Update README.md
Browse files
README.md
CHANGED
|
@@ -15,7 +15,8 @@ pipeline_tag: text-generation
|
|
| 15 |
# Overview
|
| 16 |
|
| 17 |
ToolMind-web-3B is a specialized lightweight agent built on top of the **Nanbeige4-3B** foundation model.
|
| 18 |
-
Following extensive **SFT (Supervised Fine-Tuning)** and **RL (Reinforcement Learning)** focused on search behaviors,
|
|
|
|
| 19 |
|
| 20 |
# Highlights
|
| 21 |
|
|
|
|
| 15 |
# Overview
|
| 16 |
|
| 17 |
ToolMind-web-3B is a specialized lightweight agent built on top of the **Nanbeige4-3B** foundation model.
|
| 18 |
+
Following extensive **SFT (Supervised Fine-Tuning)** and **RL (Reinforcement Learning)** focused on search behaviors,
|
| 19 |
+
our model attains leading performance among small-scale models on multiple long-horizon leaderboards, enabling reliable execution of up to hundreds of consecutive tool invocations.
|
| 20 |
|
| 21 |
# Highlights
|
| 22 |
|