Let’s drink to love.. Liang Wenfeng.
A fun little game i made with Qwen3.8-27b-Q4 with DeepSeek Harness as driver.
Wanted to see how good it is, so decided to lazy prompt mixing in Pagoda world + fps game. It's amazing how much visually stunning world is. DeepSeek Harness continued working on this untill the game
DeepSeek harness + Qwen3.8-27B
With Open-webui, it became slow due to longcontext, but DSH is fine. Agent behavior is also easy to see.
Today we're launching Miles v0.1, an open-source RL framework for LLMs and multimodal models.
RL training is easy to start and hard to debug. Miles helps you ensure your run is correct, use hardware efficiently, and keep RL running at scale. Over the past 9 months, 72
You have reached the end of the archive
All of deepseek