DiffusionGemma 改造 Gemma 4 实现每秒 1500 token，但推理能力滞后
google deepmind 发布了 diffusiongemma 的技术报告，这是一个从 gemma 4 转换而来的文本扩散模型，在 h100 上可达到每秒 1,500 token 的速度。虽然它通过并行去噪带来了巨大的速度提升，但在原始推理质量上仍落后于自回归模型。
google deepmind detailed diffusiongemma, a text diffusion model retrofitted from the existing gemma 4 checkpoint rather than trained from scratch. the conversion utilizes a two-phase process combining reinforcement learning and sampler distillation to balance quality with the massive speed gains of parallel denoising. this method allows the model to generate about 1,500 tokens per second on an h100 and improve at tasks requiring bidirectional context, like sudoku or structured json output. however, the report acknowledges that quality benchmarks fall short of the autoregressive base model, partly because the architecture was not originally designed for diffusion.