Since I lack the hardware to run the ORQA fine tuning code on local machine, I was trying to run the fine tuning code on Colab using its GPU runtime.
The thing I noticed is that Conda workaround I used for colab does not use colab's GPU resource at all.
Even then, I tried to run the the code for fine tuning and got error after some time.
I managed to run the REALM pre training with my own dataset and wanted to do fine tune ORQA to see its performance with my REALM pretraining. My question now is, is it even possible to run ORQA fine tuning on Colab? Are there any workarounds or things I'm missing?
Since I lack the hardware to run the ORQA fine tuning code on local machine, I was trying to run the fine tuning code on Colab using its GPU runtime. The thing I noticed is that Conda workaround I used for colab does not use colab's GPU resource at all. Even then, I tried to run the the code for fine tuning and got error after some time.
I managed to run the REALM pre training with my own dataset and wanted to do fine tune ORQA to see its performance with my REALM pretraining. My question now is, is it even possible to run ORQA fine tuning on Colab? Are there any workarounds or things I'm missing?