unslothai/unsloth

Request for Notebook to Fine-Tune Qwen TTS (or Alternatives Using Existing Notebooks) 6 days ago

feature request help wanted good first issue

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature] add aarch64 container images 7 days ago

feature request help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature] Add FT support for the Qwen3-TTS model. 15 days ago

feature request help wanted good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

Apple Silicon Support 15 days ago

on roadmap help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature Request] fast inference for LFM (and Mamba models) 17 days ago

help wanted good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

Docs: current Codex local llama.cpp guide does not work as written with Responses API 18 days ago

good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature] Unsloth studio: Set dataset streaming mode 24 days ago

feature request help wanted good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature] Running Usloth Studion in kaggle 27 days ago

feature request help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature] Draft model / speculative decoding 27 days ago

feature request help wanted good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature] Request permission before downloading model on first run 27 days ago

feature request help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Qwen3.5 vision training , error "only for image model" 27 days ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature] Unsloth/ Whisper/Large-v3 - S3 Bucket connection 27 days ago

feature request help wanted good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature] Model Management in Unsloth Studio - Delete & Update Models 27 days ago

feature request help wanted good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Qwen 2.5 VL 7B full fine tuning 27 days ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Error Patching SFTTrainer 27 days ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Forced coupling between num_generations and per_device_train_batch_size in GRPOTrainer resulting in OOM 27 days ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

Batch inference seems to have gibberish 27 days ago

currently fixing good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature] - "merged float32" - Float32 precision safetensors, with low VRAM/RAM requirements. 2 months ago

feature request help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature] Add Unsloth Studio UI value to tune the llama-server --fit-target flag for squeezed extra performance. 3 months ago

feature request help wanted good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature] K8s Chart? 3 months ago

feature request help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature]Request: 4-bit Quantized Export (AWQ/GPTQ) for vLLM in Unsloth Studio 3 months ago

feature request help wanted good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature]Terminal command and Python script execution within session specific Docker Container 3 months ago

feature request help wanted good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

Request: Add Support for OLMo3 (OLMo2 Already Supported) 3 months ago

feature request help wanted good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Docs] Update DPO example to use DPOConfig instead of TrainingArguments 4 months ago

help wanted good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Installation on Windows with Conda fails due to aggressive PyTorch version replacement 4 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[BUG] can't install with pip3 4 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Hermes Lora chat test bug: does not have a {% if add_generation_prompt %} for generation purposes 4 months ago

help wanted good first issue bug

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

How to set "reasoning_effort" of GPT-OSS during GRPO rollouts? 4 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature Request] Add Idefics3 architecture support (Granite Docling VLM) 4 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

Why are you lying about intel support? 5 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

trainer.evaluate() fails with wandb 5 months ago

help wanted bug

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Cannot load qwen3-vl series with lora adapter on vllm. 5 months ago

good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Tool_Calling does not work properly 5 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

GPT-oss-20b-BF16: torch.OutOfMemoryError: CUDA out of memory. 5 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

How can I increase the GPU utilization while training / finetune an LLM with unsloth 5 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] import unsloth failed and shows UnicodeDecodeError 5 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Docs Improvement] Improve documentation on how to export model from Colab 6 months ago

feature request help wanted good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

OOM for GPT OSS 120b on 183GB of VRAM (B200) 8 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

GRPOTrainer fails after SFTTrainer - Gemma-3-1b-it (matmul shape mismatch) 8 months ago

help wanted unsure bug? inactive

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] After adding import unsloth to the first line of the script, the GRPOTrainer fails to run properly; however, it works normally again once this import is removed. The Sophia optimizer interface being used was generated by an AI. 8 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Please fill in your issue title here. 8 months ago

help wanted Discussion

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

Notebook to show vLLM inference 8 months ago

currently fixing good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

OOM-ing on Nvidia Jetson Orin Nano 8 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Feature Request] Raw txt file training 8 months ago

on roadmap feature request good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

reinforce(gspo) training didn't yield any improments 9 months ago

help wanted unsure bug?

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

Kaggle - GPT OSS Data set 9 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

Bug: Text-only model (gemma-3n-E4B-it) incorrectly tries to load vision model (mobilenetv5_300m_enc) 10 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Multi-gpu finetuning in Unsloth 10 months ago

help wanted multigpu

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

Evaluation loss becomes constant 10 months ago

currently fixing good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Vision not works for mistral-small3.2 10 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] GRPO has error in latest update (Exception: Invalid prefix encountered) 10 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Huge difference in Training when using FastLanguageModel and FastModel with same parameters on an llm 11 months ago

AI Summary: The user reports a significant bug where training an LLM with `FastLanguageModel` yields vastly different results compared to using `FastModel`, even when both are configured with identical parameters. They have confirmed their `unsloth` installation is updated and provided environment details including Colab T4 and local RTX A4500 GPU information.

Complexity: 4/5

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

AutoModelForSequenceClassification 11 months ago

feature request help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] CUDA error: no kernel image is available for execution on the device 11 months ago

AI Summary: The user is encountering a "CUDA error: no kernel image is available for execution on the device" when attempting to load a Qwen model using `unsloth`'s `FastLanguageModel.from_pretrained` on Kaggle. This issue occurs early in the process, before any trainer is initialized, despite the user updating `unsloth` and trying various dependency versions.

Complexity: 3/5

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

Number of images does not match number of special image tokens in the input text. Got 976 image tokens in the text but 1024 tokens from image embeddings. 11 months ago

fixed - pending confirmation good first issue

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

please give t5 support. 11 months ago

feature request help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

no speedup even a little slower for training qw3 8b with lora 11 months ago

AI Summary: The task is to investigate why Unsloth, a library designed to accelerate large language model fine-tuning, is not providing a speedup, and in fact shows a slowdown, when training a Qwen-8B model with LoRA on a single 4090 GPU. The issue needs debugging, focusing on the configuration and usage of Unsloth within the existing transformers Trainer setup. The investigation should compare performance with and without Unsloth to pinpoint the source of the slowdown.

Complexity: 4/5

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Inconsistent responses using gemma3 12B model 11 months ago

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] Loss & gradient norm zero when full finetuning custom model & most layers frozen 11 months ago

AI Summary: The task is to debug a loss and gradient norm issue in a custom PyTorch model fine-tuned using Unsloth. The model uses a modified LoRA architecture with learnable scaling factors, and all layers except the scaling factors are frozen. The issue occurs when using Unsloth's `full_finetuning=True` mode, and the goal is to identify why the loss and gradient norm are zero.

Complexity: 4/5

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

[Bug] train_on_responses_only is broken with new Unsloth releases (worked in March) 11 months ago

AI Summary: Debug a broken `train_on_responses_only` feature in the Unsloth library. The feature, crucial for fine-tuning LLMs on datasets with overlapping inputs, stopped working in recent releases, causing a `ZeroDivisionError`. The task involves investigating the code changes between the working (March) and broken releases to identify the source of the error, potentially involving examining the dataset processing and loss calculation within the Unsloth library and its interaction with the provided training notebook.

Complexity: 4/5

help wanted

unslothai/unsloth

67.7K

Unsloth Studio is a web UI for training and running open models like Gemma 4, Qwen3.6, DeepSeek, gpt-oss locally.

Python

#agent#deepseek#fine-tuning#gemma#gemma3#gpt-oss#llama#llama3#llm#llms#mistral#openai#qwen#reinforcement-learning#self-hosted#text-to-speech#tts#ui#unsloth

Open Issues Need Help