inference-cli
Run OmegaFold CLI inference from FASTA to PDB safely, including devices, weights, resources, dry-run validation, and troubleshooting.
- License 许可
- Apache-2.0
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- inference-cli
- license
- Apache 2.0
- metadata
- { "disco-role": "operating" }
- description
- Run OmegaFold CLI inference from FASTA to PDB safely, including devices, weights, resources, dry-run validation, and troubleshooting.
- disable-model-invocation
- true
02 资源文件 4
- 参考文档references/cli-reference.md4.2 KB
- 参考文档references/inference-workflow.md4.2 KB
- 参考文档references/troubleshooting.md4.7 KB
- 脚本scripts/omegafold_cli_smoke.py7.8 KB
03 SKILL.md 正文
OmegaFold inference CLI
Use this sub-skill when a task needs to run or validate OmegaFold command-line inference from a FASTA file to PDB outputs. Prefer the installed omegafold console script; use the installed-package fallback python -m omegafold only when the console script is unavailable but the package import works.
Fast path
- Validate the CLI without downloading model weights:
python scripts/omegafold_cli_smoke.py - Inspect a FASTA and print a planned command without running inference:
python scripts/omegafold_cli_smoke.py --fasta INPUT_FILE.fasta --output-dir OUTPUT_DIRECTORY --model 2 --device cuda - Run full inference only when weights, device, memory, and runtime are acceptable:
omegafold INPUT_FILE.fasta OUTPUT_DIRECTORY --model 1 --device cuda
Route by need
- Use CLI reference for flags, defaults, model choices, command templates, and installed-module fallback syntax.
- Use inference workflow for safe validation, weights/cache decisions, device/resource planning, and output expectations.
- Use troubleshooting for install/import errors, Python and Torch compatibility, missing weights, device mismatch, OOM, and output mistakes.
- Use scripts/omegafold_cli_smoke.py for no-download CLI help checks and FASTA dry-run planning.
- For FASTA normalization details, output file naming, PDB confidence fields, and validation of generated structures, route to data-and-outputs.
- For direct Python use of
OmegaFold,make_config,fasta2inputs, and tensor contracts, route to model-api.
Safety defaults
omegafold --helpis safe: it prints parser help and does not load weights.omegafold INPUT_FILE.fasta OUTPUT_DIRECTORY ...is not a smoke test: argument parsing selects a model, may download weights, then loads a largetorchcheckpoint before inference.- Avoid full inference unless the user has approved network/download cost or has provided a local
--weights_file, and the selected--devicehas enough memory.
04 同仓库更多 Skills
passnet-triton-opt
Make a single PassNet Triton kernel fast AND numerically faithful: performance model (when a replacement can win at all), block/grid/warp tuning, autotune policy, launch overhead, and per-op numeric recipes to pass the dtype baseline tolerances. Use when a pass matches and is correct but speedup below expected, or when correctness fails by small numeric margins.
passnet-solve
END-TO-END playbook for solving one PassNet sample: analyze the computation graphs, decide the optimization strategy (what to fuse, what to replace, what to leave alone), drive the iteration loop, and maximize the sample score. This is the ENTRY skill — invoke it first for any PassNet optimization task; it tells you when to use passnet-pattern-fusion, passnet-triton-opt and passnet-feedback.
passnet-skill
PassNet GPU kernel optimization via compiler passes. Design and implement Triton-based optimization passes, create pass files under ./pass_dir/, self-evaluate with pass_evaluator, and iterate to maximize GPU speedup.
passnet-pattern-fusion
Author PassNet pass files: write patterns that actually MATCH the FX graph, pick fusion regions (which ops to absorb into one Triton kernel), and structure multi-pass submissions with the mandatory shared-dispatch architecture. Use when creating or fixing pattern()/replacement_args()/replacement_func() files, when a pass "failed to match", or when deciding how to fuse multiple kernels/ops.
passnet-orchestrate
Round/batch planning and decision support for PassNet graph optimization. Use for multi-sample triage, worker allocation, eval-budget policy, and post-eval keep/revert decisions. For solving one concrete sample end to end, use passnet-solve as the entry skill; consult this skill only when you need broader planning or a specific decision gate.
passnet-feedback
Fast, deterministic feedback for PassNet work: pre-flight pattern/match verification WITHOUT burning a GPU evaluation, per-node bottleneck analysis of a sample's graphs, and eval-log parsing into per-variant status + estimated score + failure classification. Use BEFORE every GPU evaluation (check_pattern), at the START of a sample (analyze_graph), and AFTER every evaluation (parse_eval_log).