kaggle-fleet 0.1.0__tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (34) hide show
  1. kaggle_fleet-0.1.0/LICENSE +21 -0
  2. kaggle_fleet-0.1.0/PKG-INFO +364 -0
  3. kaggle_fleet-0.1.0/README.md +332 -0
  4. kaggle_fleet-0.1.0/pyproject.toml +67 -0
  5. kaggle_fleet-0.1.0/setup.cfg +4 -0
  6. kaggle_fleet-0.1.0/src/kaggle_fleet/__init__.py +21 -0
  7. kaggle_fleet-0.1.0/src/kaggle_fleet/__main__.py +12 -0
  8. kaggle_fleet-0.1.0/src/kaggle_fleet/account_manager.py +687 -0
  9. kaggle_fleet-0.1.0/src/kaggle_fleet/cli.py +860 -0
  10. kaggle_fleet-0.1.0/src/kaggle_fleet/config.py +117 -0
  11. kaggle_fleet-0.1.0/src/kaggle_fleet/kaggle_runner.py +399 -0
  12. kaggle_fleet-0.1.0/src/kaggle_fleet/models.py +187 -0
  13. kaggle_fleet-0.1.0/src/kaggle_fleet/notebook_manager.py +205 -0
  14. kaggle_fleet-0.1.0/src/kaggle_fleet/server/__init__.py +7 -0
  15. kaggle_fleet-0.1.0/src/kaggle_fleet/server/api_handlers.py +439 -0
  16. kaggle_fleet-0.1.0/src/kaggle_fleet/server/http_server.py +334 -0
  17. kaggle_fleet-0.1.0/src/kaggle_fleet/server/static/app.js +704 -0
  18. kaggle_fleet-0.1.0/src/kaggle_fleet/server/static/index.html +306 -0
  19. kaggle_fleet-0.1.0/src/kaggle_fleet/server/static/styles.css +310 -0
  20. kaggle_fleet-0.1.0/src/kaggle_fleet/storage.py +132 -0
  21. kaggle_fleet-0.1.0/src/kaggle_fleet.egg-info/PKG-INFO +364 -0
  22. kaggle_fleet-0.1.0/src/kaggle_fleet.egg-info/SOURCES.txt +32 -0
  23. kaggle_fleet-0.1.0/src/kaggle_fleet.egg-info/dependency_links.txt +1 -0
  24. kaggle_fleet-0.1.0/src/kaggle_fleet.egg-info/entry_points.txt +3 -0
  25. kaggle_fleet-0.1.0/src/kaggle_fleet.egg-info/requires.txt +6 -0
  26. kaggle_fleet-0.1.0/src/kaggle_fleet.egg-info/top_level.txt +1 -0
  27. kaggle_fleet-0.1.0/tests/test_cli_json.py +120 -0
  28. kaggle_fleet-0.1.0/tests/test_core.py +431 -0
  29. kaggle_fleet-0.1.0/tests/test_isolation_fixes.py +186 -0
  30. kaggle_fleet-0.1.0/tests/test_link_login.py +111 -0
  31. kaggle_fleet-0.1.0/tests/test_server_and_cli.py +427 -0
  32. kaggle_fleet-0.1.0/tests/test_server_e2e_verification.py +349 -0
  33. kaggle_fleet-0.1.0/tests/test_verification.py +389 -0
  34. kaggle_fleet-0.1.0/tests/test_web_api.py +92 -0
@@ -0,0 +1,21 @@
1
+ MIT License
2
+
3
+ Copyright (c) 2026 The KaggleFleet contributors
4
+
5
+ Permission is hereby granted, free of charge, to any person obtaining a copy
6
+ of this software and associated documentation files (the "Software"), to deal
7
+ in the Software without restriction, including without limitation the rights
8
+ to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
9
+ copies of the Software, and to permit persons to whom the Software is
10
+ furnished to do so, subject to the following conditions:
11
+
12
+ The above copyright notice and this permission notice shall be included in all
13
+ copies or substantial portions of the Software.
14
+
15
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
16
+ IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
17
+ FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
18
+ AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
19
+ LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
20
+ OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
21
+ SOFTWARE.
@@ -0,0 +1,364 @@
1
+ Metadata-Version: 2.4
2
+ Name: kaggle-fleet
3
+ Version: 0.1.0
4
+ Summary: Run local Jupyter notebooks on Kaggle GPUs across a team of Kaggle accounts - from a CLI, a local web panel, or an AI agent.
5
+ Author: KaggleFleet contributors
6
+ License: MIT
7
+ Project-URL: Homepage, https://github.com/towhaEL/KaggleFleet
8
+ Project-URL: Issues, https://github.com/towhaEL/KaggleFleet/issues
9
+ Project-URL: Changelog, https://github.com/towhaEL/KaggleFleet/blob/main/CHANGELOG.md
10
+ Keywords: kaggle,gpu,jupyter,notebook,kernels,research,multi-account,cli,ai-agents
11
+ Classifier: Development Status :: 4 - Beta
12
+ Classifier: Environment :: Console
13
+ Classifier: Environment :: Web Environment
14
+ Classifier: Intended Audience :: Science/Research
15
+ Classifier: Intended Audience :: Developers
16
+ Classifier: License :: OSI Approved :: MIT License
17
+ Classifier: Operating System :: POSIX :: Linux
18
+ Classifier: Operating System :: MacOS
19
+ Classifier: Programming Language :: Python :: 3
20
+ Classifier: Programming Language :: Python :: 3 :: Only
21
+ Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
22
+ Classifier: Topic :: Utilities
23
+ Requires-Python: >=3.9
24
+ Description-Content-Type: text/markdown
25
+ License-File: LICENSE
26
+ Requires-Dist: kaggle>=1.7
27
+ Requires-Dist: six
28
+ Provides-Extra: dev
29
+ Requires-Dist: pytest>=7; extra == "dev"
30
+ Requires-Dist: ruff>=0.5; extra == "dev"
31
+ Dynamic: license-file
32
+
33
+ <p align="center">
34
+ <img src="https://raw.githubusercontent.com/towhaEL/KaggleFleet/main/docs/images/banner.png"
35
+ alt="KaggleFleet: run local Jupyter notebooks on Kaggle GPUs across your research team's own accounts" width="100%">
36
+ </p>
37
+
38
+ <p align="center">
39
+ <a href="https://pypi.org/project/kaggle-fleet/"><img src="https://img.shields.io/pypi/v/kaggle-fleet?color=6d5dfc" alt="PyPI version"></a>
40
+ <a href="https://pypi.org/project/kaggle-fleet/"><img src="https://img.shields.io/pypi/pyversions/kaggle-fleet" alt="Python versions"></a>
41
+ <a href="https://github.com/towhaEL/KaggleFleet/actions/workflows/ci.yml"><img src="https://github.com/towhaEL/KaggleFleet/actions/workflows/ci.yml/badge.svg" alt="CI"></a>
42
+ <a href="https://github.com/towhaEL/KaggleFleet/blob/main/LICENSE"><img src="https://img.shields.io/badge/license-MIT-green" alt="MIT license"></a>
43
+ <img src="https://img.shields.io/badge/platform-Linux%20%7C%20macOS%20%7C%20WSL-lightgrey" alt="Linux, macOS, WSL">
44
+ </p>
45
+
46
+ <p align="center">
47
+ <b>Run your local Jupyter notebooks on Kaggle GPUs, across your research team's own Kaggle accounts,<br>
48
+ from the terminal, a local web panel, or an AI agent.</b>
49
+ </p>
50
+
51
+ <p align="center">
52
+ <a href="#install">Install</a> ·
53
+ <a href="#quick-start">Quick start</a> ·
54
+ <a href="#web-panel">Web panel</a> ·
55
+ <a href="#using-kagglefleet-from-ai-agents-and-scripts">AI agents</a> ·
56
+ <a href="https://github.com/towhaEL/KaggleFleet/blob/main/CHANGELOG.md">Changelog</a>
57
+ </p>
58
+
59
+ <p align="center">
60
+ <img src="https://raw.githubusercontent.com/towhaEL/KaggleFleet/main/docs/images/demo.gif"
61
+ alt="Terminal demo: list accounts and notebooks, push a notebook to a co-author's account, check status and logs, download the outputs" width="860">
62
+ </p>
63
+
64
+ ```bash
65
+ pipx install kaggle-fleet
66
+ kfleet account login --label "Me"
67
+ kfleet run push train.ipynb --watch # runs on a Kaggle GPU, downloads outputs when done
68
+ ```
69
+
70
+ ## Why KaggleFleet?
71
+
72
+ Kaggle gives every account a free weekly GPU quota. A research group with several
73
+ co-authors has several quotas, but using them means logging in and out of Kaggle,
74
+ copying notebooks around, and tracking who ran what. KaggleFleet does that part for you:
75
+
76
+ - **Keep your notebooks local.** Edit in your own editor and git repo; KaggleFleet stages and pushes them as Kaggle kernels.
77
+ - **One machine, many co-author accounts.** Each Kaggle account lives in its own credential sandbox, so runs can go to different accounts at the same time without leaking into each other.
78
+ - **Push, watch, collect.** Pick an account and accelerator (T4, P100, TPU or CPU), push, follow status and logs, and download the outputs when the run finishes.
79
+ - **A web panel for the team.** Accounts, approximate GPU hours used per account this week, live job status, logs, and one-click launches.
80
+ - **Built for AI agents.** Every CLI command has `--json` output and meaningful exit codes, and [`AGENTS.md`](https://github.com/towhaEL/KaggleFleet/blob/main/AGENTS.md) documents the workflow for coding agents.
81
+ - **No runtime dependencies** besides the official `kaggle` CLI. The tool itself is pure Python standard library.
82
+
83
+ | Accounts | Jobs |
84
+ |---|---|
85
+ | ![Accounts view](https://raw.githubusercontent.com/towhaEL/KaggleFleet/main/docs/images/panel-accounts.png) | ![Jobs view](https://raw.githubusercontent.com/towhaEL/KaggleFleet/main/docs/images/panel-jobs.png) |
86
+
87
+ > [!IMPORTANT]
88
+ > KaggleFleet is for **teams of real people**, each using **their own** Kaggle account with their consent
89
+ > (for example co-authors on a paper). Kaggle's [Terms of Use](https://www.kaggle.com/terms) allow one account
90
+ > per person. Do not create extra accounts to get more GPU quota. KaggleFleet is not affiliated with Kaggle or Google.
91
+
92
+ ---
93
+
94
+ ## Contents
95
+
96
+ - [Install](#install)
97
+ - [Quick start](#quick-start)
98
+ - [Adding co-author accounts](#adding-co-author-accounts)
99
+ - [Running notebooks](#running-notebooks)
100
+ - [Web panel](#web-panel)
101
+ - [Using KaggleFleet from AI agents and scripts](#using-kagglefleet-from-ai-agents-and-scripts)
102
+ - [Configuration](#configuration)
103
+ - [How credential isolation works](#how-credential-isolation-works)
104
+ - [CLI reference](#cli-reference)
105
+ - [Limitations](#limitations)
106
+ - [Development](#development)
107
+ - [Contributing and releases](#contributing-and-releases)
108
+
109
+ ## Install
110
+
111
+ You need Python 3.9 or newer on **Linux or macOS**. On Windows, use WSL.
112
+
113
+ ```bash
114
+ pipx install kaggle-fleet # recommended: isolated, puts `kfleet` on your PATH
115
+ # or
116
+ pip install kaggle-fleet
117
+ ```
118
+
119
+ To try the latest unreleased code: `pipx install git+https://github.com/towhaEL/KaggleFleet.git`.
120
+
121
+ This installs two commands, `kfleet` and its longer alias `kaggle-fleet`, and pulls in the official `kaggle` CLI.
122
+ Browser sign-in (`kfleet account login`) requires Kaggle CLI 2.x. API-key accounts work with 1.7 and newer.
123
+
124
+ Check your setup:
125
+
126
+ ```bash
127
+ kfleet doctor
128
+ ```
129
+
130
+ ## Quick start
131
+
132
+ ```bash
133
+ cd my-research-project/ # a folder with your notebooks, or a ./notebooks subfolder
134
+
135
+ kfleet account login --label "Me" # sign in with your Kaggle account in the browser
136
+ kfleet notebook list # see what KaggleFleet found
137
+ kfleet run push train.ipynb --watch
138
+ ```
139
+
140
+ To try it on something small first, push [`examples/hello_gpu.ipynb`](https://github.com/towhaEL/KaggleFleet/blob/main/examples/hello_gpu.ipynb), which reports the GPU it got.
141
+
142
+ With `--watch`, KaggleFleet polls the run until it finishes and then downloads the kernel's
143
+ output files to `./outputs/<kernel-slug>/`. Without it, the push returns right away and you can
144
+ check on the run later:
145
+
146
+ ```bash
147
+ kfleet run list
148
+ kfleet run status # latest job; or pass a job id / kernel slug
149
+ kfleet run logs
150
+ kfleet run download
151
+ ```
152
+
153
+ Or open the web panel:
154
+
155
+ ```bash
156
+ kfleet serve --open # http://127.0.0.1:8080
157
+ ```
158
+
159
+ ## Adding co-author accounts
160
+
161
+ Each co-author adds their own account once, on the machine that runs KaggleFleet.
162
+
163
+ ### Option A: browser sign-in (recommended)
164
+
165
+ ```bash
166
+ kfleet account login --label "Co-author B"
167
+ ```
168
+
169
+ This opens Kaggle's sign-in page in your normal browser. **Whichever Kaggle account is signed in
170
+ there is the account that gets added**, so switch the browser to the co-author's account first,
171
+ or use `--no-browser` to get a link you can open in a private window or on another device and
172
+ paste the verification code back.
173
+
174
+ The login runs inside a throwaway sandbox. Your own `~/.kaggle` login is never read or changed,
175
+ and adding an account that is already configured is refused.
176
+
177
+ In the web panel, **Sign in with browser** offers the same flow as a link plus a code field, which
178
+ works well when a co-author signs in from their own laptop.
179
+
180
+ ### Option B: API key (`kaggle.json`)
181
+
182
+ The co-author creates a token at **kaggle.com → Settings → API → Create New Token** and sends you
183
+ the `kaggle.json` file, or its username and key:
184
+
185
+ ```bash
186
+ kfleet account add --username alice --key 0123456789abcdef... --label "Alice"
187
+ ```
188
+
189
+ ### Option C: import what is already on this machine
190
+
191
+ ```bash
192
+ kfleet account import # picks up ~/.kaggle/kaggle.json and ~/.kaggle/credentials.json
193
+ ```
194
+
195
+ ### Managing accounts
196
+
197
+ ```bash
198
+ kfleet account list
199
+ kfleet account test alice # checks that the credentials authenticate as *alice*, not just as someone
200
+ kfleet account switch alice # change the default account for pushes
201
+ kfleet account remove alice # deletes the local sandbox only; the Kaggle account is untouched
202
+ ```
203
+
204
+ ## Running notebooks
205
+
206
+ ```bash
207
+ kfleet run push 02_train.ipynb \
208
+ --account bob \
209
+ --accelerator NvidiaTeslaP100 \
210
+ --dataset bob/preprocessed-data \
211
+ --kernel-source alice/nb-00-data-prep \
212
+ --max-runtime 7200 \
213
+ --watch
214
+ ```
215
+
216
+ | Option | Meaning |
217
+ |---|---|
218
+ | `--account`, `-a` | Account id or username. Defaults to the active account. |
219
+ | `--accelerator` | `NvidiaTeslaT4` (default), `NvidiaTeslaP100`, `TpuV5E8` or `None` |
220
+ | `--no-gpu` | Run on CPU |
221
+ | `--no-internet` | Disable internet access inside the kernel |
222
+ | `--dataset`, `--competition`, `--kernel-source`, `--model` | Attach Kaggle inputs. Each flag can be repeated. |
223
+ | `--title`, `--slug` | Override the kernel title and slug, which are derived from the notebook by default |
224
+ | `--max-runtime SECONDS` | Ask Kaggle to stop the kernel after this long. The default is Kaggle's own session limit. |
225
+ | `--watch`, `-w` | Wait for completion, then download outputs (turn this off with `--no-download`) |
226
+ | `--timeout SECONDS` | How long `--watch` waits. The kernel itself is not affected. |
227
+
228
+ **Kernel naming.** The kernel title comes from the notebook's first `# heading`. The slug comes from the
229
+ file name: `02_train.ipynb` becomes `<username>/nb-02-train`. File names that start with a digit get
230
+ a prefix, `nb-` by default; set `KAGGLE_FLEET_SLUG_PREFIX=myproject` to use your project's name instead.
231
+ Kernels are always created as **private**.
232
+
233
+ **Re-running a notebook** pushes a new version of the same kernel. Every push starts a new run on
234
+ Kaggle and uses GPU quota.
235
+
236
+ ## Web panel
237
+
238
+ ```bash
239
+ kfleet serve # http://127.0.0.1:8080
240
+ kfleet serve --port 9000 --open
241
+ ```
242
+
243
+ The panel has three views:
244
+
245
+ - **Accounts:** add, test, switch and remove accounts, with each account's estimated GPU hours for the last 7 days.
246
+ - **Launch notebook:** pick a notebook, an account and hardware, then push.
247
+ - **Jobs:** live status, logs, refresh and download for every run.
248
+
249
+ The panel has **no login**. It listens on `127.0.0.1` and rejects requests from other websites
250
+ (it checks the `Origin` and `Host` headers and requires JSON request bodies). To use it from another
251
+ machine, prefer an SSH tunnel over exposing it:
252
+
253
+ ```bash
254
+ ssh -L 8080:127.0.0.1:8080 your-gpu-box # then open http://127.0.0.1:8080 locally
255
+ ```
256
+
257
+ The panel is backed by a small JSON REST API, documented in [`docs/rest-api.md`](https://github.com/towhaEL/KaggleFleet/blob/main/docs/rest-api.md).
258
+
259
+ ## Using KaggleFleet from AI agents and scripts
260
+
261
+ Add `--json` to any command, before or after the subcommand, to get a single JSON document on stdout.
262
+ Failures print `{"error": "..."}` and exit with a non-zero code. Account secrets are never printed.
263
+
264
+ ```bash
265
+ $ kfleet run push 02_train.ipynb --account bob --json
266
+ {
267
+ "success": true,
268
+ "job": {
269
+ "id": "job_20261121_101500_nb-02-train_3f2a",
270
+ "kernel_id": "bob/nb-02-train",
271
+ "status": "RUNNING",
272
+ ...
273
+ }
274
+ }
275
+
276
+ $ kfleet run status job_20261121_101500_nb-02-train_3f2a --json
277
+ $ kfleet run logs job_20261121_101500_nb-02-train_3f2a --json
278
+ $ kfleet run download job_20261121_101500_nb-02-train_3f2a --json # lists the downloaded files
279
+ ```
280
+
281
+ [`AGENTS.md`](https://github.com/towhaEL/KaggleFleet/blob/main/AGENTS.md) is a guide written for coding agents such as Claude Code or Codex:
282
+ the workflow, the JSON shapes, how to pick an account with spare GPU quota, and what not to do.
283
+ Copy it, or link to it, from your research repo's own `AGENTS.md` / `CLAUDE.md`.
284
+
285
+ ## Configuration
286
+
287
+ | Environment variable | Default | Purpose |
288
+ |---|---|---|
289
+ | `KAGGLE_FLEET_DIR` | `~/.kaggle_fleet` | State: accounts, credential sandboxes, job history, staging |
290
+ | `KAGGLE_FLEET_NOTEBOOKS_DIR` | `./notebooks` if it exists, else `.` | Where notebooks are discovered |
291
+ | `KAGGLE_FLEET_OUTPUTS_DIR` | `./outputs` | Where kernel outputs are downloaded |
292
+ | `KAGGLE_FLEET_KAGGLE_BIN` | auto-detected | Path to the `kaggle` executable |
293
+ | `KAGGLE_FLEET_SLUG_PREFIX` | `nb` | Prefix for slugs derived from file names that start with a digit |
294
+
295
+ `--notebooks-dir` and `--outputs-dir` override the matching variables for a single command.
296
+
297
+ ## How credential isolation works
298
+
299
+ Every account gets a private directory, `~/.kaggle_fleet/envs/<account-id>/` (mode `0700`, files `0600`).
300
+ Each `kaggle` command for that account runs with:
301
+
302
+ - `KAGGLE_CONFIG_DIR` pointing at that directory, and
303
+ - `HOME` set to a per-account home whose `.kaggle` links to that directory. The Kaggle SDK reads OAuth tokens
304
+ from the hard-coded path `~/.kaggle/credentials.json` and ignores `KAGGLE_CONFIG_DIR`, so without a
305
+ separate `HOME`, every OAuth account would silently run as whoever is logged in to your real `~/.kaggle`.
306
+ - Any `KAGGLE_KEY` or `KAGGLE_API_TOKEN` from your shell removed, because those take priority over all other credentials.
307
+
308
+ When the Kaggle CLI refreshes an OAuth token, the new token is copied back into `accounts.json`.
309
+ `kfleet account test` checks the identity Kaggle reports (`kaggle config view`), so a credential
310
+ leaking from another account would be caught instead of passing silently.
311
+
312
+ Credentials are stored **unencrypted** on disk and protected only by file permissions, the same way
313
+ the Kaggle CLI stores them. Run KaggleFleet on a machine you trust. See [`SECURITY.md`](https://github.com/towhaEL/KaggleFleet/blob/main/SECURITY.md).
314
+
315
+ ## CLI reference
316
+
317
+ ```
318
+ kfleet [--json] [--notebooks-dir DIR] [--outputs-dir DIR] <command>
319
+
320
+ account list | add | add-oauth | login | switch | test | remove | import
321
+ notebook list | inspect <notebook>
322
+ run push <notebook> | list | status [job] | logs [job] | download [job]
323
+ serve [--host] [--port] [--open] [--allow-host HOST]
324
+ doctor
325
+ ```
326
+
327
+ Run `kfleet <command> <subcommand> --help` for every option.
328
+ `python -m kaggle_fleet` works the same way as `kfleet`.
329
+
330
+ ## Limitations
331
+
332
+ - **The GPU hours shown are estimates.** They count only runs launched through KaggleFleet, and a run's end time is when KaggleFleet noticed it finished. Kaggle's own quota page is authoritative.
333
+ - **Status is polled.** Status comes from `kaggle kernels status`, so runs are refreshed while you watch them (CLI `--watch` or the web panel) or when you ask (`run status`).
334
+ - **Kernels cannot be cancelled from here** because the Kaggle CLI does not support it. Stop a run on kaggle.com.
335
+ - **Linux and macOS only** (symlink-based sandboxes). On Windows, use WSL.
336
+
337
+ ## Development
338
+
339
+ ```bash
340
+ git clone https://github.com/towhaEL/KaggleFleet.git
341
+ cd KaggleFleet
342
+ python -m venv .venv && source .venv/bin/activate
343
+ pip install -e ".[dev]"
344
+
345
+ pytest # or: python -m unittest discover -s tests -t .
346
+ ruff check src tests
347
+ ```
348
+
349
+ The test suite never contacts Kaggle: the `kaggle` CLI is mocked or replaced with a fake script.
350
+ See [`CONTRIBUTING.md`](https://github.com/towhaEL/KaggleFleet/blob/main/CONTRIBUTING.md).
351
+
352
+ ## Contributing and releases
353
+
354
+ Bug reports, feature ideas and pull requests are welcome. Please read
355
+ [`CONTRIBUTING.md`](https://github.com/towhaEL/KaggleFleet/blob/main/CONTRIBUTING.md) first, and report
356
+ security issues privately as described in [`SECURITY.md`](https://github.com/towhaEL/KaggleFleet/blob/main/SECURITY.md).
357
+
358
+ Releases are listed on the [GitHub releases page](https://github.com/towhaEL/KaggleFleet/releases) and
359
+ [PyPI](https://pypi.org/project/kaggle-fleet/). Maintainers: see
360
+ [`docs/RELEASING.md`](https://github.com/towhaEL/KaggleFleet/blob/main/docs/RELEASING.md).
361
+
362
+ ## License
363
+
364
+ [MIT](https://github.com/towhaEL/KaggleFleet/blob/main/LICENSE)