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: re-measure the startup bundle budget, which main is red on by itself (#11013)
Daniel Hana58537ed07eb
diff --git a/studio/frontend/scripts/check-bundle-budget.ts b/studio/frontend/scripts/check-bundle-budget.tsindex 96d9e06a4e4..b27280e2323 100644--- a/studio/frontend/scripts/check-bundle-budget.ts+++ b/studio/frontend/scripts/check-bundle-budget.ts@@ -32,24 +32,14 @@ const DIST = resolve(HERE, "..", "dist");* machine.*/export const BUDGET = {- // Measured 1,496.2 KB transfer / 5,207.2 KB raw at 17363f8a2. Raised for the audio placement control: same- // build both sides, merge base 1,560.9 KB transfer against branch 1,562.6 KB, so it crossed the old 1,562.5 KB- // ceiling by a tenth of a kilobyte.- //- // Raised again after a second squeeze. main measures 1,585.6 KB transfer and passed- // 1,645,000 by 20.9 KB, which is 1.3% and about six days at the 3.7 KB/day this file- // has drifted since 17363f8a2. That is the same margin the previous raise left, and it- // lasted six days before main went red again on a commit whose only frontend change was- // one English sentence rewritten shorter.- //- // Nothing became eager that should not be: across the whole window the eager chunk set- // gained no member, so there is nothing here to lazy-load. What runs out is headroom.- // 1,690,000 leaves 64.8 KB (4.1%), the proportion #8964 shipped with, which absorbed 17- // days. rawBytes stays put at 64.7 KB spare so both halves come up for one re-measure- // together instead of each dragging main red on its own.- transferBytes: 1_690_000,- // same-toolchain merge comparison: 5381.0 to 5384.0 kib raw, with 84 eager chunks.- rawBytes: 5_520_000,+ // Measured, one machine and one build per side, so the pair is comparable to itself rather+ // than to a runner's: 5,384.0 KB raw / 1,610.2 KB transfer at the merge base, leaving+ // 211.7 KB and 64.6 KB spare. Both halves are re-measured TOGETHER, or each drags main red+ // on its own; the previous four raises each bought a few kilobytes and were spent within+ // days, charging the next PR for drift it did not cause. The eager set has not gained a+ // member across any of it, so what runs out is headroom, not laziness.+ transferBytes: 1_715_000,+ rawBytes: 5_730_000,};// The chunk count is reported but not budgeted. Splitting a page out of the entry raises it while lowering the