Fine-tune and deploy foundation models with Azure Machine Learning 

Thanks for the explanations

greate one Swati. its super helpful. small question, even though some of the models can be finetuned outside ml studio (e.g. flan t5 base) why can it not be done in ml studio.

Thank you Swati! This was a supremely helpful tutorial... Quick question - is it possible to: 1) Fine tune a base model with your own data, register it, and then... 2) Fine tune the newly fine tuned model with a different data set, (re-register it?), and deploy it? Also, can step 2 be repeated in perpetuity?

we need an tutorial on how to import open source models from hugging face such as mistral 7b?

❤

Thank you Swati. Is it possible to upload any other small language model which is not present model catalog. If yes then how to fine tune this model with our custom data set. Can we download these fine tune models. Could you please point us to any link which covers this

Democratizing but I can’t download my fine tuned model¿

Can we create same using Pay as you go subscription ?

I am planning to finetune GPT-3.5 Turbo with Azure AI. I have a PDF document consisting of 10 pages, containing approximately 17,000 characters and 3,200 tokens. Now, I intend to fine-tune it using GPT-3.5 Turbo. How long will it actually take to train the model on the specified dataset?

At around 3:48 you show the sample inference space, with a query box, I cannot see that. Any idea as to why that might be?

Why is Microsoft limiting data sources for Azure ML to only other Azure tools? Very disappointing.