You can finetune any of the supported models on your datasets. Here’s how you can do it.

Note : The data set needs to be processed before you start the finetuning job.Once the data set is processed navigate to finetuning and start a job

Here is what each of the fields mean:

  1. Job Name: Name of the job for your reference.
  2. Finetune Type: Type of finetuning you want to do. You can learn more here.
  3. Base Model: The base model for your finetuning. Read more about the supported models here.
  4. File ID or File name: The file ID or name of the dataset you want to train your model on.
  5. (Advanced) Epochs: Number of epochs you want to train your model for. It is set to 2 by default, but you can increase it if the amount of data you have is less.
  6. (Advanced) Learning rate: Learning rate for the finetune. Higher learning rate has a higher chance of catastrophic forgetting (for full-finetunes especially), so please be careful when tuning this parameter.

Once you have filled in the details, you will be able to see the estimated cost for the finetuning. Please make sure you have enough credits in your account to start the finetuning.

[photo]

Manage datasetsQuery finetuned models