facebookresearch / ov-seg

This is the official PyTorch implementation of the paper Open-Vocabulary Semantic Segmentation with Mask-adapted CLIP.
Other
689 stars 61 forks source link

Loss plots #9

Closed goncayilmaz closed 1 year ago

goncayilmaz commented 1 year ago

Hello,

I was wondering if you can provide the loss graphs from W&B. I am trying to train OVSeg myself and I am wondering how the graphs should look like.

Thank you in advance for your response.

Jeff-LiangF commented 1 year ago

Hi @goncayilmaz, As the checkpoints were trained half a year ago, unfortunately, I cannot find the W&B logs anymore. Just contact me if you have any questions about replicating the results. ThanksE

BKINGING commented 1 year ago

你好

我想知道您是否可以提供W&B的损失图表。我正在尝试自己训练 OVSeg,我想知道图形应该是什么样子。

提前感谢您的回复。

bash /media/hubu/data1/202131116023006_bky/ov-seg-main/open_clip_training/src/scripts/coco_gt_171cls_finetune_VitL.sh 2023-10-01,13:03:34 |资讯 |使用单个进程运行。设备库达:0。 2023-10-01,13:03:34 |资讯 |从OpenAI加载预训练的ViT-L-14。 2023-10-01,13:03:45 |资讯 |型号: 2023-10-01,13:05: 35 |资讯 |完成零镜头图像网。 2023-10-01,13:05:35 |资讯 |评估纪元: 0 ade150-零镜头-val-top1: 0.2742 ade150-零镜头-val-top5: 0.5465 2023-10-01,13:05:35 |资讯 |开始纪元 0 回溯(最近一次调用): 文件 “/home/hubu/anaconda3/envs/bpy38/lib/python3.8/runpy.py”,第 194 行,在_run_module_as_main 返回_run_code(code, main_globals, None, 文件 “/home/hubu/anaconda3/envs/bpy38/lib/python3.8/runpy.py”, 第 87 行, 在 _run_code exec(code, run_globals) 文件 “/media/hubu/Data1/202131116023006_bky/ov-seg-main/open_clip_training/src/training/main.py”, 第 320 行,在 main() 文件中 “/media/hubu/Data1/202131116023006_bky/ov-seg-main/open_clip_training/src/training/main.py”,第 268 行,在 main train_one_epoch(model, data, epoch, optimizer, scaler, scheduler, ema, args, writer) 文件 “/media/hubu/Data1/202131116023006_bky/ov-seg-main /open_clip_training/src/training/train.py”,第 69 行,在 train_one_epoch for i 中,批处理枚举(数据加载器): 文件 “/home/hubu/anaconda3/envs/bpy38/lib/python3.8/site-packages/torch/utils/data/dataloader.py”,第 521 行,在下一个 数据中 = self._next_data() 文件 “/home/hubu/anaconda3/envs/bpy38/lib/python3.8/site-packages/torch/utils/data/dataloader.py”,第 1203 行,_next_data 返回 self._process_data(data)

文件 “/home/hubu/anaconda3/envs/bpy38/lib/python3.8/site-packages/torch/utils/data/dataloader.py”, 第 1229 行,在 data.reraise() _process_data 文件中 “/home/hubu/anaconda3/envs/bpy38/lib/python3.8/site-packages/torch/_utils.py”中,第 434 行,在重新引发 异常 FileNotFoundError: 在 DataLoader 工作进程 0 中捕获 FileNotFoundError。 原始回溯(最近一次调用): 文件 “/home/hubu/anaconda3/envs/bpy38/lib/python3.8/site-packages/torch/utils/data/_utils/worker.py”,第 287 行,在 _worker_loop data = fetcher.fetch(index) 文件 “/home/hubu/anaconda3/envs/bpy38/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py”,第 49 行,在 fetch

data = [self.dataset[idx] for idx in possibly_batched_index] 文件 “/home/hubu/anaconda3/envs/bpy38/lib/python3.8/site-packages/torch/utils/data/_utils/fetch.py”,第 49 行,在 data = [self.dataset[idx] for idx in possibly_batched_index] 文件 “/media/hubu/Data1/202131116023006_bky/ov-seg-main/open_clip_training/src/training/data.py”,第 47 行,在 getitem images = self.transforms(Image.open(str(self.images[idx] ))) 文件 “/home/hubu/anaconda3/envs/bpy38/lib/python3.8/site-packages/PIL/Image.py”,第 3236 行,在 open 中 fp = builtins.open (文件名, “rb”) FileNotFoundError: [errno 2] 没有这样的文件或目录: 'coco_gt_171cls/2/000000129379.jpg'

这张照片在数据集中,但找不到错误,还能找到其他一些照片?