Ucas-HaoranWei / GOT-OCR2.0

Official code implementation of General OCR Theory: Towards OCR-2.0 via a Unified End-to-end Model
6k stars 514 forks source link

Request for Assistance Regarding Attention Mask and Pad Token ID Warning Affecting Model Performance #97

Open AlexandraDobrescu opened 1 month ago

AlexandraDobrescu commented 1 month ago

I encountered the following warning while running the model: "The attention mask and the pad token ID were not set. As a result, you might experience unexpected behavior. To ensure reliable results, please provide an attention_mask with your input. The pad_token_id has been set to eos_token_id: 151643 for open-end generation." I suspect this may be impacting the model's performance. Could you offer guidance on how to address this issue? Any assistance would be greatly appreciated.

TaoSeekAI commented 4 weeks ago

Same issue Any ideas ?

tairenchen commented 3 weeks ago

Same problem, any solutions?

LifeRainSunshine commented 2 weeks ago

一样的问题 求解决

450711420 commented 2 weeks ago

same problem please share the solution if u guys solve this already thanks :)

Ucas-HaoranWei commented 2 weeks ago

I think it is because of the tokenizer version. My tokenizer version is 0.15.2.

KlausFan commented 1 week ago

same problem image