Repository Issue Activity (beta)

#139 Verify evals on Papers with Code

#138 Unable to reproduce the ScanNet 3D object detection F1 scores reported in Table 6

#137 请问不是真实尺度的点云，spatiallm能处理吗？

#136 Questions about fine-tuning the Structured3D dataset

#79 Trying to reproduce structured3d finetuning results

#114 相同的输入多次执行结果不一样，有办法保持一致吗？

#135 Questions About Reproducing the SceneScript Results Reported in the Paper

#131 Questions About Sample Acquisition per Scene in the SpatialLM Dataset

#134 SpatialLM inference crashes: KeyError: 'shape'

#132 Bug in code that calculates layout F1 score

#125 How do we train our own dataset?

#129 请问有方法控制生成某一类别的数量吗？

#130 模型出现重复识别一面墙的情况

#128 如何导出onnx

#127 When feeding rotated point clouds as input, the model performance degrades significantly.

#111 How to Reproduce Pre-Training Results

#123 推理问题

#103 训练，3块显卡总显存90g不够用，显存爆了。如何能控制显存占用

#122 Model detect non-existent walls and furniture on aligned data

#124 布局识别不对

#119 Details about the data annotation

#118 OOM

#121 用手机拍的视频出来的结果不对

#120 Would non-axis aligned data work?

#115 输出结果都是相同的数据是什么原因？