CASE FILE / Games & simulation

The Jev Doom demo by @CompleteSkeptic is really cool, so while I’m waiti…

The Jev Doom demo by @CompleteSkeptic is really cool, so while I’m waiting for access I tried the same thing with Gemma 4 E2B from @GoogleGemma. Mine takes 640×480 frames directly, with no intermediate text description of the scene. It shares the image prefill across batched action queries, then reads the controls directly from logits. It sort of works.

By Simon Edwardsson @SimEdw

● ● ●CASE FILEGUIDE

Simon Edwardsson · @SimEdw

Original post

The Jev Doom demo by @CompleteSkeptic is really cool, so while I’m waiting for access I tried the same thing with Gemma 4 E2B from @GoogleGemma. Mine takes 640×480 frames directly, with no intermediate text description of the scene. It shares the image prefill across batched action queries, then reads the controls directly from logits. It sort of works.

Views 521Likes 10Reposts 6

Chinese translation

@CompleteSkeptic的Jev Doom演示非常酷，所以在我等待访问时，我在@GoogleGemma的Gemma 4 E2 B上尝试了同样的事情。我的直接拍摄640 x 480帧，没有场景的中间文本描述。它在批处理操作查询中共享图像预填充，然后直接从日志中读取控件。这有点有效。

What this case shows

Game state maps cleanly onto typed choices
Let JEV pick; keep generation elsewhere
The original post usually has a playable demo