First of all, I would like to express my gratitude for your excellent research.
I have a question about using your code for inference or evaluation.
Initially, looking at your code, it seems that training on the Mistral model should take precedence because it appears that the pre-trained weights are not being shared.
So, if I want to infer using your model, should I proceed with training on the Mistral model to obtain the weights and then execute the evaluation code?
Because with slight modifications to the evaluation code, I've seen a few models capable of inference.
If I want to infer using your model, what should I do?
I haven't been able to look at all the code, but I couldn't find the inference-related code.
Once again, thank you for your excellent research. Thank you!
First of all, I would like to express my gratitude for your excellent research.
I have a question about using your code for inference or evaluation. Initially, looking at your code, it seems that training on the Mistral model should take precedence because it appears that the pre-trained weights are not being shared.
So, if I want to infer using your model, should I proceed with training on the Mistral model to obtain the weights and then execute the evaluation code? Because with slight modifications to the evaluation code, I've seen a few models capable of inference.
If I want to infer using your model, what should I do? I haven't been able to look at all the code, but I couldn't find the inference-related code.
Once again, thank you for your excellent research. Thank you!