Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2509.24229 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908565016412160 |
|---|---|
| author | Wang, Kangxu Chen, Ze Wei, Chengcheng Zheng, Jiewen He, Jiarong Gao, Max |
| author_facet | Wang, Kangxu Chen, Ze Wei, Chengcheng Zheng, Jiewen He, Jiarong Gao, Max |
| contents | This paper presents the opdainlp team's solution for the GPU track of the CPDC 2025 challenge. The challenge consists of three tasks, aiming to build an in-game conversational AI that adheres to character personas, aligns with the game's worldview, and supports function calling. Considering both effectiveness and resource/time constraints during inference, we synthesized data for some of the tasks based on the datasets provided by the competition organizers. We employed Qwen3-14B with LoRA fine-tuning and model fusion, and utilized a base model integrated with multiple LoRA adapters during inference. Specifically, in the competition, we used three distinct LoRA adapters to handle tool calling, response generation with tool call results, and response generation without tool call results, respectively. MultiLoRA inference was implemented using vLLM. Our solution achieved the first place in Task 1 and Task 3, and the second place in Task 2 of the GPU track. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_24229 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Model Fusion with Multi-LoRA Inference for Tool-Enhanced Game Dialogue Agents Wang, Kangxu Chen, Ze Wei, Chengcheng Zheng, Jiewen He, Jiarong Gao, Max Computation and Language This paper presents the opdainlp team's solution for the GPU track of the CPDC 2025 challenge. The challenge consists of three tasks, aiming to build an in-game conversational AI that adheres to character personas, aligns with the game's worldview, and supports function calling. Considering both effectiveness and resource/time constraints during inference, we synthesized data for some of the tasks based on the datasets provided by the competition organizers. We employed Qwen3-14B with LoRA fine-tuning and model fusion, and utilized a base model integrated with multiple LoRA adapters during inference. Specifically, in the competition, we used three distinct LoRA adapters to handle tool calling, response generation with tool call results, and response generation without tool call results, respectively. MultiLoRA inference was implemented using vLLM. Our solution achieved the first place in Task 1 and Task 3, and the second place in Task 2 of the GPU track. |
| title | Model Fusion with Multi-LoRA Inference for Tool-Enhanced Game Dialogue Agents |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2509.24229 |