vanishingradient commited on
Commit
a5b37a7
·
verified ·
1 Parent(s): a4f498d

Add kaggle/KAGGLE_SETUP.md

Browse files
Files changed (1) hide show
  1. kaggle/KAGGLE_SETUP.md +11 -11
kaggle/KAGGLE_SETUP.md CHANGED
@@ -5,20 +5,20 @@
5
  - **Internet:** On
6
  - **Language:** Python
7
 
8
- ## 2. Secrets (Add-ons → Secrets → Attach to notebook)
9
- | Secret name | Value |
10
- |-------------|--------|
11
- | `HF_TOKEN` | same as local `.env` |
12
- | `WANDB_API_KEY` | same as local `.env` |
13
- | `LANGFUSE_PUBLIC_KEY` | same as local `.env` |
14
- | `LANGFUSE_SECRET_KEY` | same as local `.env` |
15
- | `LANGFUSE_HOST` | `https://cloud.langfuse.com` (optional) |
16
 
17
- Kaggle does **not** allow uploading secrets via API from this machine — paste once in the UI.
18
 
19
- **Attach toggles:** After creating each secret, open the notebook’s Secrets panel and ensure the **switch is ON** next to every label. Creating a secret without attaching it causes `HTTP Error 400` / `ConnectionError` on `get_secret`.
 
 
 
 
 
 
 
20
 
21
- **OpenTelemetry / google-adk warnings:** The notebook uninstalls `google-adk` (unused). Safe to ignore if you still see a residual pip conflict message after that.
22
 
23
  ## 3. Checkpoint / observability contract
24
  Every `save_steps` (default 5):
 
5
  - **Internet:** On
6
  - **Language:** Python
7
 
8
+ ## 2. Secrets (API path — recommended)
 
 
 
 
 
 
 
9
 
10
+ Kaggle **cannot** create/attach User Secrets via API. We use a **private dataset** instead:
11
 
12
+ ```bash
13
+ python scripts/push_kaggle_secrets_dataset.py # uploads saqlain22/deepagents-rl-secrets (private)
14
+ kaggle kernels push -p notebooks/ # attaches it via dataset_sources + enable_gpu
15
+ ```
16
+
17
+ Notebook loads `/kaggle/input/.../secrets.json` first, then falls back to Add-ons → Secrets if present.
18
+
19
+ **Do not** make `deepagents-rl-secrets` public.
20
 
21
+ **OpenTelemetry / google-adk warnings:** Notebook uninstalls unused `google-adk`.
22
 
23
  ## 3. Checkpoint / observability contract
24
  Every `save_steps` (default 5):