Microsoft's new Phi-4 model in all formats. Includes GGUF, 4-bit bnb and original versions. Includes Unsloth's bug fixes.
Unsloth AI
Enterprise
company
Verified
AI & ML interests
Hey! We're focusing on making AI more accessible to everyone!
Recent Activity
View all activity
Organization Card
🦥Unsloth makes fine-tuning of LLMs & Vision LMs 2.2x faster and use 80% less VRAM!
Join our Discord server! Our open source Github: https://github.com/unslothai/unsloth
| Unsloth supports | Free Notebooks | Performance | Memory use |
|---|---|---|---|
| Llama 3.1 (8B) | ▶️ Start on Colab | 2.1x faster | 60% less |
| Llama 3.2 (1B & 3B) | ▶️ Start on Colab | 2.1x faster | 60% less |
| Llama 3.2 (11B) Vision | ▶️ Start on Colab | 2x faster | 60% less |
| Gemma 2 (9B) | ▶️ Start on Colab | 2x faster | 63% less |
| Mistral Nemo (12B) | ▶️ Start on Colab | 2x faster | 60% less |
| Phi-3.5 (mini) | ▶️ Start on Colab | 2x faster | 50% less |
| Qwen 2.5 (7B) | ▶️ Start on Colab | 2x faster | 50% less |
Collections
13
Unsloths Dynamic 4bit Quants selectively skips quantizing certain parameters; greatly improving accuracy while only using <10% more VRAM than BnB 4bit
-
unsloth/phi-4-unsloth-bnb-4bit
Text Generation • Updated • 109 • 9 -
unsloth/Llama-3.2-11B-Vision-Instruct-unsloth-bnb-4bit
Image-Text-to-Text • Updated • 40k • 14 -
unsloth/Qwen2-VL-7B-Instruct-unsloth-bnb-4bit
Image-Text-to-Text • Updated • 15.7k • 6 -
unsloth/Qwen2-VL-2B-Instruct-unsloth-bnb-4bit
Image-Text-to-Text • Updated • 8.65k • 4
models
2
datasets
8
unsloth/notebooks
Updated
•
90
•
30
unsloth/LaTeX_OCR
Viewer
•
Updated
•
76.3k
•
941
•
7
unsloth/Radiology_mini
Viewer
•
Updated
•
2.31k
•
2.05k
•
11
unsloth/llava-instruct-mix-vsft-mini
Viewer
•
Updated
•
9.92k
•
210
unsloth/RLAIF-V-Dataset
Viewer
•
Updated
•
2.49k
•
51
•
3
unsloth/studio_mmlu
Viewer
•
Updated
•
14k
•
34
•
2
unsloth/studio_dataset
Viewer
•
Updated
•
300
•
34
•
1
unsloth/datasets
Viewer
•
Updated
•
891
•
43
•
2