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
// SPDX-License-Identifier: AGPL-3.0-only
// Copyright 2026-present the Unsloth AI Inc. team. All rights reserved. See /studio/LICENSE.AGPL-3.0
import assert from "node:assert/strict";
import { readFileSync } from "node:fs";
import path from "node:path";
import { test } from "node:test";
import { fileURLToPath } from "node:url";
const HERE = path.dirname(fileURLToPath(import.meta.url));
const read = (rel: string) => readFileSync(path.join(HERE, "..", rel), "utf8");
const ADAPTER = read("src/features/chat/api/chat-adapter.ts");
const RUNTIME = read("src/features/chat/hooks/use-chat-model-runtime.ts");
test("a startup auto-load sends the remembered split for the GPUs it resolved", () => {
assert.match(
ADAPTER,
/reconcileTensorSplit\(\s*config\.tensorSplit,\s*config\.selectedGpuIds,\s*effectiveGpuIds,\s*\)/,
);
assert.match(ADAPTER, /tensor_split: effectiveTensorSplit \?\? undefined,/);
assert.doesNotMatch(ADAPTER, /split ratio is never remembered/);
});
test("a picker load sanitizes its stored extra args before sending them", () => {
const start = RUNTIME.indexOf(
"let loadLlamaExtraArgs = pendingLoadConfig?.llamaExtraArgs;",
);
assert.ok(start > 0);
const block = RUNTIME.slice(start, start + 1400);
assert.match(block, /isGguf &&\s*!targetIsDiffusion &&/);
assert.match(
block,
/loadLlamaExtraArgs != null &&\s*loadLlamaExtraArgs\.length > 0/,
);
assert.match(block, /await loadManagedLlamaFlags\(\)/);
assert.match(
block,
/sanitizeStoredExtraArgs\(\s*loadLlamaExtraArgs,\s*managed\?\.managed \?\? new Set<string>\(\),\s*\{\s*maxBytes: managed\?\.maxBytes,\s*windowsCommandBudget: managed\?\.windowsCommandBudget,/,
);
assert.match(block, /loadLlamaExtraArgs = cleaned;/);
});