When training with the Totaltext dataset, I encountered the following error:
ValueError: matrix contains invalid numeric entries.
In a previous issue, it was suggested to utilize the annotations provided in this repository. However, the link seems to be inaccessible. This issue extends to other annotations such as ICDAR2015, ICDAR2013, MLT, etc. Attempting to access these annotations results in an error message stating:
"Something went wrong, we're having some trouble. Please try again."
I've attempted to address this by exploring different techniques from the Onedrive community, but the error persists.
When training with the Totaltext dataset, I encountered the following error:
ValueError: matrix contains invalid numeric entries.
In a previous issue, it was suggested to utilize the annotations provided in this repository. However, the link seems to be inaccessible. This issue extends to other annotations such as ICDAR2015, ICDAR2013, MLT, etc. Attempting to access these annotations results in an error message stating:
"Something went wrong, we're having some trouble. Please try again."
I've attempted to address this by exploring different techniques from the Onedrive community, but the error persists.