Batch size in Evaluation Strategy

[TCBpenta8]

Hi,
Good to see a new idea in image captioning!

Here is my question. I noticed that in your evaluation stage, you input the sample one by one, that is in 'train_visualGPT.py' line 62. And also I found that you cast the batch size in validation and evaluation 5 times smaller than one in training. Will there be a specific reason for these? As I notice that the evaluation actually supports mini-batch strategy, and your strategy will cost tons of time if the evaluation set is huge.

I am kind of a freshman in this area so my question might be silly. Feel free to let me know what you think.

Regards,

[junchen14]

Hi,
Good to see a new idea in image captioning!

Here is my question. I noticed that in your evaluation stage, you input the sample one by one, that is in 'train_visualGPT.py' line 62. And also I found that you cast the batch size in validation and evaluation 5 times smaller than one in training. Will there be a specific reason for these? As I notice that the evaluation actually supports mini-batch strategy, and your strategy will cost tons of time if the evaluation set is huge.

I am kind of a freshman in this area so my question might be silly. Feel free to let me know what you think.

Regards,

you input the sample one by one, that is in 'train_visualGPT.py' line 62
-> it is not iterating the sample one by one, instead it is to iterate the samples batch by batch
also I found that you cast the batch size in validation and evaluation 5 times smaller than one in training.
-> it is not very necessary to control the batch size into a very smaller number in the evaluation stage, we make it 5 times smaller just in order to avoid some potential memory issues, but actually it does not happen in the real cases

[TCBpenta8]

😂 Thank you for your reply!