Card: #1 under 300M parameters on Decision Index 0.4; drop version history
Browse files
README.md
CHANGED
|
@@ -15,18 +15,15 @@ tags:
|
|
| 15 |
|
| 16 |
# tasksource-decider-nano
|
| 17 |
|
| 18 |
-
This repository was `tasksource/tasksource-jev-nano-v1`. Its `main` (tag `v5`) is the joint cross-encoder trained for
|
| 19 |
-
64,000 steps on an updated mix. Earlier versions stay at their tags: [`v4`](https://huggingface.co/tasksource/tasksource-decider-nano/tree/v4)
|
| 20 |
-
(48,000 steps, public index 30.90), [`v3`](https://huggingface.co/tasksource/tasksource-decider-nano/tree/v3) (24,000 steps, 29.05),
|
| 21 |
-
[`v2`](https://huggingface.co/tasksource/tasksource-decider-nano/tree/v2) (12,000 steps, 26.77) and the previous multi-vector model at
|
| 22 |
-
[`v1`](https://huggingface.co/tasksource/tasksource-decider-nano/tree/v1).
|
| 23 |
-
|
| 24 |
A 150M-parameter typed-decision model for the [Decision Index](https://huggingface.co/spaces/multimodalart/jev-decision-index).
|
| 25 |
It takes a `state` and typed questions (`choice`, `noul`) and returns a probability for every option. It does not
|
| 26 |
generate text.
|
| 27 |
|
|
|
|
|
|
|
| 28 |
| | |
|
| 29 |
|---|---|
|
|
|
|
| 30 |
| Decision Index 0.3, public index | **33.51** |
|
| 31 |
| Requests | 140,171 `ok`, 7 `unsupported` (pairs above 8,192 tokens) |
|
| 32 |
| Latency, 1脳 NVIDIA A30 24 GB (shared), one request at a time | median 23 ms 路 mean 33 ms 路 p80 37 ms |
|
|
|
|
| 15 |
|
| 16 |
# tasksource-decider-nano
|
| 17 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 18 |
A 150M-parameter typed-decision model for the [Decision Index](https://huggingface.co/spaces/multimodalart/jev-decision-index).
|
| 19 |
It takes a `state` and typed questions (`choice`, `noul`) and returns a probability for every option. It does not
|
| 20 |
generate text.
|
| 21 |
|
| 22 |
+
**#1 among models under 300M parameters on the [Decision Index 0.4](https://huggingface.co/spaces/multimodalart/jev-decision-index)** (Full 19.85, October 2026).
|
| 23 |
+
|
| 24 |
| | |
|
| 25 |
|---|---|
|
| 26 |
+
| Decision Index 0.4, Full (private same-skills + new-domains) | **19.85** |
|
| 27 |
| Decision Index 0.3, public index | **33.51** |
|
| 28 |
| Requests | 140,171 `ok`, 7 `unsupported` (pairs above 8,192 tokens) |
|
| 29 |
| Latency, 1脳 NVIDIA A30 24 GB (shared), one request at a time | median 23 ms 路 mean 33 ms 路 p80 37 ms |
|