UMass-Foundation-Model / 3D-LLM

Code for 3D-LLM: Injecting the 3D World into Large Language Models

UMass-Foundation-Model/3D-LLM Issues

Can you share the script of prompting ChatGPT to generate the description of 3D data?
Closed 5 months ago2
Difference between scannet features for finetuning and for pretraining?
Closed 5 months ago4
`final_scene_map_dict_scan_v2` and `v3` have conflicting key: value pairs
Closed 5 months ago2
How to evaluate the model?
Closed 5 months ago2
Some questions about the code
Closed 5 months ago1
How to process scannet dataset?
Closed 5 months ago1
How to map the 'val_1_vqa_result.json' with 'all_questions.json' for val set in pretrain.yaml?
Closed 5 months ago1
For the json files used in pretraining, did you split them to train, val and test?
Closed 5 months ago1
HW spec for Inference
Closed 5 months ago1
How to ensure the accuracy of the data
Closed 5 months ago1
The held-in datasets
Closed 5 months ago1
How to evaluate?
Closed 5 months ago4
Train 3D-LLM from scratch
Closed 5 months ago
Visual grounding data format and training process
Updated 6 months ago
Render objaverse, the operating system will restart.
Updated 7 months ago
How to align scene_ids in `data_part2_scene.json` from those in ScanNet?
Closed 9 months ago9
Sample dataset for 3DMV-VQA
Updated 7 months ago
About the annotation of Scannet
Updated 7 months ago1
m@kIoU Metric Results for 3D Dense Captioning
Closed 8 months ago1
Question about 3D Localization Mechanism
Updated 8 months ago
Categories of Questions
Updated 8 months ago
the processing of objaverse feature
Closed 8 months ago4
Get "" answer when run evaluate.py with scanqa.pth as resume_ckpt_path on scannet dataset
Closed 8 months ago2
How is scene data rendered from point clouds into images before three steps?
Closed 8 months ago2
The training goes error when you updated the 'blip2_t5.py' for optimizer parameters loading
Closed 8 months ago1
ValueError: loaded state dict contains a parameter group that doesn't match the size of optimizer's group
Closed 9 months ago4
How many obj files in all in your Object Dataset?
Closed 9 months ago5
What is `objaverse_frame_cap3d`?
Closed 9 months ago3
Are there any model checkpoints available to directly load the models? Or do I need to train them.
Closed 9 months ago3
The scene data is not accesed in google drive now because to many peple have downloaded it
Closed 9 months ago5
Where is the files such as './examples/all_refer_questions_train.json'
Closed 9 months ago1
How to inference the model given a 3D scene or object data?
Closed 9 months ago1
Some questions
Updated 9 months ago3
How to generate the data for 3D Grounding?
Updated 10 months ago
I can't find the data for 3D Grounding and 3D Dense Captioning in file 'data/data_part2_scene.json'
Updated 10 months ago2
license
Closed 10 months ago1
The google cloud links are the same for language annotations on the object data and the scene data.
Closed 10 months ago1
dataset
Closed 10 months ago4
Do I need to provide some external datasets like COCO dataset in order to run the models?
Closed 10 months ago2
About hardware source
Closed 10 months ago1
About point cloud data
Closed 10 months ago1
Code for multiple tasks in the project page
Closed 10 months ago1
Training batch size
Closed 10 months ago1