Using Adapters
Last updated
Last updated
If the fine-tuning job finishes successfully, you can view additional information about the fine-tuned adapter by clicking on the adapter with the trained
status. For each adapter of this type, you can see the following details:
Adapter Name: The name of the fine-tuned adapter.
Adapter Status: The status of the fine-tuning job (trained
, training
, failed
).
Adapter Alias: The unique identifier for the fine-tuned adapter.
Created By: The user who created the fine-tuned adapter.
Creation Date: The date and time when the fine-tuned adapter was created.
When clicking on the adapter, you can see the details of the fine-tuned adapter, including the adapter alias, the model it was fine-tuned on, and when it was deployed. You can also see the code example for using this adapter in your API request.
Here is an example of the adapter details page:
To use a fine-tuned adapter, modify the model
parameter in your API request to reference the specific adapter's name. You should follow the standard format dynamiq/adapters/{adapter-alias}
for each of the adapters you want to use.
You can use the code example available for each adapter on the adapter details page (when clicking on it), or use the one provided for the base model on the deployment page. Either way, make sure to change the model
parameter to the adapter alias you want to use in a format of dynamiq/adapters/{adapter-alias}
.
For instance,
Adapter alias: mistral-lora-test-v2v3e7op
Model parameter: dynamiq/adapters/mistral-lora-test-v2v3e7op
Request for querying the adapter:
Or
Adapter alias: llama-8b-fine-tuning-test-djxmqmwj
Model parameter: dynamiq/adapters/llama-8b-fine-tuning-test-djxmqmwj
Request for querying the adapter: