Spaces:
Running
Running
card: a model ships only if it clears zero
Browse files
README.md
CHANGED
|
@@ -10,10 +10,11 @@ pinned: false
|
|
| 10 |
|
| 11 |
**Turn agent mistakes into training data.** Production traces become a dataset, the
|
| 12 |
dataset post-trains an open model you own, and a held-out test says whether the agent
|
| 13 |
-
stopped repeating the mistake.
|
| 14 |
|
| 15 |
Everything here is produced by our open-source SDK and published with the numbers
|
| 16 |
-
behind it.
|
|
|
|
| 17 |
|
| 18 |
```bash
|
| 19 |
pip install whileai
|
|
|
|
| 10 |
|
| 11 |
**Turn agent mistakes into training data.** Production traces become a dataset, the
|
| 12 |
dataset post-trains an open model you own, and a held-out test says whether the agent
|
| 13 |
+
stopped repeating the mistake. The last step is the one most pipelines skip.
|
| 14 |
|
| 15 |
Everything here is produced by our open-source SDK and published with the numbers
|
| 16 |
+
behind it. A model ships when it beats its base on a held-out set with an interval
|
| 17 |
+
that clears zero. If it does not, it does not ship.
|
| 18 |
|
| 19 |
```bash
|
| 20 |
pip install whileai
|