docs: describe as arena benchmark repository; Apache-2.0
Browse files
README.md
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
---
|
| 4 |
+
|
| 5 |
+
# ovos-intent-bench-massive-templates-train
|
| 6 |
+
|
| 7 |
+
An OVOS Plugin Arena benchmark repository. The arena's prediction runner
|
| 8 |
+
publishes each plugin's raw output on the `massive-templates-train` dataset here, one JSON-lines
|
| 9 |
+
file per plugin under `predictions/<lang>/`, and where a sample-set manifest
|
| 10 |
+
governs scoring it lives under `sample_sets/`. The public leaderboard at
|
| 11 |
+
https://openvoiceos.github.io/ovos-plugin-arena/ is computed from these rows,
|
| 12 |
+
and anyone can recompute an entry from them without access to the arena's own
|
| 13 |
+
infrastructure.
|
| 14 |
+
|
| 15 |
+
It contains no benchmark audio or text corpus; those live in their own
|
| 16 |
+
datasets under their own licences. Rows appear here as sweeps run, so an
|
| 17 |
+
empty repository means no plugin has been scored on this dataset yet.
|
| 18 |
+
|
| 19 |
+
Row and manifest formats: `docs/SPECIFICATION.md` in
|
| 20 |
+
https://github.com/OpenVoiceOS/ovos-plugin-arena.
|