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
Raise the startup bundle budget for the FLUX.3 controls (#14312) (#14319)
Daniel Han360cd9ace2fb
diff --git a/studio/frontend/scripts/check-bundle-budget.ts b/studio/frontend/scripts/check-bundle-budget.tsindex f7410578e43..3436c6b9606 100644--- a/studio/frontend/scripts/check-bundle-budget.ts+++ b/studio/frontend/scripts/check-bundle-budget.ts@@ -287,8 +287,12 @@ export const BUDGET = {// chunk. Raised to the measured build plus the same 160 B raw / 189 B transfer margin.// Reconciled with main at 87c6408455: the combined build measures 6,844,101 B raw / 2,060,934 B transfer// (Node 24, 96 chunks). The API Settings button stays in its lazy page. Keep the same 160 B / 189 B margin.- transferBytes: 2_061_123,- rawBytes: 6_844_261,+ // FLUX.3 workflows, controls and LoRAs (#14312) add 426 B raw / 138 B transfer (6,844,101 -> 6,844,527 B raw,+ // 2,060,887 -> 2,061,025 B transfer against 6a19cb20b9, one machine and build, Node 24, 96 chunks), and #14228+ // adds 95 B raw / 15 B transfer on top, measured at d3a82576a4 as 6,844,622 B raw / 2,061,040 B transfer. Raised+ // to that build plus the same 160 B raw / 189 B transfer margin.+ transferBytes: 2_061_229,+ rawBytes: 6_844_782,};// The chunk count is reported but not budgeted. Splitting a page out of the entry raises it while lowering the