ShapeLLM

ShapeLLM copied to clipboard

Hello, dear friend! First, thanks for sharing so outstanding work ! ! I notice that Tabel 8 of paper report the recognition accuracy (%) **averaged** under the instruction-typed prompt “What...

TangYuan96

Problems about the pretrained model

2

Thank you for your excellent work! Here are some problems with the pre-trained model of [Recon++-L](https://huggingface.co/qizekun/ReConV2/blob/main/zeroshot/large/best_lvis.pth). As the image of fine-tuning logs shows, the encoder of the checkpoint is small...

jerryfeng2003

Thank you for sharing so great work!! It is very powerful. Regarding the checkpoints for ShapeLLM, there seem to be two types of [checkpoints](https://github.com/qizekun/ShapeLLM/blob/main/docs/MODEL_ZOO.md) used in the evaluation scripts When...

TangYuan96

Prediction seems inaccurate in Embodied Object Understanding 😭

2

 Thank you for the excellent work! I encountered problem that I can't get ideal prediction for **Embodied Object Understanding**. Here is an example of the prediction: - **USER:** Where...

Bokai-Ji

Thank you for your awesome work! I got a question when I use "finetune.sh". When I set with_color=False, my code will report the following problem: RuntimeError: shape '[8192, 32, 6]'...

Chunmian-art