print([torch.cuda.getdevicename(i) for i in range(torch.cuda.devicecount())])" to confirm which device you're picking. VRAM: the default --batch-size 256 is sized for A100-class hardware (80 GB VRAM). …
How to use it
Reference it in AdaL, Claude Code, Cursor or any coding agent — nothing to install:
@skills nvidia/bionemo-kermt-continue-pretrain