DCAgent/a1-mind2web: Fine-tuned for Web Interaction

DCAgent/a1-mind2web is an 8 billion parameter model, derived from the Qwen/Qwen3-8B architecture. It has undergone specialized fine-tuning on the neulab-mind2web-sandboxes_glm_4.7_traces_jupiter dataset, indicating a focus on tasks related to web environments and interactions.

Key Capabilities

• Web-centric Processing: Optimized for understanding and interacting with web-based data and interfaces due to its specific training dataset.
• Qwen3-8B Foundation: Benefits from the robust capabilities of the Qwen3-8B base model, providing a strong linguistic and reasoning foundation.

Training Details

The model was trained with a learning rate of 4e-05 over 7 epochs, utilizing a multi-GPU setup with 16 devices. The training process involved an AdamW optimizer and a cosine learning rate scheduler with a 0.1 warmup ratio.

Good for

• Applications requiring automated web navigation.
• Tasks involving data extraction from web pages.
• Developing agents that interact with web interfaces.