AgPerry/Qwen3-8B-fim-v2v3pt-swe-lego-posttrain

Overview

AgPerry/Qwen3-8B-fim-v2v3pt-swe-lego-posttrain is an 8 billion parameter model built upon the Qwen3 architecture. It is a fine-tuned iteration of the /mmu-vcg-hdd/multimodal/models/Qwen3-8B-fim-v2v3pt base model, specifically adapted through post-training on specialized datasets. The training process involved a learning rate of 0.0001, a total batch size of 48, and utilized a cosine learning rate scheduler over 4 epochs.

Key Capabilities

• Specialized Fine-tuning: The model has undergone targeted fine-tuning on swe_lego_real_data_resolved_trajectories and swe_lego_synthetic_data_resolved_trajectories datasets, indicating a focus on tasks related to these specific data types.
• Qwen3 Architecture: Benefits from the underlying capabilities of the Qwen3 model family.
• Context Window: Supports a context length of 32,768 tokens, allowing for processing of relatively long sequences.

Good For

• SWE-Lego Domain Tasks: Ideal for applications requiring understanding or generation within the specific swe_lego data contexts it was trained on.
• Research and Development: Suitable for researchers exploring the impact of post-training on specialized datasets using a Qwen3 base.