DepthAnything/Depth-Anything-V2 issues and pull requests

#168 - How to train with our own data

Issue - State: open - Opened by serna01 1 day ago

#167 - Question about Metric Depth Training

Issue - State: open - Opened by shilpaullas97 2 days ago

#166 - DepthAnythingV2Preprocessor EERO

Issue - State: open - Opened by CPLACKY 2 days ago

#165 - Unit of measure for metrics depth ?

Issue - State: open - Opened by avanmalleghem 3 days ago

#164 - 训练和推理结果不一致原因？

Issue - State: open - Opened by tangjunjun966 3 days ago

#163 - Scaling problem of metric depth when inferring on the NYUv2 dataset

Issue - State: open - Opened by YkiWu 4 days ago - 1 comment

#162 - Depth-Anything-V2-Base license

Issue - State: open - Opened by juntaosun 6 days ago

#161 - inquiry on the future

Issue - State: open - Opened by donrikk 7 days ago

#160 - When the Depth-Anything-V2-Giant 1.3B model will be released？

Issue - State: open - Opened by krrr8s 8 days ago - 2 comments

#159 - feat: add pyproject toml

Pull Request - State: open - Opened by badayvedat 8 days ago

#158 - Why does the hf example use image.size[::-1] for interpolate

Issue - State: open - Opened by ZihaoZheng98 8 days ago - 1 comment

#157 - [Question] How to align the predicted depths to a real depth map.

Issue - State: closed - Opened by Master-cai 8 days ago - 3 comments

#156 - S

Issue - State: open - Opened by FASOOS1 10 days ago

#155 - Why abandon ZoeDepth in metric depth fine-tuning of V2?

Issue - State: open - Opened by liboyu02 10 days ago - 1 comment

#154 - Your code has sigmoid output for metric depth

Issue - State: open - Opened by adizhol-str 11 days ago - 4 comments

#153 - Questions about the teacher model

Issue - State: open - Opened by li12138cmd 15 days ago - 1 comment

#152 - Metric depth, focal depth estimation output/input

Issue - State: open - Opened by calledit 16 days ago - 2 comments

#151 - 推理出来全是0

Issue - State: open - Opened by hf62580 18 days ago - 3 comments

#150 - key error when using finetune model to infer

Issue - State: closed - Opened by nuanxinqing 22 days ago - 2 comments

#149 - About depthmap scale

Issue - State: open - Opened by wuqun-tju 23 days ago - 1 comment

#148 - Output Image Completely Black Using 'depth-anything/Depth-Anything-V2-Metric-Indoor-Large-hf' Model

Issue - State: closed - Opened by jamesbaker1 24 days ago - 1 comment

#147 - Metric Depth Scale Issue

Issue - State: open - Opened by LaFeuilleMorte 29 days ago - 1 comment

#146 - 仿射不变损失与SiLogLoss

Issue - State: open - Opened by zsw1456419654 29 days ago - 1 comment

#145 - How effective is this model in thermal infrared images?

Issue - State: open - Opened by puyiwen about 1 month ago

#144 - Details when training on TartanAir

Issue - State: open - Opened by luisdecker about 1 month ago - 2 comments

#143 - Metric Depth model FOV depth issue

Issue - State: open - Opened by daegyu93 about 1 month ago - 1 comment

#142 - How Should I Adjust the Focal Length Values in depth_to_pointcloud.py?

Issue - State: open - Opened by sergenerbay about 1 month ago

#141 - bad metric depth resuls when using huggingface transformers pipeline

Issue - State: open - Opened by MoAbbasid about 1 month ago - 2 comments

#140 - 3d point cloud from a video

Issue - State: open - Opened by gigasurgeon about 1 month ago - 3 comments

#139 - 你好，请问相对深度怎么微调到绝对深度？

Issue - State: open - Opened by Hlll-huang about 1 month ago - 1 comment

#138 - Virtual KITTI 2 datasets download

Issue - State: closed - Opened by nuanxinqing about 1 month ago

#137 - loss on Relative depth Model

Issue - State: open - Opened by PardisTaghavi about 1 month ago - 2 comments

#136 - Hugging face metric models dont produce depths

Issue - State: open - Opened by Eirikalb about 1 month ago - 4 comments

#135 - after download the vkitti2 datasets and use dish_tran.sh to start train, the traing phase looks normal, but when test, the predicted depth vaules are all 0

Issue - State: closed - Opened by visonpon about 1 month ago - 1 comment

#134 - 得到的一张深度图转点云，可视化异常，希望得到帮助！

Issue - State: open - Opened by chenqi13814529300 about 1 month ago

#133 - [Fix] image size before feeding into model

Pull Request - State: open - Opened by zafirshi about 1 month ago

#132 - wrong depth map of facial image

Issue - State: open - Opened by Time-Lord12th about 1 month ago - 1 comment

#131 - 获取超过8位的深度图或者数据

Issue - State: open - Opened by tang915915 about 1 month ago

#130 - Why is the default epoch much larger than V1?

Issue - State: open - Opened by 1171000410 about 1 month ago - 1 comment

#129 - Question on Quality of Peudo-Labels for Training Student Model

Issue - State: closed - Opened by nkinnaird about 1 month ago - 2 comments

#128 - Issue when converting depth to point cloud

Issue - State: open - Opened by mikami520 about 1 month ago - 4 comments

#127 - There is a large additive bias in the metric depth outputs

Issue - State: open - Opened by jbrownkramer about 1 month ago

#126 - Metric3D v2 predates Depth Anything v2 and is SOTA for some metrics but is not listed in the papers tables

Issue - State: open - Opened by jbrownkramer about 1 month ago

#125 - "xFormers not available" error when running the sample code

Issue - State: open - Opened by 1412-KID about 1 month ago - 4 comments

#124 - Feature Request - Fine tuning script/information

Issue - State: closed - Opened by RelaxingVR about 1 month ago - 2 comments

#123 - Size Mismatch for Depth-Anything-V2-Small Model

Issue - State: open - Opened by SuDaxia-kai about 1 month ago - 2 comments

#122 - Problem with numpy values of depth map

Issue - State: open - Opened by 520jz about 2 months ago - 1 comment

#121 - Why is the size of the predicted depth value reversed

Issue - State: open - Opened by 520jz about 2 months ago - 1 comment

#120 - The depth map obtained from the RGB-D camera dataset. Why is the depth map calculated from RGB images inconsistent?

Issue - State: open - Opened by chenqi13814529300 about 2 months ago

#119 - Why use max_depth in metric depth prediction?

Issue - State: open - Opened by 1punch3coins about 2 months ago - 3 comments

#118 - Sharper depth when interpolating to image size

Issue - State: open - Opened by felipecadar about 2 months ago - 1 comment

#117 - about metric indicators

Issue - State: open - Opened by lizhuqi233 about 2 months ago - 1 comment

#116 - operate on a live video

Issue - State: open - Opened by Kafka157 about 2 months ago - 5 comments

#115 - Question About Metric Depth Fine-tune

Issue - State: open - Opened by CDchenlin about 2 months ago - 2 comments

#114 - the precision of infering result

Issue - State: closed - Opened by SoulProficiency about 2 months ago - 3 comments

#113 - add inference script

Pull Request - State: closed - Opened by dvksn about 2 months ago

#112 - The predicted depth maps are not consistent

Issue - State: open - Opened by YacineDeghaies about 2 months ago - 3 comments

#111 - Harsh white lines appear around center object

Issue - State: open - Opened by fewjative about 2 months ago - 2 comments

#110 - Bad maxium depth for no reason

Issue - State: closed - Opened by SlamMate about 2 months ago

#109 - Outdoor datasets fine-tuning

Issue - State: closed - Opened by Edric-star about 2 months ago - 2 comments

#108 - Training student model on pseudo-labeled real images

Issue - State: closed - Opened by luochunfen about 2 months ago - 2 comments

#107 - Why choice of max-depth does not work？

Issue - State: open - Opened by Blissy-32 about 2 months ago - 1 comment

#106 - Normal model behavior ? or scaling ?

Issue - State: closed - Opened by YacineDeghaies about 2 months ago - 1 comment

#105 - Normal model behavior ? or scaling ?

Issue - State: open - Opened by YacineDeghaies about 2 months ago

#104 - segmentation fault running example

Issue - State: open - Opened by jeezrick about 2 months ago - 1 comment

#103 - cuda out of memory

Issue - State: open - Opened by xuan-cloud 2 months ago - 8 comments

#102 - How to reproduce finetuning relative depth model to metric depth mentioned in the paper?

Issue - State: open - Opened by Bearick 2 months ago - 3 comments

#101 - merge

Pull Request - State: closed - Opened by akshaygopalkr 2 months ago

#100 - I had problem fine-tuned model inference.

Issue - State: closed - Opened by Ahnsungmoon 2 months ago - 3 comments

#99 - Fine-tuning problems

Issue - State: open - Opened by Edric-star 2 months ago - 18 comments

#98 - How to deal with the too large value while transforming relative depth into absolute depth?

Issue - State: open - Opened by suzdl 2 months ago - 6 comments

#97 - Can't run in transformer.js

Issue - State: closed - Opened by gwendall 2 months ago - 1 comment

#96 - Is the output of fine-tuning the outdoor model using a depth map also a depth map?

Issue - State: open - Opened by HaosenZ 2 months ago - 1 comment

#94 - 请问怎么根据自己的数据集来微调模型呢？

Issue - State: open - Opened by lizhuqi233 2 months ago - 1 comment

#93 - What's the meaning of the model's output?How can I get the true distance from the model output?

Issue - State: open - Opened by duyuwen-duen 2 months ago - 17 comments

#92 - Problem about fine tune to get Metric Depth Model

Issue - State: open - Opened by command-z-z 2 months ago - 3 comments

#89 - Updated metric depth to point cloud code

Pull Request - State: open - Opened by 1ssb 2 months ago - 4 comments

#88 - The performance measurement results for metric depth are different from the values in the paper.

Issue - State: open - Opened by dnflslwlq 2 months ago - 4 comments

#87 - about the relative depth

Issue - State: open - Opened by QingshuiL 2 months ago - 3 comments

#86 - use bash dist_train.sh to train on my own datasets, the results are bad

Issue - State: open - Opened by visonpon 2 months ago - 4 comments

#85 - Temporal consistency

Issue - State: open - Opened by vitacon 2 months ago - 7 comments

#84 - HuggingFace checkpoints for fine-tuned Metric Depth Estimators

Issue - State: open - Opened by gchhablani 2 months ago - 2 comments

#82 - about pipeline

Issue - State: open - Opened by xuan-cloud 2 months ago - 2 comments

#70 - Custom depth map output format

Issue - State: closed - Opened by ThreeDeeJay 2 months ago - 11 comments

#69 - what min_depth and max_depth values to use for relative depth

Issue - State: open - Opened by YacineDeghaies 2 months ago - 4 comments

#68 - CUDA out of memory

Issue - State: open - Opened by alexdesion 2 months ago - 3 comments

#58 - Bad evaluation results.

Issue - State: closed - Opened by indu1ge 3 months ago - 1 comment

#48 - About Depth-Anything-V2-Giant 1.3B

Issue - State: open - Opened by Albertchamberlain 3 months ago - 2 comments

#46 - Training on my own dataset using relative depth maps instead

Issue - State: open - Opened by YacineDeghaies 3 months ago

#44 - how to get the grayscale output?

Issue - State: closed - Opened by MistyDragon7 3 months ago - 4 comments

#40 - Seems like the model outputs are disp image rather than depth image.

Issue - State: open - Opened by InFistLee 3 months ago - 7 comments

#36 - Error when trying to use sample code on CPU

Issue - State: closed - Opened by M3nxudo 3 months ago - 3 comments

#31 - Fine-tuning with my own dataset

Issue - State: open - Opened by XYQ0121 3 months ago - 4 comments

#26 - Worse metric depth than V1?

Issue - State: open - Opened by BaderTim 3 months ago - 6 comments

GitHub / DepthAnything/Depth-Anything-V2 issues and pull requests