data-media-ml-and-vision
Route scraping, media conversion, plotting, vision, and ML demo tasks.
- License 许可
- MIT
- 最近更新
- 2026-09-03
- Stars
- ★ 351
- Forks
- 27
01 前言 Front Matter
- name
- data-media-ml-and-vision
- license
- MIT
- metadata
- { "disco-role": "operating" }
- description
- Route scraping, media conversion, plotting, vision, and ML demo tasks.
- disable-model-invocation
- true
02 资源文件 4
- 参考文档references/data-media-recipes.md7.5 KB
- 参考文档references/ml-vision-dependency-map.md6.5 KB
- 参考文档references/troubleshooting.md4.9 KB
- 脚本scripts/check_heavy_project_requirements.py25.3 KB
03 SKILL.md 正文
data-media-ml-and-vision
Use this sub-skill for python-mini-project folders whose main job is one of these:
- scraping or tabular collection
- PDF, image, audio, QR, or ASCII conversion
- notebook-driven plotting or ML demos
- OpenCV, MediaPipe, TensorFlow, Keras, Ultralytics, Folium, Plotly, or NLTK work
- Windows-only Excel COM conversion
Representative folders:
- Web scraping for book names
- Web Scraping IPhone from Flipkart
- Slideshare to PDF
- Demerge_pdfs
- Image_compressor
- Simple_Image_resize
- images_to_pdf_converter
- image_comparator
- img_to_ascii
- QR Code Generator
- qr_with_logo
- Download Audio
- Youtube_video_download
- TextToSpeech
- Speech_To_Text
- Speaking_Dictionary
- Speaking_Wikipedia
- Face_Recognition
- Finding_Lanes
- Motion_Detection
- Object_Detection
- Shape_Recognition
- Geo_Plot_Using_Folium
- Plotter
- Regression using ANN
- digit-recognizer
- mnist_digit_recognizer
- Text_Predication
- xls_to_xlsx
Route to web-network-and-automation instead when the main task is deploying or operating a Flask/FastAPI service, handling credentials or secrets, or maintaining long-lived API/server behavior.
Route to cli-algorithms-and-utilities when the project is really a pure stdlib helper with no media, scraping, notebook, or ML dependency.
Working rules:
- Stay inside the current checkout.
- Read the project README, requirements, and notebook files as static evidence.
- Run scripts/check_heavy_project_requirements.py before changing heavy folders.
- Assume network, camera, microphone, speakers, GPU, model download, notebook kernel, and Windows COM dependencies are optional unless the task explicitly asks for them.
- Do not execute project code by default. Use a safe sample file or fixture only when a live check is explicitly requested and the environment is ready.
What to produce:
- a short task summary
- a dependency/backend risk note
- a safe next-step plan or handoff to another sub-skill when the boundary is crossed
See:
- references/data-media-recipes.md
- references/ml-vision-dependency-map.md
- references/troubleshooting.md
- scripts/check_heavy_project_requirements.py
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).