Local UI to run and train LLMs and diffusion models. Supports GGUF, MLX, Qwen3.8, DeepSeek-V4.1, MiniMax-H3, Gemma 4, FLUX and more.
Python agentaichatgptdeepseekfine-tuninggemmaimage-generationllamallmllmsopenaipythonqwenreinforcement-learningself-hostedstable-diffusiontext-to-speechttsuiunsloth
Studio: nudge the dictate glyph down 1px instead of 2px (#13504)
Michael Han3cea7ec89298
diff --git a/studio/frontend/src/index.css b/studio/frontend/src/index.cssindex 6d7113322f0..b00ba0c87c0 100644--- a/studio/frontend/src/index.css+++ b/studio/frontend/src/index.css@@ -5750,7 +5750,8 @@ html.force-reduced-motion .sidebar-section-drop-line {& svg.unsloth-dictate-icon {width: calc(var(--ui-icon-size) * 1.2);height: calc(var(--ui-icon-size) * 1.2);- transform: translateY(2px);+ /* Optical nudge for the top-heavy capsule. 2px read as low, most of all in the 28px circle. */+ transform: translateY(1px);}/* Flex centers both symmetric glyphs; pixel offsets vary by display scale. */& svg.unsloth-send-icon {diff --git a/studio/frontend/tests/composer-icon-alignment.test.ts b/studio/frontend/tests/composer-icon-alignment.test.tsindex d48f55e1459..775dd97461c 100644--- a/studio/frontend/tests/composer-icon-alignment.test.ts+++ b/studio/frontend/tests/composer-icon-alignment.test.ts@@ -52,3 +52,9 @@ test("neither composer action glyph carries a device-pixel nudge", () => {assert.doesNotMatch(source, RETIRED_CANCEL_CLASS);}});++test("the dictate glyph keeps a 1px optical nudge, not the 2px that read as low", () => {+ const body = css.match(/& svg\.unsloth-dictate-icon \{([\s\S]*?)\n\t\}/)?.[1];+ assert.ok(body, "dictate icon rule");+ assert.match(body, /transform:\s*translateY\(1px\);/);+});