npu-torchair-infer
ascend-ai-coding/awesome-ascend-skillsMigrate any HuggingFace model to Ascend NPU torchair graph mode (torch.compile) and benchmark it for accuracy and performance against NPU eager and CPU eager.…
Scores out of 100 · grade B+
2026-08-09Works
40% of the score90/100
- References 1 file that is not in the repo: set_env.sh.
Maintained
25% of the score94/100
- no commits in the last 12 weeks
- no license file
Adopted
20% of the score30/100
- 155 stars on the source repo.
Documented
15% of the score93/100
- 550 words with worked examples.
- Ships 6 bundled files.
Install
npx skills add ascend-ai-coding/awesome-ascend-skills/npu-torchair-inferWhat the check found
| Finding | What it means |
|---|---|
| Missing files | It points at files that are not in the repository, so those steps will fail. |
| No license | The repository ships no license file, so the reuse terms are unclear. |
Files it references that are not in the repository
What it says it does
Migrate any HuggingFace model to Ascend NPU torchair graph mode (torch.compile) and benchmark it for accuracy and performance against NPU eager and CPU eager. Use when running, compiling, or benchmarking HF models (vision, text, image-text encoders such as SigLIP2, DINOv3, ViT, CLIP, Qwen-VL, SAM) on Ascend 910B/CANN with torch_npu and torchair; when a torch.compile graph-mode run on NPU fails (Dynamo TorchRuntimeError, unsupported op, interpolate/contiguous errors); or when comparing torchair vs npu_eager vs cpu with cosine similarity, max abs diff, and p50/p95/p99 latency.
Also in ascend-ai-coding/awesome-ascend-skills
| Artifact | Score | What the check found | Type | Reach | Last commit |
|---|---|---|---|---|---|
| npu-smiascend-ai-coding/awesome-ascend-skills | No license | Skill | 155 stars | today | |
| gitcode-merge-flowascend-ai-coding/awesome-ascend-skills | No license | Skill | 155 stars | today | |
| github-issue-summaryascend-ai-coding/awesome-ascend-skills | No license | Skill | 155 stars | today | |
| npu-docker-launcherascend-ai-coding/awesome-ascend-skills | No license | Skill | 155 stars | today | |
| vllm-ascend-serverascend-ai-coding/awesome-ascend-skills | No license | Skill | 155 stars | today | |
| vllm-daily-pr-issue-trackerascend-ai-coding/awesome-ascend-skills | No license | Skill | 155 stars | today |
Other media skills
Browse all| Artifact | Score | What the check found | Category | Reach | Last commit |
|---|---|---|---|---|---|
| media-useheygen-com/hyperframes | clean | Media | 298,846 installs | today | |
| talking-head-recutheygen-com/hyperframes | clean | Media | 130,756 installs | today | |
| hyperframes-creativeheygen-com/hyperframes | clean | Media | 282,614 installs | today | |
| remotion-to-hyperframesheygen-com/hyperframes | clean | Media | 236,529 installs | today | |
| slideshowheygen-com/hyperframes | clean | Media | 147,126 installs | today | |
| imagecoreyhaines31/marketingskills | clean | Media | 59,761 installs | today |
Put this measurement in your README
A badge carrying how many listings this index holds from the repository and how many pass every static structural check. It reads from this index every time somebody loads your page, so it changes when the measurement changes and there is nothing to keep up to date. Free, no account, and the value is not something you or we can set by hand.
[](https://skillworks.kynth.studio/?q=ascend-ai-coding%2Fawesome-ascend-skills)Would rather not hotlink us? Every badge is also served in shields.io’s endpoint schema, so shields renders the image and your readers never talk to our domain:
Published by Toolproof, the masthead over this index and eight others. The method behind the number is at toolproof.kynth.studio/methodology, and the whole thing is readable as JSON with no key at /api.
