Hi, About fine-tuing, after finishing the fine-tunes, later I want to add some other examples to my previous training data file and want to update my model data. Is it possible? If yes, do I have to pay for the new added examples for the update or for the whole data again?
yes,you will be charged for that also.
Yes, it is possible to fine-tune your fine-tuned models in your account.
You will only need to pay when you “fine tune again” for the new fine-tuning data. Your old / prior fine-tuning data has already been billed and processed into your saved / prior fine-tune model.
If an already optimized model is used, of course, only new data is paid.
I tested the training of a user model several times, the first time the trained model did not look bad, however, training that model again gave very bad results.
I personally decided to train each model with the base model from Open AI. I think it is better to pay a few euros more and get a better model. By the way, I used the Babbage model for the test
I would like to hear if anyone has better experiences than me.
Thank you for the clarification.
I have tried with curie, ada, but the result was bad. I don’t know if I missed some information but only davinci works fine for my project.
I have the same problem
Let me explain my problem.
I uploaded a JSONL file for fine-tuning data using this API: https://api.openai.com/v1/files. After uploading the file, I created the fine-tuning job using the API: https://api.openai.com/v1/fine_tuning/jobs. This process successfully created the fine-tuning model.
I have more training data to further train this model, but I am uncertain about the exact procedure for doing so. Additionally, I would like to know if I need to pay for the previous data in this process.
In this scenario, I would like to replace all the old training data with the new training data for this model and then proceed with the training. What is the correct process for achieving this?
Is anyone helping me to find the best way to do this?
I guess that would help - although it wouldn’t really help.
You have bumped a very old thread that no longer refers to the recommended models.
https://api.openai.com/v1/fine_tuning/jobs is the new endpoint for models such as gpt-3.5-turbo.
It does not support continuing training of a model. You must begin a new training job with your altered data or hyperparameters.
Continuing to update fine-tune models (likely still producing a new model name) is expected in the future.