finetuner
Open source
freemium
4.6/5
Unsloth
2x faster QLoRA fine-tuning with hand-tuned Triton kernels. Free OSS for single-GPU; commercial Pro for multi-GPU.
Overview
2x faster QLoRA fine-tuning with hand-tuned Triton kernels. Free OSS for single-GPU; commercial Pro for multi-GPU.
Pros
- 2x speed vs HF default
- Lower VRAM use
- Great Colab support
Cons
- NVIDIA only
- Multi-GPU is paid
Compatibility
| Operating systems | Linux |
| GPU backends | NVIDIA CUDA |
| License | Open source · freemium |
Get Unsloth
Frequently asked
Is Unsloth free?
Unsloth has a paid tier (freemium). Check the pricing page for current terms.
What operating systems does Unsloth support?
Unsloth supports Linux.
Which GPUs work with Unsloth?
Unsloth supports NVIDIA CUDA. CPU-only inference is also possible but slow.
Reviewed by RunLocalAI Editorial. See our editorial policy for how we evaluate tools.