Closed aoji0606 closed 1 month ago
We use a better strategy to train Phi-3-Mini-based and Llama-3-8B-based Bunny, which would be open-sourced soon.
BTW, do you use Bunny-pretrain-laion-2m and Bunny-695k? It seems that your results are a little bit worse than expectation.
Yes, I use Bunny-pretrain-laion-2m for pre-train, and Bunny-695k for fine-tune
What's your pre-training learning rate?
1e-3 for pre-train 2e-5 for fine-tune
Do you use lora or full fine-tuning?
I use full fine-tuning
LoRA would convey a better performance.
ahhh, ok, thank you
hi~ What the learning rate should be set when using lora to fine-tune the llama3 model
As default, 2e-4 and 2e-5.
ok thank u
Close the issue for now if there's no further discussions. Feel free to reopen it if there's any other questions.
Thanks for your work.
I use the siglip and llama3 for pretrain and fine-tune,but the test results were poor,here are the results:
Why is that?