kaggle

Name: kaggle
Author: Borda

23stars3forks

How to install kaggle?

borda/ai-rig/kaggle

$npx -y skills add Borda/AI-Rig --skill kaggle

Installs into the current project.

Summary

Generate a Kaggle competition notebook as a Jupytext # %% Python script following the user's established ML research style: PTL for DNN training, best-fit tool selection, EDA→Baseline→Train→Inference pipeline with per-stage lens cells, small single-purpose cells each carrying a

Files · 1

View on GitHub

SKILL.md

1	<objective>
2
3	Generate Kaggle competition notebook script, Jupytext `# %%` format.
4
5	Two goals, equal weight — neither traded for other:
6	- Win — leaderboard-competitive: leakage-safe CV, metric-aligned loss/model choice, tuning/ensembling when it moves the score, not style theater
7	- Teach — read top to bottom like a university/seminar lecture on solving this competition: reader new to it follows the full reasoning chain, every decision motivated, nothing left as unexplained code
8
9	Follows user's ML research style distilled from past notebooks:
10	- PTL always for DNN training (PyTorch Lightning + torchmetrics) — even simple baselines
11	- Tool agnostic — best-fit library for problem; PTL when training loop needed
12	- Stages with lenses — each major stage: quick sanity check cell (show one batch, print shapes, verify submission format)
13	- Small, single-purpose cells — one action per cell (load, one transform, one plot, one check); never bundle setup + run + verify to save cell count
14	- Every cell earns its place — one-line why (comment or markdown sentence) before/in each cell: the specific reason this step happens now — never a restatement of what the code does
15	- Section markdown is extensive and structured — full explanation of what/why/how-it-advances-the-goal per section, formatted as tables/lists/blockquotes over dense prose paragraphs; markdown before a plot sets up the question, markdown after states the finding and its implication — plot and prose flow as one beat, never an orphaned chart
16	- `# !` bash over subprocess — package installs, `nvidia-smi`, `ls -lh`, `# ! head submission.csv`
17	- EDA is visual — distribution plots, sample grids, dimension scatters before any model
18	- Inference included — model save pattern + separate load-and-infer cells
19	- CSVLogger + seaborn — metrics plotted from `metrics.csv` after every training run
20
21	NOT for writing Python packages, modules, production code — notebook scripts only.
22	NOT research literature survey — use `/research:topic` for SOTA literature search.
23
24	</objective>
25
26	<inputs>
27
28	- $ARGUMENTS: one of:
29	- `<competition-name>` — short slug for output filename; generates blank template
30	- `<competition-name> <url>` — fetches competition overview from URL before generating
31	- `<competition-name> "<description>"` — inline description of problem and data
32	- `--type <type>` — hint: `classification`, `regression`, `segmentation`, `detection`, `tabular` (auto-detected when omitted)
33	- `--eda-only` — generate only EDA sections (no model/training/submission); always online (no offline setup)
34	- `--inference-only` — generate inference notebook from checkpoint (no EDA, no training); always offline (frozen packages pattern); loads checkpoint from `PATH_CHECKPOINT` constant; output suffix `-inference.py`
35	- `--offline-setup` — include offline package setup (frozen_packages pattern) in setup cell; auto-applied when `--inference-only`; ignored when `--eda-only` (EDA always online)
36	- `--resume <path>` — read existing `.py` script, extend/improve it
37
38	Output: `.experiments/kaggle/<competition-name>.py`
39
40	</inputs>
41
42	<constants>
43
44	```yaml
45	OUTPUT_DIR: .experiments/kaggle/
46	CELL_MARK: "# %%"
47	MD_CELL_MARK: "# %% [markdown]"
48	# NOTE: documentation-only — not referenced as shell vars across separate Bash() calls (state doesn't persist); keep values in sync with literal use sites (Steps 1, 3, 4).
49	```
50
51	</constants>
52
53	<compaction>
54
55	Key boundary: end of Step 3 — notebook script generated by `foundry:sw-engineer`, written to OUTFILE.
56	Preserve: OUTFILE path (derived from TMPDIR keys), COMPETITION_NAME (TMPDIR key), mode flags (EDA_ONLY, INFERENCE_ONLY, OFFLINE_SETUP).
57	Clear at Step 1 start (stale prior run) and after Step 4 package-distillation gate resolves.
58
59	</compaction>
60
61	<workflow>
62
63	Task hygiene: call `TaskList` first; close orphaned tasks. Create tasks per phase.
64
65	## Step 1: Parse arguments and gather context
66
67	```bash
68	# loads: compaction-contract.md
69	export CSID="${CLAUDE_CODE_SESSION_ID:-$PPID}"
70	ARGS="$ARGUMENTS"
71	COMPETITION_NAME=$(echo "$ARGS" \| awk '{print $1}')
72	RESUME_FLAG=""
73	EDA_ONLY=false
74	INFERENCE_ONLY=fal

People who liked this also viewed

6 picks

Type

kaggle

23stars3forks

View on GitHub

How to install kaggle?

borda/ai-rig/kaggle

$npx -y skills add Borda/AI-Rig --skill kaggle

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	<objective>
2
3	Generate Kaggle competition notebook script, Jupytext `# %%` format.
4
5	Two goals, equal weight — neither traded for other:
6	- Win — leaderboard-competitive: leakage-safe CV, metric-aligned loss/model choice, tuning/ensembling when it moves the score, not style theater
7	- Teach — read top to bottom like a university/seminar lecture on solving this competition: reader new to it follows the full reasoning chain, every decision motivated, nothing left as unexplained code
8
9	Follows user's ML research style distilled from past notebooks:
10	- PTL always for DNN training (PyTorch Lightning + torchmetrics) — even simple baselines
11	- Tool agnostic — best-fit library for problem; PTL when training loop needed
12	- Stages with lenses — each major stage: quick sanity check cell (show one batch, print shapes, verify submission format)
13	- Small, single-purpose cells — one action per cell (load, one transform, one plot, one check); never bundle setup + run + verify to save cell count
14	- Every cell earns its place — one-line why (comment or markdown sentence) before/in each cell: the specific reason this step happens now — never a restatement of what the code does
15	- Section markdown is extensive and structured — full explanation of what/why/how-it-advances-the-goal per section, formatted as tables/lists/blockquotes over dense prose paragraphs; markdown before a plot sets up the question, markdown after states the finding and its implication — plot and prose flow as one beat, never an orphaned chart
16	- `# !` bash over subprocess — package installs, `nvidia-smi`, `ls -lh`, `# ! head submission.csv`
17	- EDA is visual — distribution plots, sample grids, dimension scatters before any model
18	- Inference included — model save pattern + separate load-and-infer cells
19	- CSVLogger + seaborn — metrics plotted from `metrics.csv` after every training run
20
21	NOT for writing Python packages, modules, production code — notebook scripts only.
22	NOT research literature survey — use `/research:topic` for SOTA literature search.
23
24	</objective>
25
26	<inputs>
27
28	- $ARGUMENTS: one of:
29	- `<competition-name>` — short slug for output filename; generates blank template
30	- `<competition-name> <url>` — fetches competition overview from URL before generating
31	- `<competition-name> "<description>"` — inline description of problem and data
32	- `--type <type>` — hint: `classification`, `regression`, `segmentation`, `detection`, `tabular` (auto-detected when omitted)
33	- `--eda-only` — generate only EDA sections (no model/training/submission); always online (no offline setup)
34	- `--inference-only` — generate inference notebook from checkpoint (no EDA, no training); always offline (frozen packages pattern); loads checkpoint from `PATH_CHECKPOINT` constant; output suffix `-inference.py`
35	- `--offline-setup` — include offline package setup (frozen_packages pattern) in setup cell; auto-applied when `--inference-only`; ignored when `--eda-only` (EDA always online)
36	- `--resume <path>` — read existing `.py` script, extend/improve it
37
38	Output: `.experiments/kaggle/<competition-name>.py`
39
40	</inputs>
41
42	<constants>
43
44	```yaml
45	OUTPUT_DIR: .experiments/kaggle/
46	CELL_MARK: "# %%"
47	MD_CELL_MARK: "# %% [markdown]"
48	# NOTE: documentation-only — not referenced as shell vars across separate Bash() calls (state doesn't persist); keep values in sync with literal use sites (Steps 1, 3, 4).
49	```
50
51	</constants>
52
53	<compaction>
54
55	Key boundary: end of Step 3 — notebook script generated by `foundry:sw-engineer`, written to OUTFILE.
56	Preserve: OUTFILE path (derived from TMPDIR keys), COMPETITION_NAME (TMPDIR key), mode flags (EDA_ONLY, INFERENCE_ONLY, OFFLINE_SETUP).
57	Clear at Step 1 start (stale prior run) and after Step 4 package-distillation gate resolves.
58
59	</compaction>
60
61	<workflow>
62
63	Task hygiene: call `TaskList` first; close orphaned tasks. Create tasks per phase.
64
65	## Step 1: Parse arguments and gather context
66
67	```bash
68	# loads: compaction-contract.md
69	export CSID="${CLAUDE_CODE_SESSION_ID:-$PPID}"
70	ARGS="$ARGUMENTS"
71	COMPETITION_NAME=$(echo "$ARGS" \| awk '{print $1}')
72	RESUME_FLAG=""
73	EDA_ONLY=false
74	INFERENCE_ONLY=fal

People who liked this also viewed

6 picks

Type