XiaoqianShen

Results 12 comments of XiaoqianShen

Since we could not share the Llama weight publicly, the API token lead to a private hugging face model. Here is a tutorial on how to duplicate our hugging face...

Yes, our model should be fittable for 24GB for vicuna13B and 12GB for vicuna7B. You can set the low_resource to false to fully utilize your gpu.