Heykuki News
Top
New
Best
Ask
Show
Jobs
Toggle theme
Login
Top
New
Best
Ask
Show
Jobs
Show HN: Finetune Llama 3.2 Vision in a Colab | Heykuki News
Show HN: Finetune Llama 3.2 Vision in a Colab
colab.research.google.com
10 points
danielhanchen
2 years ago
Hey HN! Just added vision finetuning support in Unsloth! VLMs like Llama 3.2 11B, 90B, Pixtral 12B, Qwen2 VL 2B, 7B, 72B and all Llava variants. Unsloth finetunes them up to 2x faster with up to 70% less VRAM without any accuracy degradation!
No comment yet