Fine tune GPT 3.5 Turbo in Python. Step by step instructions for entire process

step by step instruction on how to fine tune GPT 3.5 Turbo in Python using OpenAI API. Everything you need to know about Fine tuning.
Code files are available to download for Patreon members along with 120+ other projects' files:
/ 88125041
Detailed overview of fine tuning:
kzbin.infoD4ncJbfKkmM?...
AUTO AGI agent video: • GPT agent that actuall...
Source code for AUTO AGI agent: / code-files-for-87530987
Voice controlled Auto AGI with swarm and multi self launch capabilities
• Voice controlled Auto ...
Search 140+ echohive videos and code download links:
www.echohive.live/
Chat with us on Discord:
/ discord
LLM Paper Summaries:
llmpapers.up.railway.app/
Try the GPT-4 Auto Coder app:
gpt4autocoder.streamlit.app/
GPT-4 web UI:
gpt-4.up.railway.app/
CHAPTERS:
00:00 What we will learn
01:20 Generating fine tuning dataset
04:08 Uploading your dataset
05:12 Starting a Fine Tuning job
06:11 Listing Fine Tuning jobs
07:04 Cancelling Fine Tuning jobs
07:20 Retrieving the progress of a Fine Tuning job
08:31 Deleting a fine tuned model
08:44 Tracking Fine Tuning progress in real time
12:07 Plotting Fine Tuning progress
#openai #openaitutorial #gpt #gpt4 #finetuning #artificialintelligence #largelanguagemodels #openaiapi #gpt4api

[@nickaboka8079]

Thank you, can’t wait to try this

[@echohive]

Fine tuning is a lot of fun! Glad you enjoyed the video.

[@dawn_of_Artificial_Intellect]

Memo, I can honestly say that you are underpaid. Great video, your instructions are always on point. Thanks again

[@echohive]

Thank you 🙏 I am glad you find the video useful. I think fine tuning will be a big change for many use cases. It is exciting :) Happy coding 🎸

[@DirkTomandl]

Awesome! I know this is just a 'hello world' type example, but did the fine-tuned model actually perform better on multiplications than the original model or GPT4? Besides: I just became a Patreon, simply because your content and channel are outstanding!

[@echohive]

I haven’t tested it in a meaningful way but with a few examples it still was making mistakes with 3 and 4 digits. I will do a more extensive test and let you know.

[@R0cky0]

Great content. Thanks for sharing. I sincerely wish your channel grow and prosper substantially!

[@echohive]

Thank you kindly for the encouraging words! 🙏❤️

[@georgesanchez8051]

As always, thank you for your work!

[@echohive]

Thank you as well :)

[@atrocitus777]

saw the other video about embedding the langchain docs would fine tunning it have better results at code completitions of libraries it hasn't seen before compared to using embeddings + gpt4?

[@echohive]

Fine tuning is different than and is not opposed to fine tuning. Think of fine tuning similar to building the model vs extra knowledge is welcome by the knowledge regardless of how it was trained.

[@samfights]

i want to Fine tune for writing articles.. what kind of data should i use to train it? is long text okay? like 800 words per article is ok?

[@echohive]

As long each training sample fits into context window then it is ok

[@samfights]

thank you @@echohive

[@lochuynh6734]

How can we set hyperparams for this gpt-3.5-turbo finetune? It seems to can not specify epoch, batch_size, learning_rate, ...

[@echohive]

You can specify epochs as far as I know. I don’t think you can specify the other hyper parameters. But I am not sure entirely.