Sunflower7788

icon indicating an email [email protected]

icon location Segmentation; Detection; Image-Text Foundation Models;

Results 107 comments of Sunflower7788

试了这个openvino，报了错，不知道怎么解决 https://github.com/zhuo-yoyowz/openvino_notebooks/tree/main/notebooks/405-paddle-ocr-webcam

推荐用FD 您使用的repo不是我们官方维护的哈

ppocr INFO: best metric, hmean: 0, is_float16: False

您好，是训练自己的数据集吗？训练过程中损失是否正常下降呢？

cudaErrorInitializationError

您好，可以再检测一下图像是否正确，是否有损坏。另外可能是安装包和环境不对应，请自查下CUDA和Paddle版本

OCR检测v4server微调后可能出现【BGR通道或者RGB通道的图片】有其一检测不出结果

您好，OCR检测模型配置文件中的图像读入模式是 img_mode: BGR 只有其一能检测是正常的

请问训练ocr模型时使用ch_PP-OCRv4_rec.yml和rec_svtrnet_ch.yml的区别在哪里，哪个效果更好点？

是两个算法 svtrnet 的精度会更高一些，但是推理不支持变长，我们推荐用v4

IndexError: list index out of range

> > My recent commit should fix the problem. [81714a3](https://github.com/KaihuaTang/Scene-Graph-Benchmark.pytorch/commit/81714a337ada463f428b2d0a1521730c5c597571) > > Hi, I met the same problem, and I used the newest commit [1b955c6](https://github.com/KaihuaTang/Scene-Graph-Benchmark.pytorch/commit/1b955c608aa5c93ce25ee861bb60dea066f45f55#diff-a3e17e5e7b489553693b88e1f4f743564d8fcff78091374ad07df1b23c35aef9) > > > File "/home/xxx/Project/Scene-Graph-Benchmark.pytorch/maskrcnn_benchmark/data/datasets/visual_genome.py",...

how to get caption filter data for training stage1 of BLIP2 ?

Thanks. Will these filtered data be released?

how to get caption filter data for training stage1 of BLIP2 ?

thanks. By the way, do you use COCO Caption and VG these two dataset to train stage1? Or just CC3M CC12M SBU and LAION115M for stage1?

how to get caption filter data for training stage1 of BLIP2 ?

Is the code for CapFilter method released? no matter in BLIP or BLIP2.

表征模型中对标签列数据处理bug

您好，抱歉，我们暂未发现您所提到的问题，请您提供下运行代码，我们跟进复现下。