seconds-0/qwen3-8b-bs4-rl

BS4-Qwen3-8B Step 395

Qwen3-8B fine-tuned on BeautifulSoup HTML parsing tasks using reinforcement learning.

Training Details

• Base model: Qwen/Qwen3-8B
• Training method: RL with prime-rl
• Training steps: 395
• Hardware: 2x H100 80GB
• LoRA config: rank=8, alpha=32
• Target modules: q_proj, k_proj, v_proj, o_proj, gate_proj, up_proj, down_proj

License

Apache 2.0 (same as base model)