unslothai/unsloth v0.1.903-beta on GitHub

This release adds a browser inside Unsloth (browser use coming very soon), so files, web pages and pages the model writes open right beside your chat. It also brings EmbeddingGemma 2, Google's new multimodal embedding model, plus new Audio pages and lower-memory training.

Highlights

Browser beside chat for files, web pages and model-written HTML, as tabs
EmbeddingGemma 2: run and fine-tune Google's new model for text, code, images, video and audio Guide
Audio pages: Speak, Clone, Music, Transcribe, Edit, Separate, Convert, with 70+ GGUF models
Keep multiple models loaded, plus vLLM and SGLang engines (opt-in)
Sandbox for code the model runs, now with a Settings tab and Colab and Docker support
Live preview during image and video generation

Browser

Unsloth now has a built-in browser panel next to any chat. Attached files, web pages and HTML the model writes all open in it as tabs.

Files open as tabs: PDF, Office, HTML, Markdown, code and text.
Real browsing: address bar, search, tabs, find in page, zoom, bookmarks, history and downloads.
Model-written HTML and code open as tabs with a preview, console and Request edits, replacing Canvas.
Ask about this page: highlight part of a page and ask the model about it. Right-click a tab to pin it or fork it into a chat.
Settings > Browser: pick a search engine, set the default zoom, and import bookmarks from Chrome, Safari, Firefox or Edge.
Kept separate: web pages cannot reach your Unsloth account, chats or API.

EmbeddingGemma 2

EmbeddingGemma 2 is Google DeepMind's new multimodal embedding model. It turns text, code, images, video and audio into vectors for search, RAG, classification and clustering. Read our guide.

740M parameters, or 270M for text only, with optional vision and audio encoders. 100+ languages and an 8K context window.
About 14% better on code than EmbeddingGemma 1 (78.68 vs 68.76, MTEB code).
Run it locally: import GGUF or safetensors into Unsloth, or serve text embeddings with llama.cpp. GGUF safetensors
Fine-tune the text model with FastSentenceTransformer, using 4-bit QLoRA or BF16 LoRA.
Prompts are kept. Fine-tunes of EmbeddingGemma and Qwen3-Embedding now save the model's built-in query and document prompts, not empty ones.

Audio

Clone a voice from a short sample and save it.
Edit words in a recording and keep the speaker's voice.
Separate a song into vocals, drums and other parts, or Convert one voice into another.
Music makes songs and sound effects, and can edit parts of a clip.
Transcribe can show timestamps and speakers, and exports subtitles.

Sandbox

Code the model runs with the Python and Terminal tools stays inside a sandbox, with limited access to your computer.

OS sandbox on Linux and macOS, plus an opt-in Windows 11 preview. Now works in Colab and Docker.
Settings > Sandbox shows this computer's protection and can install a missing sandbox.
Low or High: High (default) uses the OS sandbox. Low uses software checks only.

Models + chat

Keep multiple models loaded (off by default, in Settings > Resources). Each request is answered by the model it asks for.
vLLM and SGLang: opt-in, experimental engines for Linux with NVIDIA GPUs.
Auto-compact now works for API models like Claude and OpenAI, and for MLX chats.
Import your Cursor, Claude Code and Codex chats into projects, or chats saved as Markdown.
unsloth start --app lets the Codex, OpenCode, OpenClaw and Hermes desktop apps use your local model.

Training + speed

Layer offload (opt-in, LoRA, NVIDIA and AMD graphics cards): offload_layers keeps frozen layers in system RAM, so Llama-3.1-8B 4-bit trains with VRAM capped at 7 GiB.
Faster padding-free training without flash-attn, the usual Colab case: Qwen3-0.6B drops from 5.45 GB to 1.87 GB of VRAM, with steps 5x to 8x faster.
gpt-oss-20b QLoRA is 2x to 3x faster on NVIDIA.
Images and video (NVIDIA): live preview while rendering, up to 20x faster Z-Image-Turbo on older GPUs, and the same seed gives the same image after a restart.
AMD: export fine-tuned models for the Ryzen AI NPU, and a much faster first Qwen-Image-2.1 image on newer Radeon cards.
Faster on Apple Silicon (MLX): Qwen3.6-35B-A3B replies about 10% faster, and long prompts start up to 10% sooner.
Int8 Prefill on M5 chips (MLX, experimental, opt-in): long prompts read up to 1.5x faster, with slightly less accurate answers.

What's Changed

feat(studio): complete embedded image recipe and Recipe popover by @Souravrajvi0 in #12091
Bump install.sh / install.ps1 pin to unsloth>=2026.9.14 by @danielhanchen in #12439
Count the final EOS token within the raw-text chunk budget by @vineethsaivs in #11040
fix(studio): serve desktop SPA on loopback listener by @Souravrajvi0 in #10847
Format the desktop contract test to the ruff-format fixed point by @danielhanchen in #12440
Fix three tests red on main after #12431, #12408 and #12382 by @danielhanchen in #12441
Studio: disclose required assets before media downloads by @Imagineer99 in #12387
Stop enable_padding_free_metadata writing seq_lengths into the caller's examples by @vineethsaivs in #8049
Studio: persist Data Recipes and run history server-side by @Etherll in #7108
Fix the tests left red on main by #12408, #12382 and #12436 by @danielhanchen in #12447
Read supervision from labels or the mask column in the padding-free filter tests by @danielhanchen in #12452
fix(studio): include loaded llama extra args in active model baseline by @Souravrajvi0 in #10870
Read auto_map at every config level in the remote code gate by @danielhanchen in #12453
Record why the permission pill is missing after a reload, and reload once more only if the app never booted by @danielhanchen in #12456
Add approved image attachment inputs for MCP tools by @Etherll in #10871
Studio: log a pending embedder download once instead of a traceback on every compaction by @danielhanchen in #12443
Studio: stop a recovered chat crashing on a duplicate tool call key by @danielhanchen in #12442
Give the NVIDIA probe control case a budget a loaded runner can meet by @danielhanchen in #12462
Wait for the chat-only export gate instead of counting it as the form appears by @danielhanchen in #12463
Pin the DeepSeek OCR module fetch and import it from that fetch by @danielhanchen in #12454
Check a sentence-transformers module config on every version, and confirm the resolved class by @danielhanchen in #12444
Studio: keep the app's titlebar chrome off the desktop update screen by @oobabooga in #12465
Studio: stop a reopened chat from saving every replayed event, and let Stop reach it by @oobabooga in #12464
Studio: blur modal titlebar background while keeping window controls sharp by @Imagineer99 in #12433
Gate the sentence-transformers module types on the routes that delegate the load by @danielhanchen in #12446
Retry the NVIDIA probe control case past the Linux pwsh redirect race by @danielhanchen in #12477
Studio: keep diffusion LoRA loading working on torchao 0.18 with peft 0.18 by @danielhanchen in #12494
Studio: retry Lemonade on a new port when its free port is taken before it binds by @danielhanchen in #12498
Studio: move Butterfly Pea and Earl Grey up the More themes list, and rename Cherry Cola to Cherry by @shimmyshimmer in #12500
fix(studio): prevent IME confirmation from submitting chat renames by @kei-yamazaki in #12475
Wait for vite to serve before the titlebar, find, shortcuts and settings smokes navigate by @danielhanchen in #12501
Studio: top-level Models block group in picker by @Souravrajvi0 in #10736
fix(studio): keep distinct symlink aliases for per-model settings by @Souravrajvi0 in #10644
Studio: JSON and Markdown validator blocks by @Souravrajvi0 in #10710
fix(studio): reconcile externally updated saved assistant messages by @Souravrajvi0 in #11500
Studio: save a merged Whisper export that can be loaded again by @NilayYadav in #12483
fix(studio): allow Enter to send with idle macOS Pinyin IME by @Souravrajvi0 in #12138
Studio: keep a Data Recipes publish private when the Hub repo already exists by @NilayYadav in #12482
Baseline the nine unsloth-zoo 2026.9.9 findings after review by @danielhanchen in #12515
Studio: put llama-server warnings and errors in the default session log by @Souravrajvi0 in #10848
Studio: stop leftover dataset columns becoming the system prompt by @NilayYadav in #12485
Studio: restore markdown chat import and bulk export by @Souravrajvi0 in #12513
Add install_missing_dependencies to opt out of the FP8/FP4 llm-compressor auto-install by @Souravrajvi0 in #9405
perf(studio): enter the MLX routed-experts and norm-handoff fusions per request by @Lyxot in #12422
Feat/UI rework by @Sneakr in #12479
Studio: run MiniMax-H3's Diffusers INT8 path on 24 / 16 / 12 GB cards and near the host RAM floor by @danielhanchen in #12409
Studio: extend the ROCm MIOpen cutoff and Qwen-Image-2.1 query chunking beyond gfx1151 by @oobabooga in #12471
fix(studio): block-split long backslash lines without inline lex (#11376) by @Souravrajvi0 in #11501
Studio: train vision models on images in a list column or inside the messages by @NilayYadav in #12484
Studio: resume a reply that stopped mid-thought on MLX by @Lyxot in #12504
Studio: canonicalize and validate web_search arguments by @Souravrajvi0 in #9716
Studio: optional systemd user service for Linux auto-start by @Souravrajvi0 in #9308
SentenceTransformer: opt-in encoder unpadding via shared attention dispatch by @Etherll in #4460
Studio: live dataset preview during active runs by @Souravrajvi0 in #10737
Core: pip extras and auto-install support for torch 2.13 and 2.14 by @danielhanchen in #12151
Studio: resume NPU model downloads, keep their progress, and list NPU models in the Hub by @oobabooga in #12512
Name the matching build when flash-attn or causal-conv1d was built for another torch by @danielhanchen in #12211
Studio: fused int8 GEMM with the dequant epilogue, bf16 out (Qwen-Image-2.1 14-17% faster per step) by @danielhanchen in #12448
Studio: wait for the torch warm's dynamo import before a load's first download by @danielhanchen in #12472
Studio: faster cold start for image and video loads (first image on H100 7 to 37% sooner) by @danielhanchen in #12496
Studio: let the compile cache keep denoiser graphs with sdpa_kernel blocks (restart first render 1-5.5 s faster) by @danielhanchen in #12510
Studio: keep GGUF weights and the last-run module on their host copy under whole-model offload by @danielhanchen in #12022
Studio: torch 2.13 for new Linux cu130 Python 3.13 installs, existing installs keep their torch by @danielhanchen in #12150
Studio: plan image offload on measured activations so more VRAM is never slower by @danielhanchen in #12043
studio: reuse ssl setup for local gguf requests by @mahiatlinux in #12495
Studio: run FLUX.2 RoPE as one fused kernel on fp16 GPUs (klein step 5% faster on T4, bit-identical) by @danielhanchen in #12503
Studio: load FLUX.1 and Qwen-Image on fp16-only GPUs with small host RAM by @danielhanchen in #12480
Studio: run Z-Image and Wan2.2-5B / HunyuanVideo-1.5 in fp16 on fp16-only GPUs instead of fp32 by @danielhanchen in #12476
Studio: faster first renders and resident text encoder under group offload by @danielhanchen in #12519
Studio: make unsloth start sampling and reasoning flags apply only to that agent by @NilayYadav in #12493
Studio: stop adding backslashes to Word files in Data Recipes by @NilayYadav in #12488
Studio: keep tool-call arguments when training on tool-calling datasets by @NilayYadav in #12487
Studio: keep and Vec text in research reports and file previews by @NilayYadav in #12489
Studio: keep Word equations when a .docx is added to a knowledge base or chat documents by @NilayYadav in #12490
Studio: keep MiniMax-H3 GGUF renders on a resident sd-server, released when idle by @danielhanchen in #12461
Studio: load the cached int8 checkpoint for a 5-bit-or-below GGUF pick that has to offload (Qwen-Image-2.1 about 2x faster at 12 / 8 GB) by @danielhanchen in #12455
Studio: log the real reason the Decision API model failed to load by @NilayYadav in #12491
Studio: stop CLI commands running as another account when Studio has more than one by @NilayYadav in #12492
Studio: train a dataset's system column as the system prompt by @NilayYadav in #12486
Studio: add per-model custom llama.cpp INI configuration by @Etherll in #10783
Fused Triton NF4 dequant and GEMV: bit-exact, stream-safe, torch.compile traceable by @danielhanchen in #12113
Make RMSNorm, RoPE and the input-embedding hook traceable by torch.compile by @danielhanchen in #12171
Studio: add opt-in vLLM and SGLang support with multi-GPU inference, quantization and vision by @oobabooga in #11491
Lift the transformers ceiling to 5.17.0 and align the mirrored CI caps by @danielhanchen in #11016
Studio: run vLLM and SGLang on Windows inside a private WSL2 distro by @danielhanchen in #12024
Studio: add audio.cpp as a native engine for speech, music and dictation by @Etherll in #12342
Studio: draw the command palette like chat search by @shimmyshimmer in #12511
Studio: center the sidebar row kebab and pin in their hover circle by @shimmyshimmer in #12502
Studio: clip every rounded scroller in Firefox, dropping the overflow watcher by @shimmyshimmer in #12438
Studio: keep the composer permission shield still while the plus spins by @shimmyshimmer in #12506
Parity: keep the reap-margin check off the edge of its own window by @lonexreb in #12514
Studio: use ComfyUI's default settings for FLUX.1, Qwen-Image, Z-Image, Ideogram 4, Wan2.2 and HunyuanVideo-1.5 by @danielhanchen in #12516
Studio: move torchao int8 weights back to the GPU after an oversized request streams pinned groups by @danielhanchen in #12526
Count only the lock's own waits in the unlockable-filesystem row by @lonexreb in #12529
Studio: renew the chat-run lease while long prefill is still advancing by @indrajeetapache in #12523
Studio: stop rejecting --mmproj-device CUDA1 when gpu_ids are saved by @1fanwang in #12505
fix(studio): WSL2 Windows localhost hint in startup banner (#11187) by @Souravrajvi0 in #11361
Studio: lift a carried model picker row like a sidebar chat, and lighten both drag copies by @shimmyshimmer in #12499
Postpone annotation evaluation in engine_compat.py so main's PEP 604 ratchet holds by @danielhanchen in #12556
Keep the MLX inference test stub neutral for every helper zoo enters around a model by @danielhanchen in #12548
Re-measure the Studio startup budget after the managed engine options by @danielhanchen in #12560
Studio: nudge sidebar row pin and options buttons 3px right by @shimmyshimmer in #12563
Studio: make carried rows a faintly frosted, darker copy by @shimmyshimmer in #12562
Check the sidebar pin follows the UI scale by property, not by its pinned length by @danielhanchen in #12597
Studio: search chats, projects, files and models from tabs in chat search by @shimmyshimmer in #12564
Revert 10 PRs merged before Codex review converged by @Etherll in #12569
Studio: cleaner Deep research limits dialog by @shimmyshimmer in #12632
Studio: Skills dialog scrolls on its right edge by @shimmyshimmer in #12634
Studio: Skills pill in the composer for quick toggling by @shimmyshimmer in #12631
Studio: shorter permission menu, Learn more, calmer Full access by @shimmyshimmer in #12630
Studio tests: pin Z-Image's fp16 promotion in the calibrated-activation test by @danielhanchen in #12570
Studio: keep text placed over pictures when indexing a PDF by @NilayYadav in #12577
Parity: keep the reap-margin check off the edge of its own window by @danielhanchen in #12615
Count only the lock's own waits in the unlockable-filesystem row by @danielhanchen in #12618
Studio: stop blaming Max Tokens when a connected model fills its context window by @oobabooga in #12567
Studio: keep a tool call's answer after its result in JSONL chat exports by @NilayYadav in #12579
Studio: stop unsloth start openclaw cutting every reply at 8,192 tokens by @NilayYadav in #12573
Studio: stop New chat from requesting a thread row that does not exist yet by @oobabooga in #12539
Studio: stop training Gemma 4 on tool results by @NilayYadav in #12575
Studio: say so when the Audio page stops speech at Max tokens by @NilayYadav in #12583
Studio: lift a carried model picker row like a sidebar chat, and lighten both drag copies by @danielhanchen in #12622
Studio: show NPU reply speeds and configure NPU models before loading by @oobabooga in #12595
Studio: consent dialog before FP8/FP4 llm-compressor install (Phase 2) by @Souravrajvi0 in #9554
Studio: train a dataset's system column as the system prompt by @danielhanchen in #12614
Studio: stop rejecting --mmproj-device CUDA1 when gpu_ids are saved by @danielhanchen in #12620
fix(studio): prevent IME confirmation from submitting chat renames by @danielhanchen in #12613
Chat UI: find the Full access consent dialog by its slots, not its wording by @danielhanchen in #12640
Studio: read text files saved in older Windows encodings correctly by @NilayYadav in #12581
Studio: use the reply on screen when turning chats into training data by @NilayYadav in #12584
fix(studio): WSL2 Windows localhost hint in startup banner (#11187) by @danielhanchen in #12621
Studio: let API requests that ask for a JSON reply still call their tools by @NilayYadav in #12578
Studio: keep the VRAM figure visible on the Run preview hardware row by @LeoBorcherding in #12545
Unsloth Studio / Desktop: keep release-notes table links from splitting mid-word in the update popup by @LeoBorcherding in #12543
Studio: keep the prompt queue's more menu open while queued prompts send by @LeoBorcherding in #12546
Let agent desktop apps use the model Unsloth is serving by @NilayYadav in #12582
Studio: MiniMax-H3 GGUF uses the BF16 cuBLAS matmul path by default on sm80+ (1.2 to 2.3x per step, closer to an F32 reference) by @danielhanchen in #12527
Studio: renew the chat-run lease while long prefill is still advancing by @danielhanchen in #12619
Studio: use ComfyUI's default settings for FLUX.1, Qwen-Image, Z-Image, Ideogram 4, Wan2.2 and HunyuanVideo-1.5 by @danielhanchen in #12616
Studio: keep earlier tool calls in chat history on safetensors and MLX models by @NilayYadav in #12574
Unsloth Studio / Desktop: stop spam-clicking sidebar rows from queueing a navigation per click by @LeoBorcherding in #12544
Studio: stop the Load Model panel flagging Auto context loads as over VRAM by @oobabooga in #12555
Windows ROCm: use math attention where the fused SDPA kernels fail by @danielhanchen in #12639
Studio: stream a torchao denoiser from an unpinned host copy instead of leaving it on the GPU by @oobabooga in #12551
Point five mapper rows at their upstream repo instead of an unpublished unsloth 16bit name by @LeoBorcherding in #12587
Studio small-host route: prefetch the streamed text encoder and fuse the int8 dequant (T4 FLUX.1-schnell 5.19 to 4.49 s per step, pixel-identical) by @danielhanchen in #12532
studio: allow full access for the installation owner by @mahiatlinux in #12604
Keep the load-time gradient checkpointing mode in for_training by @danielhanchen in #12629
Studio: move torchao int8 weights back to the GPU after an oversized request streams pinned groups by @danielhanchen in #12617
Studio: decode a resident Wan2.2-TI2V-5B VAE untiled when it fits by @danielhanchen in #12603
Studio: keep a load's diffusers / peft import off the post-warm import by @danielhanchen in #12572
Studio: pin inductor's dynamic_scale_rblock off so compiled renders match across servers by @danielhanchen in #12588
Studio int8 GEMM: stop the device probe from reserving 0.7 to 1.4 GB of spill memory by @danielhanchen in #12525
Studio: keep dark ink visible in transparent source images by @NilayYadav in #12576
Studio: make the fused int8 kernels bit-exact on torchao 0.17 and round int32 to bf16 twice like PyTorch by @danielhanchen in #12633
Add a fast lint gate for risky loader call sites by @danielhanchen in #12636
Studio: fused fp16 Wan block kernels on T4 (Wan2.2-5B 11.7 to 10.0 s per step, within base's spread) by @danielhanchen in #12531
Studio: serve multiple models at once by @NilayYadav in #11591
Fix text-only VLM decoder saves reloading with random weights by @1fanwang in #12561
Studio: hold cudnn.benchmark off for Wan so renders are reproducible across servers by @danielhanchen in #12565
Studio: hold cudnn.benchmark off for Qwen-Image, HunyuanVideo-1.5, FLUX.1, Z-Image and SDXL by @danielhanchen in #12568
Studio: MiniMax-H3 attention fast path and fused q/k norm + RoPE (5-7% faster per step) by @danielhanchen in #12459
Studio: MiniMax-H3 int8 Linears take the fused-dequant int8 GEMM (A100 9% faster per step, ahead of ComfyUI) by @danielhanchen in #12507
Studio: fuse MiniMax-H3's ConvRot rotation into the int8 activation quant by @danielhanchen in #12517
Keep torch deprecation warnings at the caller through the getattr wrapper by @danielhanchen in #12635
Studio: let small models read a long attached file when asked to summarize it by @NilayYadav in #12580
Studio: engage the fused int8 GEMM when the placement pins the whole denoiser, and restore torchao weights after an oversized request by @danielhanchen in #12524
studio: let vision models view workspace images in code mode by @mahiatlinux in #12591
Studio: keep the whole int8 Qwen-Image-2.1 denoiser resident at 12 GB, released only while the encoders run (298 to 212 ms per step) by @danielhanchen in #12536
Record the Wan fused block's import_module in the risky loader baseline by @danielhanchen in #12643
Studio: stop Whisper dropping sentences from clips longer than 30 seconds by @NilayYadav in #12481
Accept a list train_dataset again on TRL 1.10+ (fixes vision notebooks) by @danielhanchen in #12628
Unsloth Studio: export or convert a GGUF to FastFlowLM Q4NX for the AMD Ryzen AI NPU by @LeoBorcherding in #12541
Studio: prefetch an image load's weights and upload host tensors through a pinned ring by @danielhanchen in #12590
Studio: rounder, roomier toasts by @shimmyshimmer in #12655
Follow the multi-model refactor in two guards that went red on main by @danielhanchen in #12653
Harden torch.export .pt2 loading against CVE-2026-4538 by @danielhanchen in #12657
Guard against transformers config and chat template CVEs on older versions by @danielhanchen in #12658
Studio: bump PyJWT, urllib3, Pillow and frontend overrides for security advisories by @danielhanchen in #12656
Studio: gate repo-hosted embedder modules, MLX model_file, and _socket by @danielhanchen in #12659
Harden npm and yarn installs against supply-chain attacks by @danielhanchen in #12660
Studio: encode video mp4 with x264 frame threads (export 1.25 to 2.5x faster, same settings) by @danielhanchen in #12648
Studio: keep managed accounts off owner files and device nodes by @danielhanchen in #12661
Studio: stop the Windows event loop spinning when its self-pipe is closed by @Etherll in #12665
Studio: skip the unused Qwen-Image-2.1 text-encoder lm_head (0.6 to 1.3 GiB less VRAM, bit-identical images) by @danielhanchen in #12669
Unsloth Studio: managed accounts keep their model settings and pins across an account switch by @LeoBorcherding in #12549
Keep embedding optimizer state 32-bit when embedding_learning_rate is set by @vineethsaivs in #12458
Studio: never fail, slow down or render noise on an explicit SageAttention or FlashAttention 4 request by @danielhanchen in #12641
Studio: SageAttention 2 from the kernels hub, and FlashAttention 4 dependencies that load on a fresh install by @danielhanchen in #12654
Studio: load the hosted FP8 checkpoints for FLUX.1-dev, FLUX.2-klein-4B, FLUX.2-dev and Qwen-Image-2512, and seed Krea 2 by @danielhanchen in #12649
Studio: stream offloaded denoiser blocks without host waits (B200 FLUX.1 16 GB 0.59x, L4 0.90x per step) by @danielhanchen in #12664
Studio: pin inductor's reduction configs for LTX-2 compiles so renders match across servers by @danielhanchen in #12651
Studio: set the ROCm AOTriton opt-in from the inference package so every entry point gets fused attention by @danielhanchen in #12668
Studio frontend: bump hono, ip-address, qs, dompurify and @babel/core for security advisories by @danielhanchen in #12679
Studio: clarify image preview loading and retry failed downloads by @Imagineer99 in #12450
Studio (AMD, Windows): pin the multi-arch ROCm torch to rocm7.14.0, whose fused attention works by @danielhanchen in #12670
fix: preserve activation QAT in LoRA MLPs by @taking-lying-flat in #12558
Installers: resolve studio from the venv, never from the caller's directory by @danielhanchen in #12682
Studio: read pre-quantized diffusion checkpoints from safetensors on torchao 0.17, 0.18 and main by @danielhanchen in #12645
Studio: give the MiniMax-H3 sd.cpp engine a CUDA-12 cuDNN for fused attention on sm80+ Linux by @danielhanchen in #12662
Apply the DoRA magnitude in fast_linear_forward by @vineethsaivs in #12642
Studio: faster first start for image loads (VAE kernel prebuild, early quant probe, one FLUX.1 single-block graph) by @danielhanchen in #12666
Studio: do not reserve the OS share twice on Linux ROCm APUs by @danielhanchen in #12675
Studio: honour an explicit flash attention request on ROCm (gfx1151 8 to 12% faster per step) by @danielhanchen in #12667
Studio projects: folder-plus icon for sources by @shimmyshimmer in #12685
Studio: channels_last for 3D-conv VAEs and untiled A14B decode by @danielhanchen in #12672
[XPU] Enable FP8 training on XPU by @JoshuaL3000 in #12535
Studio: per-model automatic step skip at each model's measured step count (1.44 to 1.81x, MiniMax-H3 on max) by @danielhanchen in #12652
Studio: deterministically load explicitly mentioned skills before generation by @wasimysaid in #12538
Studio: run fp16 on ROCm GPUs without native bf16 (RDNA2 and older, Vega) by @danielhanchen in #12676
fix(studio): discover On Device GGUFs from disk before Hub by @Imagineer99 in #12460
docs(save): correct save_method values in docstrings (16bit/4bit -> merged_16bit/merged_4bit) by @simpleqt in #10394
Studio: deterministic FLUX.1, Z-Image and Qwen-Image renders on sm80, sm89 and sm120 by @danielhanchen in #12674
Studio: async block prefetch for MiniMax-H3's streamed denoiser by @danielhanchen in #12683
Studio: fused int8 GEMM for block-streamed image denoisers by @danielhanchen in #12677
Block Swap: Stream frozen transformer blocks from host RAM so a dense model can train past the card's VRAM by @LeoBorcherding in #11832
Studio: bf16 Wan VAE decode on ROCm (gfx11 / gfx12) by @danielhanchen in #12681
Studio: split Audio into Speak, Music and Transcribe workspaces by @Etherll in #12600
Studio: add the Clone page, audio inputs and saved voices by @Etherll in #12602
Studio: add timestamps, speakers and exports to the Transcribe page by @Etherll in #12607
Studio: add the Music studio with song, sound effect and edit modes by @Etherll in #12609
MLX inference test stub: accept any arguments a zoo helper is called with by @danielhanchen in #12700
Studio: add the Separate audio workflow with a synced stem mixer by @Etherll in #12610
Studio: add the Edit speech page by @Etherll in #12606
Studio: add a Convert audio page for voice conversion with A/B compare by @Etherll in #12611
Bump Desktop crates, Docker JupyterLab, setuptools and CI pins for advisories by @danielhanchen in #12704
Bump yanked chacha20 to 0.10.2 in the Desktop lockfile by @danielhanchen in #12712
Rename block_swap_layers to offload_layers by @danielhanchen in #12705
Studio: request stream usage from llama.cpp connections so the context bar fills by @jayzhou2309 in #12688
fix: preserve shared upstream gradients in RoPE backward by @taking-lying-flat in #12528
fix: preserve logits in fast cross entropy backward by @taking-lying-flat in #12351
Studio: keep FlashAttention 4 inside a fullgraph compile by @danielhanchen in #12686
Studio: fade Run settings at the edges it scrolls past by @shimmyshimmer in #12722
Studio: panel-left / panel-right icons for open sidebars, API and home-wifi icons by @shimmyshimmer in #12720
fix: allow attachment-only messages in prompt queue by @Biotrioo in #9215
fix(studio): deduplicate local models by path by @Biotrioo in #9169
fix(studio): serialize SQLite connection closes (#10022) by @Biotrioo in #10069
fix: correct easiet typo, README grammar, and duplicate issue-template frontmatter by @simpleqt in #10392
Migrate Anthropic smoke probes to v1 by @pascalandr in #9468
Allow compiled decode for gpt-oss by @danielhanchen in #12713
Studio: avoid duplicate llama-server option families by @Apoze in #9621
Revert "Studio: add per-model custom llama.cpp INI configuration (#10783)" by @danielhanchen in #12725
Studio: recount tokens when status sync adopts a resident GGUF by @claxman in #10467
Chat: show the Thinking control for Ollama models that report the thinking capability by @lonexreb in #10043
Studio: stop writing the date into user messages by @NilayYadav in #12699
tests(version_compat): AST-based symbol checks and per-model drift coverage by @Suchitra-idu in #6740
Prepare an exhausted call before the budget gate, so its replay parses by @lonexreb in #10274
Studio: in-app browser panel for chat by @shimmyshimmer in #12347
Support pre-registered MCP OAuth clients by @ousamabenyounes in #7665
docs: add 4GB VRAM QLoRA fine-tuning guide for consumer GPUs by @KafKafrnZ in #7605
feat: add support for Phi-4-multimodal-instruct by @Rishabh-git10 in #4278
feat: add Gefen-X (gefenx / gefenx_muon) optimizer integration by @thad0ctor in #7051
Fix attention support detection for trust_remote_code models by @Shaurya-M002 in #8229
Fix connected/cloud Max Tokens fallback for MiniMax M3 by @Yudeeswaran in #8914
[Feature] Reasoning effort slider / LM Studio Link Provider / Open Specific folder for project by @re4 in #8970
fix: pass force_download through to snapshot_download in _get_statistics by @arthi-arumugam-git in #8906
Studio: mention the leftover uv cache in uninstall by @claxman in #10442
studio: install torchcodec from the PyTorch cuXXX index on Linux aarch64 by @vivekvar-dl in #4456
fix(install): route Linux Intel GPU hosts away from CPU-only torch by @andomeder in #5274
Studio: search, sort, multi-select and collapse for project sources by @Rafael-Silva-Oliveira in #8850
feat(studio): type-to-activate search, composer and prompt inputs by @Fahad090NP in #10303
Studio: load S3 audio datasets — download the audio beside its manifest and point the manifest at it by @lonexreb in #10050
Studio: show the eject toast before the running-chats check by @claxman in #10469
Fix expected non-streaming cancellation handling by @Apoze in #9616
Studio: stop the SWA config probe from adding a phantom base model and refetching every load by @claxman in #10607
Studio: delete the models Studio only discovered, support files included by @atharvgaur1845 in #9313
Windows: keep uv cache under the Studio root by @Max-Reisinger in #9007
Document sharing models and projects across login users by @adamsiwiec1 in #9363
Wire UEmbed offset pooling and SPLADE sparse output into FastSentenceTransformer by @ysys143 in #9322
Don't reuse a leftover incomplete isolated Node tree by @adamsiwiec1 in #9361
Studio: bound repeated llama-server respawns after SIGKILL by @JulienJBO in #9689
studio: add keyboard shortcuts for UI zoom and scaling (Cmd/Ctrl + / - / 0) by @brickheadbs-claude in #9647
fix: [Bug] import unsloth failed and shows UnicodeDecodeError by @chensimian in #9704
Add the UEmbed unified training loss and its fine-tuning scripts by @ysys143 in #9323
Track the pinned upstream UEmbed reference behind the parity test by @ysys143 in #9324
Load oci:// models via llmman serve by @ericcurtin in #10030
Studio: keep Switch Back visible for a snapshot-path chat by @claxman in #10511
Studio: separate active context from processed tokens by @dre4moff in #9528
[CHORE] Comment out dead LoRA/GELU code paths by @elrensmin in #9265
fix: [Bug] Tool_Calling does not work properly by @chensimian in #9706
fix: OOM for GPT OSS 120b on 183GB of VRAM (B200) by @chensimian in #9703
fix: stop false link-definition probes forcing full-document renders by @apurv-1 in #10688
Studio: bake a relocatable RUNPATH into the Linux llama.cpp source build by @lonexreb in #12398
Studio browser: hide tab menu items that don't apply, use the toolbar's reload icon by @shimmyshimmer in #12728
Add Qwen3.5 model defaults for Studio by @ashzak in #5958
Studio: pipeline linked-folder RAG ingestion by @Biotrioo in #9873
Studio: find a hand-added mmproj when the GGUF repo publishes none by @yzxcj797 in #9365
tests: skip two version_compat suites where the daily sweep has no torch by @Sletch in #12274
feat(install): add Intel GPU (XPU/SYCL) auto-detection and llama.cpp SYCL compilation by @jischebeck in #9084
Studio: show the models drive in the Live Monitor when it is not the system disk by @danielhanchen in #12724
fix(studio): accessibility improvements for screen reader users (NVDA/JAWS) by @devinprater in #6601
Studio: open a pasted-text upload without its <pasted_text> wrapper by @MohammadHijjawi97 in #12198
Studio: bound the code highlighter cache for tool cells, previews and READMEs by @claxman in #10609
Studio: resolve GET /v1/models/: for an on-disk quant by @danielhanchen in #12727
studio: report why launcher recovery failed, not that it was absent by @Sletch in #9924
CLI: show a thinking model's reasoning with --think when attached to a running Unsloth by @breken-ai in #11956
Studio: show a JSON seed's values as written in the Data Recipe preview by @MohammadHijjawi97 in #12180
Studio: return a missing number in a recipe dataset page as null by @MohammadHijjawi97 in #12103
Offloaded embedding: one graph under compiled inference by @danielhanchen in #12726
Studio: int8 activations on T4 for Qwen-Image (2.5x per step) by @danielhanchen in #12690
studio: round the window button glyphs by @mahiatlinux in #12729
Studio: hide iq4_nl KV cache where llama.cpp runs its attention on the CPU (#6272) by @rodboev in #7008
Studio: seam-free VAE tiles for Qwen-Image-2.1 on low-VRAM tiers (fixes thin horizontal / vertical lines) by @danielhanchen in #12696
studio: report a denied rename as a possible ACL fault, not a scanner by @Sletch in #10075
Fix incomplete links showing [blocked] during chat streaming by @umran666 in #12104
Studio: opt-in recursive scanning for custom model folders by @aditya-786 in #6989
fix: do not permanently purge RAG when a project is recreated by @apurv-1 in #10583
Studio: support source code files in workspace folders and project knowledge bases by @harshaygadekar in #10395
Manual placement: log the --fit verdict the launch actually carries by @deepspace28 in #10831
fix(studio): allow loopback CORS origins and custom origin overrides in desktop mode by @harshaygadekar in #10084
Studio: accept a SKILL.md saved with a UTF-8 byte order mark by @MohammadHijjawi97 in #12195
Studio: make ConvRot the Z-Image-Turbo INT8 default by @danielhanchen in #12612
Studio: shared fused ConvRot act-quant kernel (L4 7.8% faster) and a G4 int8 GEMM tile table by @danielhanchen in #12693
Guard OAuth providers from legacy desktop clients by @NoahJenkins in #8747
Studio: fused RoPE for FLUX on ROCm (FLUX.2-klein 8% faster per image, pixel-identical) by @danielhanchen in #12701
Studio: flash attention by default for FLUX.2-klein on ROCm gfx11 by @danielhanchen in #12703
studio: keep the close glyph at its original size by @mahiatlinux in #12738
Studio: keep using an already-downloaded prequant .pt; new users get the .safetensors twin by @danielhanchen in #12711
Studio: "Denoising please wait..." status with step count, and a live denoising preview at no speed cost by @danielhanchen in #12692
Tell npm permission failures apart from a blocked registry (#8725) by @SergiorCode in #8729
Studio: keep a list item's first paragraph on its bullet in fetched pages by @MohammadHijjawi97 in #12179
studio: dump the backend's own thread stacks while a stall is in progress by @hellopahe in #9715
Add unsloth install-kernels for prebuilt xformers / causal_conv1d / mamba_ssm wheels by @danielhanchen in #12732
Studio: cap tool-result text at 256 KB before it reaches the model by @ItsRoy69 in #11430
fix(llama-prebuilt): print the correct hint for access-denied renames (WinError 5) instead of the scanner theory by @deepspace28 in #10182
feat(studio): show estimated context usage before model load by @InfoSage05 in #9475
Studio: keep the exit code and the first error line when training dies by @harshaygadekar in #11574
Studio: chunked VAE attention on ROCm when no fused kernel takes head dim 512 (fixes FLUX.1 2048x2048 OOM) by @danielhanchen in #12702
Studio: keep streamed int8 denoisers resident up to the measured fit by @danielhanchen in #12687
Studio: tidy the message action bars, branch picker and fork icon by @shimmyshimmer in #12735
unsloth: explain the real fix when nvidia-smi sees a GPU torch cannot use by @vivekvar-dl in #9281
Studio: resolve trust_remote_code off the event loop in /api/inference/status by @Lwrless in #10830
Studio: seam-free LTX-2.3 video decode (untiled when it fits, VRAM-sized tiles otherwise) by @danielhanchen in #12698
Windows installer: prefer uv-managed Python over global winget installs by @AryanGupta1112 in #7825
Studio: add a Sandbox settings tab with a Windows MXC opt-in and one-click host preparation by @danielhanchen in #12215
unsloth start vibe: launch Mistral Vibe against a running Unsloth server by @xavierpestel-ai in #8180
Studio browser: annotate code on demand, no idle desktop polling, smaller page cache on low-memory machines by @danielhanchen in #12734
Studio: per-block CUDA graphs under every offload mode, and compile below the offload hooks by @danielhanchen in #12694
perf(studio): keep the MLX GPU warm for a bounded time after generation by @Lyxot in #12721
Studio: whole-step CUDA graphs under offload on top of per-block graphs (L4 FLUX.1 16 GB 10% faster, HunyuanVideo-1.5 1 GiB lighter) by @danielhanchen in #12707
Studio: stop an IPv6 blackhole from killing the backend by @Lwrless in #10803
Studio: int8 ConvRot text encoder for Qwen-Image-2.1, fp8 fallback by @danielhanchen in #12684
SDPA: attend packed rows per segment instead of under a dense mask (2.9x less memory, 2.5x faster without flash-attn / xformers) by @danielhanchen in #12740
offload_layers='auto': plan for batch size 2, re-plan at Trainer init for the real batch by @danielhanchen in #12741
Studio: extend automatic context compaction to MLX by @dre4moff in #9399
Unsloth Desktop/Studio audio: follow-ups from testing the Clone, Transcribe, Edit, Separate and Convert pages by @LeoBorcherding in #12706
Studio: experimental Int8 Prefill run setting for MLX models by @Lyxot in #12550
studio: enable anthropic studio tools by @mahiatlinux in #12497
Studio: keep Wan2.2-5B's leading blocks resident under group offload by @danielhanchen in #12691
fix: enable XPU support for tests with agnostic test selection by @JoshuaL3000 in #8319
Run bitsandbytes NF4 Linear4bit through Unsloth's NF4 kernels by @danielhanchen in #12739
Deep research contract: follow the research Delete gate into the More menu item by @danielhanchen in #12751
Studio: seam-free tiled decode for every image VAE whose stock tiles fall under the floor (HunyuanImage-2.1 grid, Qwen-Image, slivers) by @danielhanchen in #12736
Live Monitor models disk: translate its labels and count every tile by @danielhanchen in #12749
Recount harness: bind the loader's msgs the history restore now estimates from by @danielhanchen in #12747
studiobench: delete through the reply's More menu, where #12735 moved it by @danielhanchen in #12748
SDPA packed-segment test: gate the CUDA memory test on a real GPU by @danielhanchen in #12761
Re-measure the Studio startup budget after the Sandbox tab and action bar rework by @danielhanchen in #12763
Transcript stream: never drop a phase change behind a newer update by @danielhanchen in #12760
version-compat latest lanes: install requests, stop shadowing torchcodec by @danielhanchen in #12769
Provider URL validation tests: count only the lookups the code under test made by @danielhanchen in #12771
Studio: use the edge Gemma 4 template for the QAT E2B/E4B repos by @PundarikakshNTripathi in #12715
Studio model picker: keep rows full width with overlay scrollbars, roomier On Device columns by @shimmyshimmer in #12746
MLX backend tests: answer zoo's _fusion_modules in the inference stub by @danielhanchen in #12783
scan-packages baseline: re-review datasets 5.1.0's readline loop by @danielhanchen in #12784
Load the compiled class when auto_model is a concrete class by @danielhanchen in #12757
Studio chat: no empty bubble under annotation-only messages by @shimmyshimmer in #12799
Studio: give toast notifications the composer's shadow by @shimmyshimmer in #12800
install-kernels: add flash-attn and install every kernel by default by @danielhanchen in #12744
Branch picker chevrons: give back the 24px hit target in the same 20px slot by @danielhanchen in #12781
studio: even the gaps between the window button glyphs by @mahiatlinux in #12782
studio: keep the linux desktop window resizable by @mahiatlinux in #12723
Studio: stop the Vulkan probe from popping "Entry point not found" on old Vulkan loaders by @oobabooga in #12762
Frontend tests: read the composer shadow and palette seeds where #12800 left them by @danielhanchen in #12810
Studio: send blank seed cells to Data Recipe prompts as empty text by @NilayYadav in #12789
Studio: make adding files to a knowledge base discoverable from the chat by @oobabooga in #12779
Studio: keep line breaks and ticked boxes in attached Word files by @NilayYadav in #12788
chore: pre-commit autoupdate by @pre-commit-ci[bot] in #12780
Studio: open Audio from More and keep it in the sidebar while it is open by @Etherll in #12807
Studio: show the close button in the mobile sidebar by @shimmyshimmer in #12806
Studio: 12px code font, context usage ring and small sizing fixes by @shimmyshimmer in #12805
Studio: keep project folders on the Docker volume by @NilayYadav in #12792
Studio: send an explicit CFG for every FLUX family on the sd.cpp engine by @danielhanchen in #12755
Use a native template icon in the macOS menu bar by @guus6457 in #10196
Studio: keep a trailing slash on Windows CUDA toolkit roots by @173787247 in #9966
Studio: chat with and export Mac LoRAs trained on base models by @NilayYadav in #12794
Studio: CI guard that fails when any family's VAE decodes in tiles below the seam floor by @danielhanchen in #12766
Studio: give API clients the same Qwen thinking sampling as Chat by @NilayYadav in #12791
Studio: route Qwen-Image-Edit / Edit-2509 / Edit-2511 and FLUX.1-Kontext GGUFs to the native sd.cpp engine by @danielhanchen in #12776
Studio: Qwen-Image-Layered on the diffusers engine, GGUF included by @danielhanchen in #12775
Studio: train Continued Pretraining on the body text column by @NilayYadav in #12790
Studio: Qwen-Image-Layered on the native sd.cpp engine by @danielhanchen in #12777
Studio: save audio clips and stems through the desktop save dialog by @Etherll in #12808
Thread browser harnesses: reach Delete through the More menu, and attack Copy in the dismissal probe by @danielhanchen in #12827
Studio: size seam-free VAE tiles from measured peaks (no slower than stock, no low-VRAM OOM) by @danielhanchen in #12764
Studio: leave MiniMax-H3's unpinned streamed blocks to diffusers' onload by @oobabooga in #12753
Studio: add Manage files to Settings > Data and reorder Settings tabs by @shimmyshimmer in #12825
Studio: dark dropdown glow, menu ticks and picker polish by @shimmyshimmer in #12826
Run narrow RMSNorm rows several per program, bit-identical (Qwen3 q/k norms) by @danielhanchen in #12754
Studio: dismiss the llama.cpp update toast once the job finishes by @indrajeetapache in #9199
fast_inference for Qwen3.5 / 3.6 MoE and Gemma-4 MoE with LoRA on the experts by @danielhanchen in #12742
Studio: let safetensors vision models see every image in a chat by @NilayYadav in #12787
Improve Korean Studio translations by @Jiye-Park01 in #9927
Studio: add custom validator blocks to Data Recipes by @kentwait in #7784
feat(studio): add model picker and reload shortcut to API monitor by @InfoSage05 in #11223
Fix Windows MXC runtime grants for uv-managed Python by @Imagineer99 in #12756
Fix MXC inherited runtime permissions and pending grant recovery by @Imagineer99 in #12758
Studio: keep a Library file's star, folder and name after an edit by @NilayYadav in #12793
feat(studio): opt-in reasoning for Custom Chat Completions connections by @wasimysaid in #12743
Studio: make the OS sandbox work in Colab and other containers by @danielhanchen in #12801
Studio: build each Qwen-Image-Edit variant on its own transformer config (2509 / original Edit no longer get 2511's zero_cond_t) by @danielhanchen in #12815
Studio: load ComfyUI-format int8_convrot / fp8 single-file DiTs instead of rendering noise by @danielhanchen in #12765
Clear the duplicate-definition backlog and gate it at zero by @rajarshidattapy in #9769
Studio: open chat HTML in the Browser and remove Canvas by @shimmyshimmer in #12804
Keep EmbeddingGemma and Qwen3-Embedding prompts when fine-tuning by @NilayYadav in #12795
fix(studio): resolve UUID-form CUDA_VISIBLE_DEVICES to physical GPU indices by @rsd-darshan in #8917
Wait on the pooled fetch itself in the closed-request browser test by @danielhanchen in #12835
Studio: run Qwen-Image-2.1 edits on 16 GB cards instead of refusing them by @oobabooga in #12752
Fix managed vLLM startup in Windows Studio by @Imagineer99 in #12750
Studio: add Swedish locale by @yeager in #10071
Studio: share one kernel install core with unsloth install-kernels by @danielhanchen in #12828
Seq2Seq LoRA task type and GA token count for T5Gemma2 by @daruoktab in #7795
Studio: compact long chats on API models too by @NilayYadav in #12786
README: document Homebrew installation on macOS by @SSakutaro in #9834
Studio: Sandbox Low/High and Permissions under Settings > Sandbox by @danielhanchen in #12802
Studio: load the ComfyUI-format twin of a hosted image prequant, and run ComfyUI fp8 on the fp8 path by @danielhanchen in #12820
Studio: do not offer a local model that is short a shard by @sts-change in #8985
Studio: fix audio mic choice, reload page, tour copy and palette search by @Etherll in #12813
Studio: discover models installed through oMLX by @sts-change in #8937
Send maskless causal flex calls to SDPA is_causal by @danielhanchen in #12773
Studio: more browser settings by @shimmyshimmer in #12829
Studio: list every runnable audio.cpp-gguf model on its Audio pages by @Etherll in #12818
Studio: add Hebrew (he) display language by @start-life in #6344
Pin qwen-image-layered's VAE config for the seam guard, and check coverage in Backend CI by @danielhanchen in #12843
Studio: send audio clips, stems and transcripts to other Audio pages by @Etherll in #12811
Studio: import Cursor, Claude Code and Codex conversations from Settings by @Tamsi in #8561
Studio: close the MLX context compaction gaps left after #9399 by @Lyxot in #12823
Studio browser: mark the panel's fetches to unsloth.ai for a Cloudflare skip rule by @danielhanchen in #12834
Studio: say when the audio runtime is not the release Studio installs by @Etherll in #12809
Call create_optimizer without a model in the embedding optim-bits test by @vineethsaivs in #12697
Runtime encoding lint: exempt the reviewed PyAV codec open by @danielhanchen in #12831
Studio: tighten the browser address bar and clear it while typing by @shimmyshimmer in #12852
Studio: clone with 13 more audio.cpp models, convert with Tone-Color VC by @Etherll in #12822
Studio: sample Qwen-Image-2.1 and FLUX.1 dev at ComfyUI's fixed sigma schedule by @danielhanchen in #12838
Studio: fix locale parity on main (Swedish sandbox strings, Manage files) by @danielhanchen in #12844
Build flex attention masks without graph breaks under torch.compile by @danielhanchen in #12837
Studio: open audio models on their Audio page by @Etherll in #12816
Studio: fine-tune Laya and Cloudflare Clef decision models and serve them by @NilayYadav in #12585
Studio: sandbox level picker and Settings > Sandbox cleanup by @shimmyshimmer in #12855
Studio: show audio clip details, playback and runs in the Library by @Etherll in #12814
install.ps1: guard the mirror env restore and run its test in Windows CI by @danielhanchen in #12849
Studio: size host RAM from the container's cgroup, and re-measure the MiniMax-H3 floor by @danielhanchen in #12854
tests: make the torchcodec provenance and mirror repair tests hermetic by @danielhanchen in #12850
Studio Desktop: stop the launch-time PATH probe from rewriting shell history by @danielhanchen in #12846
Studio: complete the OpenAI-compatible audio API by @Etherll in #12812
Read the API monitor unload button's disabled terms, not its exact spelling by @danielhanchen in #12858
Studio: add audio translations and list audio workflows in /v1/models by @Etherll in #12817
Studio: read chat replies aloud in a saved Audio voice by @Etherll in #12819
Studio: convert the decoded image to uint8 on the device by @danielhanchen in #12847
Studio: keep GLM-5.3-Flash's MTP head in Auto by @danielhanchen in #12833
Do not print the 16bit LoRA notice when a quantization_config still quantizes by @danielhanchen in #12848
Compile the Laya encoder for decision model training by @danielhanchen in #12778
EmbeddingGemma 2 support and embedding improvements by @danielhanchen in #12865
Studio: stop loading EmbeddingGemma in float16 in the RAG embedder by @oobabooga in #12864
Studio UI test: drive the sandbox level picker that replaced the menu switch by @danielhanchen in #12866

New Contributors

@kei-yamazaki made their first contribution in #12475
@1fanwang made their first contribution in #12505
@simpleqt made their first contribution in #10394
@jayzhou2309 made their first contribution in #12688
@KafKafrnZ made their first contribution in #7605
@Rishabh-git10 made their first contribution in #4278
@thad0ctor made their first contribution in #7051
@Shaurya-M002 made their first contribution in #8229
@Yudeeswaran made their first contribution in #8914
@re4 made their first contribution in #8970
@arthi-arumugam-git made their first contribution in #8906
@vivekvar-dl made their first contribution in #4456
@andomeder made their first contribution in #5274
@Rafael-Silva-Oliveira made their first contribution in #8850
@Fahad090NP made their first contribution in #10303
@atharvgaur1845 made their first contribution in #9313
@Max-Reisinger made their first contribution in #9007
@adamsiwiec1 made their first contribution in #9363
@ysys143 made their first contribution in #9322
@JulienJBO made their first contribution in #9689
@brickheadbs-claude made their first contribution in #9647
@chensimian made their first contribution in #9704
@ericcurtin made their first contribution in #10030
@dre4moff made their first contribution in #9528
@elrensmin made their first contribution in #9265
@jischebeck made their first contribution in #9084
@devinprater made their first contribution in #6601
@aditya-786 made their first contribution in #6989
@NoahJenkins made their first contribution in #8747
@SergiorCode made their first contribution in #8729
@AryanGupta1112 made their first contribution in #7825
@xavierpestel-ai made their first contribution in #8180
@PundarikakshNTripathi made their first contribution in #12715
@guus6457 made their first contribution in #10196
@173787247 made their first contribution in #9966
@Jiye-Park01 made their first contribution in #9927
@kentwait made their first contribution in #7784
@yeager made their first contribution in #10071
@daruoktab made their first contribution in #7795
@SSakutaro made their first contribution in #9834
@start-life made their first contribution in #6344
@Tamsi made their first contribution in #8561

Full Changelog: v0.1.902-beta...v0.1.903-beta

unslothai/unsloth v0.1.903-beta New Browser + Voice Cloning on GitHub