zxshamson commited on
Commit
9c79612
·
verified ·
1 Parent(s): 08f56d6

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +33 -1
README.md CHANGED
@@ -21,12 +21,16 @@ ToolACE is an automatic agentic pipeline designed to generate **A**ccurate, **C*
21
  ToolACE leverages a novel self-evolution synthesis process to curate a comprehensive API pool of 26,507 diverse APIs.
22
  Dialogs are further generated through the interplay among multiple agents, guided by a formalized thinking process.
23
  To ensure data accuracy, we implement a dual-layer verification system combining rule-based and model-based checks.
24
- More details can be found in our paper on arxiv: [*ToolACE: Winning the Points of LLM Function Calling*](https://arxiv.org/abs/2409.00920)
25
 
26
 
27
  ![image/jpeg](https://cdn-uploads.huggingface.co/production/uploads/66bf01f45bdd611f9a602087/WmyWOYtg_dbTgwQmvlqcz.jpeg)
28
 
 
29
 
 
 
 
30
 
31
 
32
  ### Usage
@@ -148,3 +152,31 @@ If you think ToolACE is useful in your work, please cite our paper:
148
  }
149
  ```
150
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
21
  ToolACE leverages a novel self-evolution synthesis process to curate a comprehensive API pool of 26,507 diverse APIs.
22
  Dialogs are further generated through the interplay among multiple agents, guided by a formalized thinking process.
23
  To ensure data accuracy, we implement a dual-layer verification system combining rule-based and model-based checks.
24
+ More details can be found in our paper: [*ToolACE: Winning the Points of LLM Function Calling*](https://arxiv.org/abs/2409.00920)
25
 
26
 
27
  ![image/jpeg](https://cdn-uploads.huggingface.co/production/uploads/66bf01f45bdd611f9a602087/WmyWOYtg_dbTgwQmvlqcz.jpeg)
28
 
29
+ More techniques are applied to further improve tool-usage ability, including:
30
 
31
+ - [*ToolACE-R: Model-aware Iterative Training and Adaptive Refinement for Tool Learning*](https://arxiv.org/abs/2504.01400)
32
+ - [*ToolACE-DEV: Self-Improving Tool Learning via Decomposition and EVolution*](https://arxiv.org/abs/2505.07512)
33
+ - [*ToolACE-MT: Non-Autoregressive Generation for Agentic Multi-Turn Interaction*](https://arxiv.org/abs/2508.12685)
34
 
35
 
36
  ### Usage
 
152
  }
153
  ```
154
 
155
+ Additionally, please check our other related works whose techniques are applied in ToolACE-2.5-8B:
156
+
157
+ ```
158
+ @article{zeng2025toolacer,
159
+ title={ToolACE-R: Model-aware Iterative Training and Adaptive Refinement for Tool Learning},
160
+ author={Zeng, Xingshan and Liu, Weiwen and Huang, Xu and Wang, Zezhong and Wang, Lingzhi and Li, Liangyou and Wang, Yasheng and Shang, Lifeng and Jiang, Xin and Tang, Ruiming and Liu, Qun},
161
+ journal={arXiv preprint arXiv:2504.01400},
162
+ year={2025}
163
+ }
164
+ ```
165
+
166
+ ```
167
+ @article{huang2025toolace,
168
+ title={ToolACE-DEV: Self-Improving Tool Learning via Decomposition and EVolution},
169
+ author={Huang, Xu and Liu, Weiwen and Zeng, Xingshan and Huang, Yuefeng and Hao, Xinlong and Wang, Yuxian and Zeng, Yirong and Wu, Chuhan and Wang, Yasheng and Tang, Ruiming and Lian, Defu},
170
+ journal={arXiv preprint arXiv:2505.07512},
171
+ year={2025}
172
+ }
173
+ ```
174
+
175
+ ```
176
+ @article{zeng2025toolacemt,
177
+ title={ToolACE-MT: Non-Autoregressive Generation for Agentic Multi-Turn Interaction},
178
+ author={Zeng, Xingshan and Liu, Weiwen and Wang, Lingzhi and Li, Liangyou and Mi, Fei and Wang, Yasheng and Shang, Lifeng and Jiang, Xin and Liu, Qun},
179
+ journal={arXiv preprint arXiv:2508.12685},
180
+ year={2025}
181
+ }
182
+ ```