Instructions to use Zhengyi/LLaMA-Mesh with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Zhengyi/LLaMA-Mesh with Transformers:
# Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("Zhengyi/LLaMA-Mesh") model = AutoModelForCausalLM.from_pretrained("Zhengyi/LLaMA-Mesh") - Notebooks
- Google Colab
- Kaggle
Inquiry on Reducing Mesh Size
DEAR Zhengyi,
I hope this email finds you well.
I am currently using SolidWorks to generate 3D models for training purposes, but I have encountered an issue with mesh density. For instance, creating just two simple circles can result in an exported mesh file containing tens of thousands of tokens (around 30,000), which often leads to out-of-memory (OOM) issues.
Could you kindly advise if your team has experience or recommended approaches for reducing the mesh size (number of tokens) while maintaining sufficient geometric accuracy? Are there specific settings, tools, or workflows you would suggest?
Thank you in advance for your time and assistance. I sincerely appreciate any guidance you can share.
Best regards,