model-upgrade
Skill188 starsChanged 11 days agoApache-2.0View on GitHub
Upgrade existing MetaX models across a confirmed upstream revision range using the independent model implementation, spec and import-audit rules from model-adaptation.
In this repository10 other files it gives its agents.
- AGENTS.mdAGENTS.md
- attention-upgradeSkill
- common-senseSkill
- csrc-upgradeSkill
- environment-checkSkill
- model-adaptationSkill
- patch-upgradeSkill
- registry-upgradeSkill
- vllm-metax-model-trimSkill
- vllm-metax-upgradeSkill
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 26
Only measured GPU-count/VRAM shortages may use a recorded resource skip.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Model Upgrade
Apply common-sense and model-adaptation. Model independence, complete local model-private implementations, concrete-import boundaries, per-model specs and model acceptance are defined there and are not duplicated in this skill.
In a full run, wait for the framework source gate before editing models. Before the framework phases, perform read-only analysis of every model's required framework interfaces, cache layouts and quantization contracts in the shared dependency plan.
For each existing model, compare its prior upstream base and exact target, inspect relevant upstream commits and moved/split/target-only implementations, and account for every required model variant and dependency. Copy the full target implementation, then reapply justified MACA differences using common's preservation rules. Update the model's canonical spec beside its code and rerun the concrete-import audit.
The minimum final validation set includes every model sample currently discovered under
tools/models/; multiple samples of the same architecture are distinct configurations.
Bind each to its spec, source-review task and exact e2e/batched rows with model-bind.
Only measured GPU-count/VRAM shortages may use a recorded resource skip; missing weights,
implementation failures or absent test bindings must not silently count as skips.
For baseline samples using dummy weights, acceptance is explicitly dummy_smoke:
startup, valid prefill/decode generation and concurrent request/batch stability, without
semantic accuracy claims. Real-weight configurations use accuracy and their semantic
cases. Declare configuration.validation_mode in model rows and load_format: dummy
where applicable. A dummy smoke pass cannot satisfy real-weight accuracy acceptance.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

