Update models quickstart

[ngrayluna]

Fixes:

• Missing context
• out of place content/markdown cells
• broken URLs
• incorrect info
• removed extra spaces from cells, etc.

[review-notebook-app]

Check out this pull request on 

See visual diffs & provide feedback on Jupyter Notebooks.

---

Powered by ReviewNB

[Copilot]

Pull request overview

Updates the run_quickstart Colab notebook to improve the quickstart flow (installation/import/login), correct and modernize links/content, and clean up markdown/cell organization.

Changes:

• Reworked the opening sections (install + imports) and added a dedicated login section.
• Updated documentation links and revised explanatory markdown around runs/config usage.
• Cleaned up cell content/spacing and adjusted notebook metadata.

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated no new comments.

Comments suppressed due to low confidence (5)

colabs/intro/run_quickstart.ipynb:9

• The notebook no longer includes the standard “Open in Colab” badge/link at the top. Most notebooks under colabs/ include this badge so users can open the notebook directly from GitHub; removing it makes the quickstart harder to discover/run in Colab.

   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Use W&B to track, visualize, and manage machine learning experiments of any size.\n",
    "\n",
    "Install W&B to track, visualize, and manage machine learning experiments of any size.\n",

colabs/intro/run_quickstart.ipynb:60

• The “Log in” step only sets WANDB_API_KEY but never actually calls wandb.login(). Other colabs/intro/* notebooks call wandb.login() to validate authentication up front; without it, users won’t get immediate feedback if the key is missing/invalid until wandb.init() later.

   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "os.environ[\"WANDB_API_KEY\"] = getpass(\"Enter your W&B API key: \")"
   ]

colabs/intro/run_quickstart.ipynb:101

• The markdown explains that metrics are logged for epochs 2–9, but the code doesn’t log epoch (and therefore charts won’t be able to use epoch as an x-axis without extra configuration). Logging epoch and using it as the step keeps the output aligned with the explanation.

    "    for epoch in range(2, config['epochs']):\n",
    "        acc = 1 - 2**-config['epochs'] - random.random() / config['epochs'] - offset\n",
    "        loss = 2**-config['epochs'] + random.random() / config['epochs'] + offset\n",
    "        print(f\"epoch={epoch}, accuracy={acc}, loss={loss}\")\n",
    "        run.log({\"accuracy\": acc, \"loss\": loss})"

colabs/intro/run_quickstart.ipynb:108

• wandb.ai/home is written as bare text without a URL scheme, which won’t be a clickable link in all notebook renderers. Use an explicit https:// link (or Markdown link) to avoid another broken/non-clickable URL.

    "Click the `\"View project at:\"` link in the prior cell's output (or go to wandb.ai/home) to view your experiment. Each run appears in your project's **Workspace** and **Runs** page with auto-generated names."

colabs/intro/run_quickstart.ipynb:130

• Notebook metadata switched from the usual kernelspec (used throughout colabs/, e.g. colabs/intro/Intro_to_Weights_&_Biases.ipynb:423-426) to a language_info block. Colab/Jupyter tooling typically relies on kernelspec; this change is inconsistent with the rest of the repo and may reduce compatibility.

  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated no new comments.

Comments suppressed due to low confidence (3)

colabs/intro/run_quickstart.ipynb:109

• This notebook metadata no longer includes a kernelspec. Most other notebooks under colabs/ include it, and removing it can make the notebook less portable in Jupyter environments outside Colab (kernel selection may be missing/ambiguous).

 "metadata": {
  "accelerator": "GPU",
  "colab": {
   "include_colab_link": true,
   "provenance": [],

colabs/intro/run_quickstart.ipynb:59

• The login cell sets WANDB_API_KEY directly from getpass() but never calls wandb.login(). Many other Colab notebooks in this repo call wandb.login() to validate auth and handle the interactive login flow consistently.

    "os.environ[\"WANDB_API_KEY\"] = getpass(\"Enter your W&B API key: \")"

colabs/intro/run_quickstart.ipynb:101

• The training loop starts at epoch 2 (range(2, config['epochs'])), so it runs only epochs-2 iterations. Also, run.log() omits the epoch value, making it harder to correlate metrics to epochs in the UI.

    "    for epoch in range(2, config['epochs']):\n",
    "        acc = 1 - 2**-config['epochs'] - random.random() / config['epochs'] - offset\n",
    "        loss = 2**-config['epochs'] + random.random() / config['epochs'] + offset\n",
    "        print(f\"epoch={epoch}, accuracy={acc}, loss={loss}\")\n",
    "        run.log({\"accuracy\": acc, \"loss\": loss})"