SysCV/sam-hq issues and pull requests

#150 - How to calculate the Boundary AP

Issue - State: open - Opened by wliu20 13 days ago

#149 - Will SAM-HQ2 code be kept up to date with official sam2 code?

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

#148 - HQ-SAM 2: Training Details

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

#147 - if i want to apply sam-hq on a 15 classes dataset ?

Issue - State: open - Opened by sym330 3 months ago - 1 comment

#146 - Update sam-hq2 eval seginw

Pull Request - State: closed - Opened by ymq2017 3 months ago

#145 - Question about HQ-Output Token and weight updates in the frozen Mask Decoder

Issue - State: open - Opened by Linn0910 5 months ago

#144 - Train Light HQ-SAM

Issue - State: open - Opened by bobo59 5 months ago

#143 - Transform samhq into tensorrt!

Issue - State: open - Opened by zhujiajian98 6 months ago

#142 - About the console output during validation.

Issue - State: open - Opened by FolkScientistInDL 6 months ago - 1 comment

#141 - Input mask prompt

Issue - State: open - Opened by theFilipko 6 months ago

#140 - High-level question: is SAM-HQ suited for fine-tuning on specific tasks?

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

#139 - Evaluation for instance segmentation

Issue - State: closed - Opened by wliu20 8 months ago

#138 - RuntimeError

Issue - State: open - Opened by hz-1 8 months ago

#137 - no such pth file

Issue - State: closed - Opened by hz-1 8 months ago - 1 comment

#136 - Positive negative point inputs for SAM-hq

Issue - State: open - Opened by varadtechx 9 months ago

#135 - Does SAM_HQ support batch input?

Issue - State: open - Opened by yunniw2001 9 months ago - 1 comment

#134 - About the issue of positive and negative sample pairs

Issue - State: open - Opened by leinusi 10 months ago - 1 comment

#133 - can't we get candidates as original sam?

Issue - State: open - Opened by TikaToka 10 months ago

#132 - Visualization of Figure 6

Issue - State: open - Opened by tqinger 10 months ago

#131 - Why can't we achieve the demonstration effect? There are three tennis rackets in the incoming image, and only one mask is returned

Issue - State: open - Opened by hjj-lmx 10 months ago

#130 - How to train vit_tiny (Light HQ-SAM for real-time need): ViT-Tiny HQ-SAM model?

Issue - State: open - Opened by Andy718811 10 months ago - 2 comments

#129 - Runtime error in Colab

Issue - State: open - Opened by gttae 10 months ago - 1 comment

#128 - About interm_embeddings

Issue - State: open - Opened by zzzyzh 11 months ago

#127 - Alternative implementation in Refiners

Issue - State: closed - Opened by hugojarkoff 11 months ago

#126 - Can sam-hq perform breakpoint training: continue training on the already trained sam-hq, or continue training on the official sam-hq weights

Issue - State: open - Opened by YUANMU227 12 months ago - 2 comments

#125 - Segment Anything CPP Wrapper for macOS

Issue - State: open - Opened by ryouchinsa 12 months ago

#124 - Training problem: During the training of sam-hq, the iou output of the val set is very high, 0.98; but in eval mode, the iou of the val set is only 0.48

Issue - State: open - Opened by YUANMU227 12 months ago

#123 - RuntimeError: Sizes of tensors must match except in dimension 1. Expected size 1 but got size 13 for tensor number 1 in the list.

Issue - State: closed - Opened by stevezkw1998 about 1 year ago - 1 comment

#122 - Can we add support for transformers?

Issue - State: open - Opened by moneyhotspring about 1 year ago - 3 comments

#121 - RuntimeError: Unexpected error from cudaGetDeviceCount(). Did you run some cuda functions before calling NumCudaDevices() that might have already set an error? Error 804: forward compatibility was attempted on non supported HW

Issue - State: closed - Opened by stevezkw1998 about 1 year ago

#120 - Question about Ablation study

Issue - State: closed - Opened by mcshih about 1 year ago - 4 comments

#119 - Can I have a more specific using example (code snippet) in README.md?

Issue - State: open - Opened by stevezkw1998 about 1 year ago - 5 comments

#118 - About the number of detect boxes

Issue - State: open - Opened by yaohusama about 1 year ago - 2 comments

#117 - The purpose of self.embedding_maskfeature

Issue - State: open - Opened by ghost about 1 year ago - 3 comments

#116 - How can I test the model

Issue - State: open - Opened by Ryanye2000 about 1 year ago - 1 comment

#115 - The ap of lvis measured by the public model is inconsistent with the ap in the paper.

Issue - State: open - Opened by yaohusama about 1 year ago - 1 comment

#114 - About obtaining uvo data set ap or boundary ap

Issue - State: open - Opened by yaohusama about 1 year ago - 5 comments

#113 - Request for evaluation code

Issue - State: open - Opened by jameslahm about 1 year ago - 7 comments

#112 - Will a fast version be released like segment-anything-fast?

Issue - State: open - Opened by yatengLG about 1 year ago - 3 comments

#111 - error in 'demo_hqsam_automatic_mask_generator.ipynb'

Issue - State: open - Opened by Kyle-RuidongLI about 1 year ago

#110 - dcm file input questions

Issue - State: open - Opened by Ryanye2000 about 1 year ago

#109 - dcm file imput

Issue - State: closed - Opened by Ryanye2000 about 1 year ago

#108 - Multi-box input To use multi-box prompt inputs

Issue - State: closed - Opened by yliuosu about 1 year ago - 1 comment

#107 - The val_iou and val_boundary_iou is very low

Issue - State: open - Opened by Ghy1209 about 1 year ago - 1 comment

#106 - Question: Is it SAM-HQ model applicable for predicting segmentation mask for the input images without boxes, point or label?

Issue - State: open - Opened by mzg0108 about 1 year ago - 2 comments

#105 - some questions about training SAM_HQ

Issue - State: open - Opened by Ghy1209 about 1 year ago - 2 comments

#104 - Tuning Segment for clothing

Issue - State: open - Opened by crapthings about 1 year ago

#103 - when i run train.py with one gpu and dataset that sam_hq used, the process stopped and didn't move

Issue - State: open - Opened by Ryanye2000 about 1 year ago - 5 comments

#102 - Has anyone attempted to deploy SAM-Hq on CVAT (following the deployment of SAM)

Issue - State: open - Opened by superNCS about 1 year ago

#101 - Grounded HQ-SAM

Issue - State: open - Opened by halqadasi about 1 year ago - 1 comment

#100 - When i tried to train the mode. There is a bug

Issue - State: open - Opened by Ryanye2000 about 1 year ago - 6 comments

#99 - I want to ask if there is a way that i give the tif image and bbox label as my dataset

Issue - State: open - Opened by Ryanye2000 about 1 year ago - 1 comment

#98 - the code of Light HQ-SAM vs MobileSAM on COCO using the FocalNet-DINO

Issue - State: open - Opened by Ghy1209 about 1 year ago

#97 - May i ask if there is a way to use dcm image to train the model? If its available, what should i do.

Issue - State: open - Opened by Ryanye2000 over 1 year ago - 2 comments

#96 - Text Prompt

Issue - State: open - Opened by sebastianopazo1 over 1 year ago - 1 comment

#95 - Questions about training class information.

Issue - State: open - Opened by sweetdream33 over 1 year ago

#94 - how to use multiple gpu?

Issue - State: open - Opened by zouxiaodong over 1 year ago

#93 - when use GroundingDINO as the detector

Issue - State: open - Opened by Ghy1209 over 1 year ago - 3 comments

#92 - The label problem

Issue - State: closed - Opened by code-797 over 1 year ago

#91 - How to output the visulization of segmentation and object detection when test on seginw?

Issue - State: open - Opened by Ghy1209 over 1 year ago

#90 - ImportError: cannot import name 'FrozenBatchNorm2d' from 'torchvision.ops'

Issue - State: open - Opened by Guangpengwgp1202 over 1 year ago - 2 comments

#89 - How to export a fully segmented model onnx

Issue - State: open - Opened by LsyAi over 1 year ago

#88 - Inconsistent with the iou of the paper

Issue - State: open - Opened by yaohusama over 1 year ago

#87 - Training loss is Nan

Issue - State: open - Opened by sweetdream33 over 1 year ago - 3 comments

#86 - Can the train.py ssupport the training for Light HQ-SAM?

Issue - State: open - Opened by rbli-john over 1 year ago - 1 comment

#85 - Support multi-classes or multi-objects segmentation?

Issue - State: closed - Opened by code-797 over 1 year ago - 1 comment

#84 - IoU Computation Problem

Issue - State: open - Opened by MrZilinXiao over 1 year ago - 2 comments

#83 - Fix inconsistent pip package name that breaks `pip install` of requirements.txt

Pull Request - State: open - Opened by rtyley over 1 year ago

#82 - the fine-tune question

Issue - State: open - Opened by zy199676 over 1 year ago - 4 comments

#81 - Are the text prompts in the demo used directly as input to sam-hq?

Issue - State: closed - Opened by hyosong86 over 1 year ago - 1 comment

#80 - compute_iou()?

Issue - State: closed - Opened by sunsunnyshine over 1 year ago - 4 comments

#79 - Question about learnable parameters.

Issue - State: closed - Opened by Nastu-Ho over 1 year ago - 2 comments

#78 - Problems with visualizing results

Issue - State: open - Opened by sjc6834 over 1 year ago - 4 comments

#77 - Correlation Between Fine-Tuning Model Files and Demo Pretrained Models

Issue - State: open - Opened by RyosukeSakaguchi over 1 year ago - 6 comments

#76 - How to find object name

Issue - State: open - Opened by victorhuzhening over 1 year ago - 1 comment

#75 - onnx export script does not support vit_tiny?

Issue - State: open - Opened by shenjun1994 over 1 year ago - 3 comments

#74 - how to train the model from previous saved checkpoints

Issue - State: closed - Opened by yliuosu over 1 year ago - 2 comments

#73 - The fine-tuned HQ-SAM model shows a significant improvement in accuracy. However, the val_iou_0 accuracy of HQ-SAM during training is very low.

Issue - State: closed - Opened by geoexploring over 1 year ago - 2 comments

#72 - how to generate a sliding comparison gif file in the Visual comparison part

Issue - State: closed - Opened by yliuosu over 1 year ago - 2 comments

#71 - How to generate embedding file using SAMHQ?

Issue - State: open - Opened by moneyhotspring over 1 year ago - 2 comments

#70 - Error when trying to load embedding image in browser using onnx

Issue - State: open - Opened by moneyhotspring over 1 year ago - 3 comments

#69 - Performing batch image detection using a for loop can lead to memory leakage.

Issue - State: closed - Opened by morestart over 1 year ago - 2 comments

#68 - How to export the tensorrt model？

Issue - State: open - Opened by Sencc over 1 year ago - 1 comment

#67 - DecompressionBombWarning

Issue - State: closed - Opened by wulele2 over 1 year ago - 1 comment

#66 - Where is the training script for Light HQ-SAM?

Issue - State: open - Opened by zpge over 1 year ago - 2 comments

#65 - When using my own data for training, the first epoch reached an IOU value of over 0.9. May I ask if there is a problem? The data is 5000

Issue - State: closed - Opened by zy199676 over 1 year ago - 1 comment

#64 - Why does my implementation of sam-hq light takes longer time?

Issue - State: open - Opened by mickyzyf over 1 year ago - 1 comment

#64 - Why does my implementation of sam-hq light takes longer time?

Issue - State: open - Opened by mickyzyf over 1 year ago - 1 comment

#63 - Killing Subprocess during training

Issue - State: closed - Opened by vishakhalall over 1 year ago - 2 comments

#62 - When I use ''multimask_output=True'', only one mask is generated

Issue - State: open - Opened by BigBigMonsterProgramer over 1 year ago - 2 comments

#61 - Hello author, how do you measure loss_mask and loss dice, do you have any references?

Issue - State: open - Opened by moliniao over 1 year ago - 2 comments

GitHub / SysCV/sam-hq issues and pull requests