forum
unsloth in DGX Spark
Unsloth Studio Website no styling
Solved
I am unable to access anything other than the chat
[FAILED] llama.cpp build failed at step: cmake configure
long context
Open
DAC dataset needs 'audio' and 'text' columns, got: ['instruction', 'input', 'output']
Unsloth studio point to existing models folder?
Open
Unsloth Documentation Changes
Qwen3.5-4B SFTTrainer ignores "text" column & VL processor breaking text-only finetuning
Open
Intel mixed setup
Website not working
Open
Ask an experience in fine tuning using synthetic data
Docs Contribution Request
Open
[Regression?] Official 500K gpt-oss-20b notebook now fails with CUDA OOM
Fine-tuning Qwen3.5 to play chess and predict the next move
How can I install qwen 3.5 4b Q8 in my M4 mac?
Newbie question, Xeon + 64GB ECC RAM and old AMD RX480 8GB VRAM
unable to convert to fine tunned model to gguf
Fix misspelling in Qwen3.5 Documentation
Solved
[Unsloth Ecosystem] DeepSeek-OCR-2 deployment 98% complete on RTX 2080 Ti (Turing) – stuck on final
Open
OpenEnv hackathon X unsloth
Training hangs at step 1 when fine-tuning Qwen 3.5
Solved
Using smart_resize from qwen_vl_utils when finetuning qwen3-vl
Is the version uploaded 4 days ago for Qwen3.5-122B already updated with the tool-using fixes?
Fast Inference OOMs with vLLM for GRPO
Solved
Phi 4 mini doesn't work with Unsloth
Tiny doco error
Open
Unable to merge lora adapter with base model, after doing CPT on a model with tied_embeddings=True
gpt-oss-(20B)-GRPO.ipynb 'ParameterModule' is not defined
Solved
Finetune support for Qwen3.5-35B-A3B
Solved
Trained Unsloth Mistral-7B with 1024 max_seq_length — now inference truncates anything >1024. How t
Solved
ValueError when running inference with Qwen3-VL-8B LoRA checkpoint (FastVisionModel)
Open
GLM Flash GGUF Optimizations
Solved
Llama-3.3-70B-Instruct Clarity
GLM 4.7 Flash - Notebook Optimization
Open
Help: Need to improve performance from a 7b deepseek model or qwen 8b
Open
Fine-tune MoE Models 12x Faster with Unsloth
GLM 4.7 Flash - Notebook finetuning - Issues
Solved
MoE 4bit quant fine tuning
Open
Request for Unsloth quants: Nanbeige4.1-3B
Request for Unsloth quants: MiniMax M2.5
Numba needs NumPy 2.2 or less. Got NumPy 2.4. on official docker image
GLM-5-GGUF disable thinking
GPT-OSS 20B (MoE) – FT → GGUF → Ollama → Triton (vLLM) | Need Guidance I’m fine-tuning GPT-OSS 20B
Request an enterprise account for reinforcement learning fine-tuning (GRPO)
GLM 4.7 Flash (30B-A3B) - bnb 4 bit?
Types of RL algorithms supported
EmbeddingGemma (300M) fine-tuning notebook throwing error while pushing to HF
GKDTrainer + Unsloth
Error(s) in loading state_dict for NextDiT