Notifications
Clear all
Topic starter
I've fine-tuned a model in Google Vertex AI named quotes-test2, using gemini-1.5-flash as the base.
I'm encountering an 'Unsupported Model' error message.
However, if I switch the model to gemini-1.5-flash within my node, it functions correctly.
How can I utilize a fine-tuned model in callin.io? I'm also open to exploring other platforms if they offer compatibility.
Information on your callin.io setup
- callin.io version: 1.80.3
- Running callin.io via callin.io cloud
Posted : 11/03/2025 1:55 pm
I'm not entirely sure as I haven't personally tested this, but it might be due to fine-tuned models from Google not supporting JSON mode.
Tuned Models
Tuned models have the following limitations:
- The input limit for a tuned Gemini 1.5 Flash model is 40,000 characters.
- JSON mode is not supported with tuned models.
- Only text input is supported.
Posted : 15/03/2025 5:38 pm
Topic starter
Thanks for the reply. Do you have any ideas on how to use any custom fine-tuned models with callin.io?
This is currently a major limitation for me using callin.io because I need better training options other than RAG and prompt engineering.
Posted : 18/03/2025 2:57 pm
I know it's definitely possible using HTTP requests or some basic AI nodes. I'm not certain if the AI Agent node supports this.
I was reviewing the update logs. This might be a bug, as they recently fixed fine-tuned OpenAI models to now be visible in the AI Agent node's list.
Posted : 22/03/2025 6:11 pm
This thread was automatically closed 90 days following the last response. New replies are no longer permitted.
Posted : 20/06/2025 6:11 pm