Update README.md
Browse files
README.md
CHANGED
|
@@ -9,4 +9,22 @@ metrics:
|
|
| 9 |
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
|
| 10 |
tags:
|
| 11 |
- code
|
| 12 |
-
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 9 |
base_model: meta-llama/Meta-Llama-3.1-8B-Instruct
|
| 10 |
tags:
|
| 11 |
- code
|
| 12 |
+
---
|
| 13 |
+
|
| 14 |
+
|
| 15 |
+
# ToolACE-8B
|
| 16 |
+
|
| 17 |
+
ToolACE-8B is a finetuned model of LLaMA-3.1-8B-Instruct with our dataset [ToolACE](https://huggingface.co/datasets/Team-ACE/ToolACE) tailored for tool usage.
|
| 18 |
+
ToolACE-8B achieves state-of-the-art performance on the [Berkeley Function-Calling Leaderboard](https://gorilla.cs.berkeley.edu/leaderboard.html#leaderboard), rivaling the latest GPT-4 models.
|
| 19 |
+
|
| 20 |
+
|
| 21 |
+
ToolACE is an automatic agentic pipeline designed to generate **A**ccurate, **C**omplex, and div**E**rse tool-learning data.
|
| 22 |
+
ToolACE leverages a novel self-evolution synthesis process to curate a comprehensive API pool of 26,507 diverse APIs.
|
| 23 |
+
Dialogs are further generated through the interplay among multiple agents, guided by a formalized thinking process.
|
| 24 |
+
To ensure data accuracy, we implement a dual-layer verification system combining rule-based and model-based checks.
|
| 25 |
+
|
| 26 |
+
|
| 27 |
+
|
| 28 |
+
|
| 29 |
+
|
| 30 |
+
|