model-management
Manage CellTypist built-in model discovery, local cache behavior, model pickle inspection, conversion, subsetting, markers, and offline test fixtures.
- License 许可
- MIT
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- model-management
- license
- MIT
- metadata
- { "disco-role": "operating" }
- description
- Manage CellTypist built-in model discovery, local cache behavior, model pickle inspection, conversion, subsetting, markers, and offline test fixtures.
- disable-model-invocation
- true
02 资源文件 5
- 参考文档references/conversion-and-subsetting.md6.3 KB
- 参考文档references/model-reference.md6.6 KB
- 参考文档references/troubleshooting.md8.0 KB
- 脚本scripts/model_cache_check.py9.3 KB
- 脚本scripts/tiny_model_factory.py6.4 KB
03 SKILL.md 正文
CellTypist Model Management
Use this sub-skill when a task is about locating, downloading, loading, writing, inspecting, converting, subsetting, or smoke-testing CellTypist models.
Route Quickly
- Read model reference for cache paths, remote downloads,
Model.load,Model.write, prediction labels/probabilities, and marker extraction. - Read conversion and subsetting before changing species, gene identifiers, or retained cell types; these methods mutate the loaded model in place.
- Use troubleshooting for cache/network surprises, invalid pickle files, missing map files, conversion column mistakes, invalid collapse modes, and subset constraints.
- Run model_cache_check.py to inspect the active
CELLTYPIST_FOLDERmodel cache or verify a local model pickle without downloading anything. - Run tiny_model_factory.py to create a tiny offline-compatible CellTypist model pickle for smoke tests and examples.
Stay In Scope
- Stay here for
celltypist.models, built-in model inventory/download/cache behavior,Modelobject persistence and metadata, marker extraction,Model.convert, andModel.subset. - Route annotation with a selected model to annotation-workflows.
- Route training a new real reference model to training-and-custom-models.
- Route result tables, AnnData insertion, UMAPs, dotplots, or interpretation of finished predictions to visualization-and-results.
Offline-First Pattern
For offline work, avoid implicit default-model resolution and pass an explicit local model path containing a forward slash, such as ./local_model.pkl or Path("local_model.pkl").resolve().as_posix().
from pathlib import Path
from celltypist import models
model_path = Path("local_model.pkl").resolve().as_posix()
model = models.Model.load(model_path)
print(model.cell_types)
print(model.features[:5])
If a task may run without network access, inspect cache state first:
python scripts/model_cache_check.py --verify-model ./local_model.pkl
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).