PaddlePaddle/PaddleSeg issues and pull requests

#3550 - [MLU] fix is_compiled_with_mlu api.

Pull Request - State: closed - Opened by ShawnNew 11 months ago
Labels: contributor

#3549 - config里的参数能否通过传参来修改

Issue - State: open - Opened by xdd12135 11 months ago - 1 comment
Labels: question

#3548 - 【Hackathon 5th No.68】轻量语义分割网络PIDNet

Pull Request - State: closed - Opened by flytocc 11 months ago - 21 comments
Labels: contributor

#3547 - Fastdeply下C++推理结果与predict.py结果差别大，该如何合理地设置export.py的参数？

Issue - State: open - Opened by pcycccccc 11 months ago - 2 comments
Labels: question

#3546 - 使用pp_liteseg模型，使用自己的数据集进行训练，crop_size的尺寸如何选择

Issue - State: open - Opened by shuaihutianxie 11 months ago - 3 comments
Labels: question

#3545 - 增加训练图像、推理图像、标签图像可视化

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

#3544 - 【Bug Fix】modnet推理问题

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

#3543 - 【Bug Fix】humanseg显存泄漏

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

#3542 - 验证并提升SAM+Clip在语义分割场景下的zero-shot分割精度

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

#3541 - 新增图文对话模型X-GPT

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

#3540 - VPD模型+下游任务（视觉感知、图像分割、深度估计）

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

#3539 - [Bug fixes]: Modify code and setting to run ACT correctly

Pull Request - State: closed - Opened by ranchongzhi 11 months ago - 1 comment
Labels: contributor, Contributor PR is Merged

#3538 - 增加类激活图

Issue - State: closed - Opened by chenjjcccc 11 months ago - 1 comment

#3537 - 新增early stop功能

Issue - State: open - Opened by chenjjcccc 11 months ago - 1 comment

#3536 - MedicalSeg增加滑窗推理功能

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

#3535 - CAT-Seg (CVPR'2023)模型复现

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

#3534 - PaddleSeg text-to-sam-clip 推理报错

Issue - State: open - Opened by SenVy-WH 11 months ago - 2 comments
Labels: bug

#3533 - Can ELseg mark cracks and damaged locations in buildings (irregular shapes)?

Issue - State: open - Opened by tanghaorang 11 months ago - 1 comment
Labels: Feature Request

#3532 - 关于模型输入大小NCW*H的疑问

Issue - State: open - Opened by pcycccccc 12 months ago - 2 comments
Labels: question

#3531 - 使用宫格标注，显示图片不全

Issue - State: open - Opened by fhb521314 12 months ago
Labels: question

#3530 - 能否在训练的可视化中，加入训练图像、推理图像、标签图像的显示？

Issue - State: open - Opened by Wst-sd 12 months ago - 2 comments
Labels: Feature Request

#3529 - 请问能否将MobileSAM的训练集成进来？

Issue - State: closed - Opened by Wst-sd 12 months ago
Labels: Feature Request

#3528 - PP-LiteSeg模型微调

Issue - State: open - Opened by shuaihutianxie 12 months ago - 3 comments
Labels: question

#3524 - Can't Download

Issue - State: open - Opened by rsinghpatawat 12 months ago - 1 comment

#3523 - pp_liteseg_stdc2 512x512的图，用3090推理800ms一张，这模型是不是有点大病

Issue - State: open - Opened by kkpssr 12 months ago - 1 comment
Labels: question

#3522 - Optimize the process to speed up inference

Pull Request - State: closed - Opened by weihuang-cs 12 months ago - 1 comment
Labels: contributor, Contributor PR is Merged

#3521 - 使用EISEG加载模型出错

Issue - State: open - Opened by shliar 12 months ago - 1 comment
Labels: bug

#3520 - 训练官方数据集报错

Issue - State: closed - Opened by suorui7 12 months ago - 2 comments
Labels: question

#3519 - [Enhancement]Move code to develop

Pull Request - State: closed - Opened by shiyutang 12 months ago - 1 comment

#3518 - 使用最原始方式开发 - 图像分割-transformer-vit-seg-notebook

Issue - State: open - Opened by gg22mm 12 months ago - 3 comments

#3517 - 问下paddleseg做语义分割二分类预测时如何划分阈值？

Issue - State: open - Opened by better464 12 months ago - 1 comment
Labels: question

#3516 - QualityInspector uad can train and inference using non-square size images?

Issue - State: open - Opened by qwzhuanghkclr 12 months ago - 1 comment
Labels: Feature Request

#3515 - Cannot train using fp16

Issue - State: open - Opened by rpl-alan-shangguan 12 months ago - 1 comment
Labels: bug

#3514 - [Fix]modify quantized ops during mkldnn int8 inference

Pull Request - State: closed - Opened by ranchongzhi 12 months ago - 1 comment
Labels: contributor, Contributor PR is Merged

#3513 - Add analysis for val dataset

Pull Request - State: closed - Opened by wuyefeilin 12 months ago - 1 comment

#3512 - 请问为什么RTFormer 模型导出后推理preprocess_time非常长？

Issue - State: open - Opened by DongLZY 12 months ago - 1 comment
Labels: question

#3511 - 请问分割模型中修改了骨干网络之后，如何训练骨干网络的预训练权重？

Issue - State: open - Opened by DongLZY 12 months ago - 1 comment
Labels: question

#3510 - 采用cityscape 训练2Dunet，在3090上，精度达不到文中给出的精度

Issue - State: open - Opened by HongweiChe about 1 year ago - 11 comments

#3509 - Update quick_start.md

Pull Request - State: closed - Opened by muare about 1 year ago - 1 comment
Labels: contributor, Contributor PR is Merged

#3508 - adapt to paddle about 0D-Tensor

Pull Request - State: closed - Opened by TingquanGao about 1 year ago - 1 comment

#3507 - Dev/adpt

Pull Request - State: closed - Opened by TingquanGao about 1 year ago - 1 comment

#3506 - 用C2fNet模型训练自己的数据集

Issue - State: open - Opened by wwt123-star about 1 year ago - 2 comments
Labels: question

#3505 - [cherry-pick #3424]fix 0D array error

Pull Request - State: closed - Opened by USTCKAY about 1 year ago

#3504 - 训练epoch计算值异常

Issue - State: closed - Opened by KKWY0909 about 1 year ago
Labels: bug

#3503 - 训练出错，错误显示：输入尺寸不一致，但是检查了图片输入尺寸是一致的，请大佬帮忙看看哪里出错了

Issue - State: open - Opened by loxoo6 about 1 year ago - 1 comment
Labels: bug

#3502 - ppmatting抠图时候如何保证去除人物背景同时，保留桌子/凳子

Issue - State: open - Opened by yeyuechen about 1 year ago - 5 comments
Labels: question

#3501 - 修改rtformer网络结构，每次跑的结果不同

Issue - State: open - Opened by aidongmandexiaowowo about 1 year ago - 6 comments
Labels: question

#3501 - 修改rtformer网络结构，每次跑的结果不同

Issue - State: closed - Opened by aidongmandexiaowowo about 1 year ago - 8 comments
Labels: question

#3500 - 删掉model中的rtformer.py文件，训练仍然能进行，想问问大家这是什么情况

Issue - State: open - Opened by Cuzyoung about 1 year ago - 2 comments
Labels: question

#3499 - 设置不同的batchsize时会导致训练开始时卡住，没有任何错误信息被记录或者打印

Issue - State: closed - Opened by BossCrab-jyj about 1 year ago - 1 comment
Labels: bug

#3498 - [cherry-pick]refine the doc and code of annotation images (#3497)

Pull Request - State: closed - Opened by juncaipeng about 1 year ago - 1 comment

#3497 - refine the doc and code of annotation images

Pull Request - State: closed - Opened by juncaipeng about 1 year ago - 1 comment

#3496 - 语义分割任务数据集中的图片是否可以shape不一致

Issue - State: closed - Opened by kaixin-bai about 1 year ago - 3 comments
Labels: question

#3495 - 使用paddleseg训练ocrnet，在验证的时候会占用极大的显存之后退出。

Issue - State: open - Opened by 1wang11lijian1 about 1 year ago - 7 comments
Labels: question

#3494 - 单机多卡训练时报错

Issue - State: closed - Opened by BossCrab-jyj about 1 year ago
Labels: question

#3493 - backbone_lr_mult can not be used. 无法使用该参数

Issue - State: closed - Opened by manfrey666 about 1 year ago - 2 comments
Labels: bug

#3492 - 请问数据增强之后，生成几张图呢？

Issue - State: closed - Opened by loxoo6 about 1 year ago - 3 comments
Labels: question

#3491 - [Feature]: Add PIDNet

Pull Request - State: closed - Opened by Asthestarsfalll about 1 year ago - 3 comments
Labels: contributor

#3490 - How to lock layers for fine-tuning

Issue - State: open - Opened by danielvandenberg95 about 1 year ago - 16 comments
Labels: question

#3490 - How to lock layers for fine-tuning

Issue - State: closed - Opened by danielvandenberg95 about 1 year ago - 5 comments
Labels: question

#3489 - tensorrt 加速报错 SM version not supported in this NVRTC version

Issue - State: open - Opened by tjdster about 1 year ago - 1 comment
Labels: question

#3488 - 学习率策略

Issue - State: closed - Opened by KKWY0909 about 1 year ago - 1 comment
Labels: question

#3487 - 使用val.py进行评估时，有多个class的指标，该如何判断是哪一个的？

Issue - State: open - Opened by MooTong123 about 1 year ago - 4 comments
Labels: question

#3487 - 使用val.py进行评估时，有多个class的指标，该如何判断是哪一个的？

Issue - State: open - Opened by MooTong123 about 1 year ago - 2 comments
Labels: question

#3486 - 对大批量数据进行人像分割推理时触发Linux OOM机制导致程序被kill

Issue - State: open - Opened by enemy1205 about 1 year ago - 3 comments
Labels: bug, question

#3485 - 剪枝之后模型的推理时延反而变慢了

Issue - State: open - Opened by censhallwe about 1 year ago
Labels: question

#3485 - 剪枝之后模型的推理时延反而变慢了

Issue - State: open - Opened by censhallwe about 1 year ago - 1 comment
Labels: question

#3484 - pre-trained

Issue - State: open - Opened by djl00 about 1 year ago
Labels: question

#3484 - pre-trained

Issue - State: open - Opened by djl00 about 1 year ago - 1 comment
Labels: question

#3483 - Bad predictions even though good validation result.

Issue - State: closed - Opened by bit-scientist about 1 year ago - 5 comments
Labels: question

#3483 - Bad predictions even though good validation result.

Issue - State: closed - Opened by bit-scientist about 1 year ago - 5 comments
Labels: question

#3482 - 训练模型和推理模型预测同一张图像结果相差太多

Issue - State: open - Opened by hunagjingwei about 1 year ago - 4 comments
Labels: bug, question

#3481 - [Feature] Update ACT

Pull Request - State: open - Opened by shiyutang about 1 year ago - 1 comment

#3481 - [Feature] Update ACT

Pull Request - State: closed - Opened by shiyutang about 1 year ago - 1 comment
Labels: contributor, Contributor PR is Merged

#3480 - How to use QualityInspector STFPM to train multiple types of data in MvTec AD dataset at the same time?

Issue - State: open - Opened by zzh805780186 about 1 year ago
Labels: question

#3480 - How to use QualityInspector STFPM to train multiple types of data in MvTec AD dataset at the same time?

Issue - State: open - Opened by zzh805780186 about 1 year ago
Labels: question

#3479 - [Feature] Add multi-label semantic segmentation support

Pull Request - State: closed - Opened by MINGtoMING about 1 year ago - 4 comments
Labels: contributor, Contributor PR is Merged

#3479 - [Feature] Add multi-label semantic segmentation support

Pull Request - State: open - Opened by MINGtoMING about 1 year ago - 4 comments
Labels: contributor

#3478 - 【模型量化】OCRNet/Segformer/PP-MobileSeg

Pull Request - State: open - Opened by marshall-dteach about 1 year ago - 1 comment
Labels: contributor

#3478 - 【模型量化】OCRNet/Segformer/PP-MobileSeg

Pull Request - State: closed - Opened by marshall-dteach about 1 year ago - 2 comments
Labels: contributor

#3477 - Coverting Modnet model for paddlelite gives Check failed: is_dispensable || is_have_output: Error

Issue - State: open - Opened by munibkhanali about 1 year ago
Labels: bug

#3476 - 请问config里面的model.pretrained和model.backbone.pretrained的区别是什么？

Issue - State: open - Opened by MooTong123 about 1 year ago - 8 comments
Labels: question

#3476 - 请问config里面的model.pretrained和model.backbone.pretrained的区别是什么？

Issue - State: open - Opened by MooTong123 about 1 year ago - 13 comments
Labels: question

#3475 - File cpp_inference_demo.sln

Issue - State: open - Opened by EthanAbitbol3 about 1 year ago
Labels: question

#3475 - File cpp_inference_demo.sln

Issue - State: open - Opened by EthanAbitbol3 about 1 year ago
Labels: question

#3474 - I want to use Docker, but authentication required issue occur.

Issue - State: open - Opened by PeterKim1 about 1 year ago
Labels: question

#3473 - Extract the lobes of the brain.

Issue - State: open - Opened by vantuyen-dev about 1 year ago - 1 comment
Labels: question

#3473 - Extract the lobes of the brain.

Issue - State: open - Opened by vantuyen-dev about 1 year ago - 1 comment
Labels: question

#3472 - 在aistudio可视化日志文件，网址打开后加载不出来。

Issue - State: open - Opened by loxoo6 about 1 year ago - 2 comments
Labels: bug

#3472 - 在aistudio可视化日志文件，网址打开后加载不出来。

Issue - State: open - Opened by loxoo6 about 1 year ago - 2 comments

#3471 - ValueError: (InvalidArgument) The type of data we are trying to retrieve (float32) does not match the type of data (uint8) currently contained in the container

Issue - State: open - Opened by bit-scientist about 1 year ago - 5 comments
Labels: question

#3471 - ValueError: (InvalidArgument) The type of data we are trying to retrieve (float32) does not match the type of data (uint8) currently contained in the container

Issue - State: open - Opened by bit-scientist about 1 year ago - 4 comments
Labels: question

GitHub / PaddlePaddle/PaddleSeg issues and pull requests