---
base_model: Qwen/Qwen3-4B-Base
library_name: transformers
tags:
- qwen3
- spsd
- sft
- board-games
---

# SPSD-RL

Qwen3-4B-Base full-parameter checkpoint trained with prompt/completion
supervision on the `LorMolf/SPSD-RL` conversation dataset.

Source artifact:
`outputs/qwen3_4b_base_spsd_rl_sft_prompt_completion_4gpu_20260603/final_bs20_accum4_ddp7200_wandb_localcache`.