●Change8

v0.1.521-beta

📦 unslothView on GitHub →
✨ 12 features🐛 31 fixes🔧 4 symbols

Summary

This release introduces support for Kimi K3 and DeepSeek v4 Flash models running locally with Unsloth Dynamic GGUFs. It also includes significant improvements to download efficiency, installation processes across different operating systems, and numerous bug fixes for the Unsloth Studio and backend.

Migration Steps

  1. Update installation scripts: macOS/Linux/WSL: `curl -fsSL https://unsloth.ai/install.sh | sh`
  2. Update installation scripts: Windows: `irm https://unsloth.ai/install.ps1 | iex`

✨ New Features

  • Run Kimi K3 & DeepSeek v4 Flash locally with Unsloth Dynamic GGUFs.
  • More efficient and faster downloading for Colab, low memory systems, and high memory/CPU systems.
  • Auto fallback to HTTP if XET is stuck.
  • Unsloth automatically detects multi-GPU setups and can offload model layers to system memory for DeepSeek V4 Flash.
  • Unsloth automatically detects multi-GPU setups and offloads layers automatically for Kimi K3.
  • Support for low, high, and max reasoning efforts for Kimi K3.
  • Use pre-built Tauri icon for macOS .app bundle.
  • Report active training status in Studio.
  • Apply saved per-model settings on API loads and add API monitor in Studio.
  • Detect an unreachable hub, not just dead DNS, for offline operations.
  • Charge the split compute-buffer rate on a multi-GPU auto fit in Studio.
  • Complete translation coverage for all supported locales in Studio.

🐛 Bug Fixes

  • Fix two pre-existing Backend CI failures.
  • Warn loudly when requested packing is silently disabled.
  • Preserve reasoning across turns for Qwen3.x.
  • Make tests/utils runnable again and cover test_packing.py in CI.
  • Uninstaller: reject unrecognized arguments instead of uninstalling.
  • Make the uninstaller test suites run on macOS and Windows PowerShell 5.1.
  • Studio: Don't disarm Ctrl+C in agents launched by unsloth start.
  • Studio: keep a slow model load alive through a proxy timeout.
  • Tighten the comments around the uninstaller argument guard.
  • Route the explicit Vulkan setup failures through the setup failure helpers.
  • Honor an explicit map_eos_token=False in get_chat_template.
  • Fix desktop repair version skew and output decoding.
  • Desktop: Improve the installation message.
  • Studio: make the API keys list scrollable.
  • Desktop: notarize the final macOS disk image on release.
  • Desktop: preserve first app window size after setup.
  • data-recipe: recognize current sampling progress log line.
  • Desktop: clean up the backend on Unix termination signals.
  • Desktop: show the app version in Settings > About.
  • Studio: Fix hub search input overlapping the format filter at narrow widths.
  • Desktop: exit the backend when the app that spawned it dies.
  • Desktop: open uploaded files via the OS instead of window.open.
  • Run OpenClaw in the launched project workspace.
  • Desktop: drop documents into a chat instead of rejecting them as non-GGUF.
  • Replace deprecated transformers attention mask imports.
  • Assert the Vulkan installer bailouts fail closed via the setup failure helper.
  • Fix multiline prompts through Windows npm shims.
  • Studio: Fix GGUF classification for custom MTP scan roots.
  • Guard VLM detection against config.architectures being None in the save paths.
  • Make the auto-appended EOS token deterministic.
  • Unsloth Studio: fix sidebar rail bottom fade rendering as a lighter block in dark mode.

Affected Symbols

unsloth-cli.pytransformers attention mask importsget_chat_templateconfig.architectures