forum
Batch inference broken with Gemma2
Mistral 3.2 Finetuning with Unsloth + Serving with vLLM
Open
Cannot Fine-Tune Gemma-3N with Images
Is the dots LLM supported?
Slow Inference with fine-tuned Qwen3
Solved
Devstral-Small-2507-GGUF:IQ2_XXS does not support tools
Data Format (JSON Input)
Multi-GPU in Unsloth.
Qwen3 notebooks use SFTTrainer instead of Unsloth Trainer any reason ??
Open
Finetuning SparkTTS0.5B not working
Open
TypeError: unsupported operand type(s) for +: 'Tensor' and 'NoneType'
Gemma 3 Fine Tuning - Tons of vram needed?
Zero Divison error during training
Gemma 3n Colab Notebook Not Working
Modelfile missing
Open
Can I use bone stock Qwen3 1.5B Instruct model for finetuning?
Test question
Worse performance while jailbreaking Llama
Having an issue with VLM qwen2.5 VL inference
i am having this error while i am trying to fine tune gemma 3
Open
GRPO training CUDA memory issue
Fine Tuning Qwen 3
Issue with Synthetic-data-kit + Unsloth + vLLM
MI300X Training Speeds
SparkTTS Colab breaks on inference
qwen 2.5 vl error
Negative loss during training
Open
Multimodal finetuning of Gemma3N
qwen3-30b-a3b finetuning
I need a skilled AI/ML consultant who has native level of English
llama.cpp error: 'invalid unordered_map<K, T> key
Open
Qwen 2.5vl RuntimeError
how to specify max_seq_length when using unsloth train_on_responses_only?
Docs website error
Qwen 2.5 Vl 7B Instruct
How to include Validation Loss?
RuntimeError: Direct module loading failed for unsloth_compiled_module_phi3: positional argument fol
Solved
Gemma3 quantization below Q8
Open
No longer "smartly offloading gradients to save VRAM"?
Open
Solved
[Gemma3] Please help with correct dataset formatting
AttributeError: 'LlamaForCausalLM' object has no attribute 'load_lora'
GRPO run with Qwen3-4B
❌ GGUF conversion failed - n_tensors = 0 during save_pretrained_gguf
Open
Been trying to try out the GRPO Notebook
Problem with long context finetuning
A dependency updated and I can't seem to get Unsloth to run anymore. What am I doing wrong?
Model Architecture Suggestions for Grammar-based Models
SFT -- error logging eval_loss
Nanonets OCR model quantizeed with unsloth in ollama
Open
Runpod template
Open