No such file or directory: 'Your dataset path/caption_rm2048_train.csv'

[userkenny]

In the train command to below, It seems it did not find a file name "caption_rm2048_train.csv" from hdvila.py called. would you pls provide the file or guid me to get thru it. Thank you.

TORCH_DISTRIBUTED_DEBUG=DETAIL accelerate launch train_followyourpose.py --config="configs/pose_train.yaml"

Portion error log to beloa:

File "/home/cc/FollowYourPose/followyourpose/data/hdvila.py", line 109, in _load_metadata
with open(caption_path, 'r',encoding="utf-8") as csvfile: #41s
FileNotFoundError: [Errno 2] No such file or directory: 'Your dataset path/caption_rm2048_train.csv'

[mayuelala]

Hi，caption_rm2048_train.csv is the file in the HDVILA, maybe you should download this dataset firstly.

[userkenny]

I overlooked the hdvila.py file under the path. From code line# 53~58, for line#53, is it looking for a customized dataset file to train. I am not sure where to download the desired dataset to feed it. Any advises? thank you.Breadcrumbs 1. FollowYourPose <https://github.com/mayuelala/FollowYourPose/tree/main> 2. /followyourpose <https://github.com/mayuelala/FollowYourPose/tree/main/followyourpose> 3. /data <https://github.com/mayuelala/FollowYourPose/tree/main/followyourpose/data> /hdvila.py self.data_dir = 'Your dataset path' if dataset_set=='train': self.text_name = 'caption_rm2048_train.csv' else: self.text_name = 'caption_2048_val_new.csv~' self.meta_path = os.path.join(self.data_dir, self.text_name) mayuema ***@***.***> 於 2023年8月14日 週一 下午11:18寫道：

…

Hi，caption_rm2048_train.csv is the file in the HDVILA, maybe you should download this dataset firstly. — Reply to this email directly, view it on GitHub <#26 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/ADIC32JY6KPBW2M3YOKDNU3XVI6UZANCNFSM6AAAAAA3PAUHAA> . You are receiving this because you authored the thread.Message ID: ***@***.***>

[userkenny]

I studied for a while, is it started from below. https://ai.google.com/research/ConceptualCaptions/download CC Tai ***@***.***> 於 2023年8月15日 週二 上午10:21寫道：

…

I overlooked the hdvila.py file under the path. From code line# 53~58, for line#53, is it looking for a customized dataset file to train. I am not sure where to download the desired dataset to feed it. Any advises? thank you.Breadcrumbs 1. FollowYourPose <https://github.com/mayuelala/FollowYourPose/tree/main> 2. /followyourpose <https://github.com/mayuelala/FollowYourPose/tree/main/followyourpose> 3. /data <https://github.com/mayuelala/FollowYourPose/tree/main/followyourpose/data> /hdvila.py self.data_dir = 'Your dataset path' if dataset_set=='train': self.text_name = 'caption_rm2048_train.csv' else: self.text_name = 'caption_2048_val_new.csv~' self.meta_path = os.path.join(self.data_dir, self.text_name) mayuema ***@***.***> 於 2023年8月14日 週一 下午11:18寫道： > Hi，caption_rm2048_train.csv is the file in the HDVILA, maybe you should > download this dataset firstly. > > — > Reply to this email directly, view it on GitHub > <#26 (comment)>, > or unsubscribe > <https://github.com/notifications/unsubscribe-auth/ADIC32JY6KPBW2M3YOKDNU3XVI6UZANCNFSM6AAAAAA3PAUHAA> > . > You are receiving this because you authored the thread.Message ID: > ***@***.***> >

[Lionelding]

Same issue for caption_rm2048_train.csv, it is not available in https://github.com/microsoft/XPretrain/tree/main/hd-vila or https://github.com/microsoft/XPretrain/tree/main/hd-vila-100m.
Could you please point me to the place where you found it?
Appreciate your help

[Lionelding]

I studied for a while, is it started from below. https://ai.google.com/research/ConceptualCaptions/download CC Tai @.> 於 2023年8月15日 週二 上午10:21寫道：
…
I overlooked the hdvila.py file under the path. From code line# 5358, for line#53, is it looking for a customized dataset file to train. I am not sure where to download the desired dataset to feed it. Any advises? thank you.Breadcrumbs 1. FollowYourPose https://github.com/mayuelala/FollowYourPose/tree/main 2. /followyourpose https://github.com/mayuelala/FollowYourPose/tree/main/followyourpose 3. /data https://github.com/mayuelala/FollowYourPose/tree/main/followyourpose/data /hdvila.py self.data_dir = 'Your dataset path' if dataset_set=='train': self.text_name = 'caption_rm2048_train.csv' else: self.text_name = 'caption_2048_val_new.csv' self.meta_path = os.path.join(self.data_dir, self.text_name) mayuema @.> 於 2023年8月14日 週一 下午11:18寫道： > Hi，caption_rm2048_train.csv is the file in the HDVILA, maybe you should > download this dataset firstly. > > — > Reply to this email directly, view it on GitHub > <#26 (comment)>, > or unsubscribe > https://github.com/notifications/unsubscribe-auth/ADIC32JY6KPBW2M3YOKDNU3XVI6UZANCNFSM6AAAAAA3PAUHAA > . > You are receiving this because you authored the thread.Message ID: > @.***> >

Hmm, I did not find it. Could you please point me to the place where you found it? Thank you.

[treesturn]

anyone has a solution to this issue, where can i find the caption_rm2048_train.csv file?