QwenDLM-4B
A diffusion language model converted from Qwen3-4B, used to test diffusion LLM acceleration techniques.
Timeline (last 90 days)
According to the author's report in the Reddit post, ORL increases the number of tokens per forward pass of QwenDLM-4B by up to 6.2 times, improves end-to-end throughput by about 6.9 times, and raises HumanEval pass@1 from 82.9% to 89.6%.
On a single H100 and at similar measured accuracy, QwenDLM-4B + ORL reaches 1,298 tokens/s, DFlash 746 tokens/s, and EAGLE3 405 tokens/s (end-to-end system comparison: ORL runs on dlmrt, speculative decoding runs on vLLM).
All Facts (2)
According to the author's report in the Reddit post, ORL increases the number of tokens per forward pass of QwenDLM-4B by up to 6.2 times, improves end-to-end throughput by about 6.9 times, and raises HumanEval pass@1 from 82.9% to 89.6%.
50%UnverifiedOn a single H100 and at similar measured accuracy, QwenDLM-4B + ORL reaches 1,298 tokens/s, DFlash 746 tokens/s, and EAGLE3 405 tokens/s (end-to-end system comparison: ORL runs on dlmrt, speculative decoding runs on vLLM).
50%