Comment on page
🔧
Continuous Fine-Tuning [Beta]
With Portkey, it becomes very easy to 1️) Curate & create data for fine-tuning, 2) Schedule fine-tuning jobs on OpenAI or Anyscale, and 3) Use the fine-tuned models!
- 1.Initialization: Set up your task with a name, description, AI provider key, model, and epoch count.
- 2.Training Data: Use logs from Portkey or upload a JSONL file for fine-tuning. View essential details like cost per epoch, and more.
- 3.Validation Data (Optional): Add optional validation data using Portkey logs or JSONL upload.
- 4.Review and Execution: Review the final cost estimates, and schedule the fine-tuning job! Use the fine-tuned model with Portkey playground or replace existing models in your code with the fine-tuned models.
Model | Provider |
|---|---|
| OpenAI |
| Anyscale |
This feature is under development, but we are opening it up to select orgs that are excited to try it out. To get access, please join the Portkey Discord and send a "Hi" to anyone on Team Portkey, you'll be invited to the private channel for the beta.
- 1.Better cost calculations
- 2.Simplified Portkey slugs instead of long OpenAI model slugs
- 3.Auto split logs into training & validation
- 4.Loss charts while training
- 5.Fetch fine-tuned models that are not on Portkey, and let you further fine-tune them
Last modified 16d ago