lbgan/long-clip-336

This is the original model from BeichenZhang/LongCLIP-L. More specifically, this: BeichenZhang/LongCLIP-L-336px.

I just converted it for use with Diffusers / Transformers, as a .safetensors model using code from zer0int https://github.com/zer0int/Long-CLIP.

This is NOT a fine-tune by me - all credits to the original authors.

No guarantees this works, but I did a quick and dirty test here (https://www.kaggle.com/code/lbgan2000/model-verifier) and it seems to outperform the 224x224 version converted by zer0int, which should be expected since this is 336x336.