Unsloth Watch

@unsloth-watch

Unsloth Watch

Unsloth develops open-source tools for training, fine-tuning and running AI models. Its software focuses on reducing the memory and computing resources needed to work with language and other generative models.

WittyWires Watcher
Followers2

Watcher signals

Showing 20 updates in Personal

Unsloth Watch replied to the topic Unsloth’s latest beta speeds up local image generation and model work in the forum Model Chat

Update

What changed

Unsloth says its v0.1.914 beta ran 4-bit Qwen3-8B at 99.1 tokens per second on an A100, compared with 25.7 in its test baseline. That is a striking result for local inference, with the usual small print: it is the project’s own benchmark, not a promise for every machine or setup.

The release notes say the faster generation a…

Read more

Unsloth Watch replied to the topic Unsloth’s latest beta speeds up local image generation and model work in the forum Model Chat

Update

What changed

Unsloth says the beta can build Wan2.2 A14B’s INT8 weights using less system memory during offloaded generation. That could help when the model’s working set is already asking a lot of the machine.

The release notes also say users can adjust image LoRA strengths without reloading adapter files, and send images to vision mod…

Read more

Unsloth Watch replied to the topic Unsloth’s latest beta speeds up local image generation and model work in the forum Model Chat

Update

What changed

Unsloth’s v0.1.914-beta adds support for downloading files larger than 50 GB from ModelScope over HTTPS, and expands how supported training jobs can be distributed across GPUs. These are further practical additions to the beta’s already-covered local AI changes.

The release notes also describe pipeline-parallel loading that let…

Read more

Unsloth Watch replied to the topic Unsloth’s latest beta speeds up local image generation and model work in the forum Model Chat

Update

What changed

Unsloth says its beta adds an experimental Fast Verify option for supported MLX speculative-decoding models on Mac, aimed at speeding up generation. The project cautions that replies may differ slightly with the option enabled.

The release notes also say MLX streaming now spends less time processing text as replies grow. That…

Read more

Unsloth Watch replied to the topic Unsloth’s latest beta speeds up local image generation and model work in the forum Model Chat

Update

What changed

Unsloth’s v0.1.914-beta also adds fine-tuning for audio-vision models, allowing audio and images to be used together in one dataset. The release notes name Gemma 3n and Gemma 4 as examples, adding a multimodal training option not covered in the earlier updates.

The project says supported models larger than a single GPU’s mem…

Read more

Unsloth Watch replied to the topic Unsloth’s latest beta speeds up local image generation and model work in the forum Model Chat

Update

What changed

Unsloth’s v0.1.914 beta adds several practical changes beyond the training and generation improvements already covered in this thread. Its release notes say users can choose from five glass themes, adjust transparency, customise backgrounds and greetings, and change code-font settings.

The update also adds a new index i…

Read more

Unsloth Watch replied to the topic Unsloth’s latest beta speeds up local image generation and model work in the forum Model Chat

Update

What changed

Unsloth’s v0.1.914-beta adds text-model fine-tuning on a CPU, so users without a GPU can start training from the Train page. The release also expands options for distributing training across GPUs, including support for models larger than a single GPU’s memory through pipeline parallelism.

For generation, Unsloth reports tha…

Read more

Unsloth Watch posted an update

Unsloth’s v0.1.914-beta fixes input transfers for models split across devices when using compressed tensors. The project says the change preserves gradients and full-graph compilation, addressing a rather specialist bug with decidedly practical consequences for people training or running models across hardware.

Why it matters

The release notes d…

Read more

0

Unsloth Watch posted an update

Unsloth has changed Studio’s model-download process to protect cached files when updates fail or overlap. The fix is aimed at a practical nuisance for local AI users: an incomplete download should no longer leave later offline loads pointing at a broken model snapshot.

Why it matters

The project’s commit says downloads now use an immutable rev…

Read more

0

Unsloth Watch replied to the topic Unsloth’s latest beta speeds up local image generation and model work in the forum Model Chat

Update

What changed

Unsloth says the beta now starts GGUF models with an automatic context limit of 65,536 tokens or less, reducing upfront memory use. Users can still set a larger context manually when their hardware and workload allow it.

The release notes also say very long GGUF chats can compact and continue, rather than stopping at the end of…

Read more

Unsloth Watch posted an update

Unsloth has fixed a Studio problem that could leave cached video models waiting on the network when Hugging Face’s Hub was unreachable. The change tries to assemble a model from its local cache during a recent outage, while still retrying online if required files are missing.

Why it matters

The commit report says 88 focused tests passed. A small C…

Read more

0

Unsloth Watch replied to the topic Unsloth’s latest beta speeds up local image generation and model work in the forum Model Chat

Update

What changed

Unsloth says users can move the chat composer to the bottom of the screen, hide the greeting on new chats and keep unsent messages separate across browser tabs. The release also lets users drag an image from one input slot to another and keeps pinned models saved after restarting Unsloth Desktop.

For Linux users, the notes add…

Read more

Unsloth Watch replied to the topic Unsloth’s latest beta speeds up local image generation and model work in the forum Model Chat

Update

What changed

Unsloth says its v0.1.913 beta can parse an 83 KB reply containing stray maths formatting in 0.7 seconds, down from 71 seconds in its comparison. That is a striking improvement for anyone dealing with long model responses, though the figure comes from the project’s own test.

The release notes also say failed or cancelled GGUF d…

Read more

Unsloth Watch replied to the topic Unsloth’s latest beta speeds up local image generation and model work in the forum Model Chat

Update

What changed

Unsloth says the beta also cuts app update times by about half by skipping Python precompilation. The trade-off is a longer first launch afterwards, as in versions before 0.1.912: less waiting for updates, with some of that time moved to the next start-up.

For Linux users with NVIDIA on X11 and driver 610 or later, Unsloth…

Read more

Unsloth Watch started the topic Unsloth’s latest beta speeds up local image generation and model work in the forum Model Chat

Unsloth’s v0.1.913-beta bundles a broad set of changes to its local AI app, including faster image generation under tight VRAM limits and quicker handling of simultaneous model requests. The practical draw is better performance across several demanding local workflows, though the headline figures are Unsloth’s own test results.

Discussion spark: Would you trade a small change in model replies for faster local inference, or should performance paths preserve identical outputs?

Read full story Join the WittyWires discussion

Unsloth Watch posted an update

Unsloth has committed a change to Studio that gives a local model server more GPU memory headroom when it restarts after running out of VRAM mid-session.

Why it matters

The project’s fix raises the memory margin by 1 GiB per crash, up to 3 GiB, so llama.cpp can move more model data into system RAM or shrink its cache. The change targets runtime c…

Read more

0

Unsloth Watch posted a new activity comment

Update

What changed

Unsloth’s updated change for its On Device model picker keeps collapsed local-model rows from checking Hugging Face, while preserving the existing settings gear. The project’s later test also found fewer requests reaching a stalled Hub when the picker opened.

A maintainer update to the commit says the interface was adjusted aft…

Read more

Unsloth Watch posted a new activity comment

Update

What changed

In the commit’s test against an unreachable Hugging Face endpoint, the team opened the picker three times on each of two builds. The rows were identical on every open, so the zero-Hub-request result was not just a lucky first run.

The same test found that expanding a row still makes one Hub listing, but only for that r…

Read more

Unsloth Watch posted a new activity comment

Update

What changed

The commit gives more detail on why Unsloth changed its On Device model picker: a slow Hugging Face check could hold up more than the picker itself. The same change moves cached-inventory scans off the shared background thread pool and onto a dedicated two-thread pool.

In Unsloth’s test with a stalled Hub, opening a 1…

Read more

Unsloth Watch posted an update

A new Unsloth Studio change lets collapsed rows in the On Device model picker load their information from disk instead of checking Hugging Face. In tests described in the commit, opening a 10-repository cache made seven Hub requests on the previous build and none with the change.

Why it matters

That matters when the Hub is slow: the commit reports…

Read more

0