# Model

Manage and train models

## Create Model

 - [POST /tailored-gen/models](https://docs.bria.ai/tailored-generation/model/create-model.md): Create new model.
Compatibility Rules:  FIBO Models: Must use a dataset with training_version = fibo.  Legacy Models: Must use a dataset with training_version in [max, light, 3.2, 2.3].  Cross-Compatibility:  Legacy datasets are generally interchangeable (e.g., a max dataset can train a light model). * FIBO is isolated. You cannot train a legacy model on a FIBO dataset or vice versa.

## Get Models

 - [GET /tailored-gen/models](https://docs.bria.ai/tailored-generation/model/get-models.md): Retrieve a list of models. If there are no models, an empty array is returned.

## Upload Model

 - [POST /tailored-gen/models/upload_model](https://docs.bria.ai/tailored-generation/model/upload-model.md): This API allows users to upload a pre-trained tailored model to Bria’s infrastructure, and run them within Bria’s ecosystem.

Important Note:: If you wish to upload a FIBO-based model, contact us for support.
Model has to be in .safetensors format. Maximum model size supported is 3GB.

You can find out the model's status by using the /tailored-gen/models/{model_id} route. Potential statuses are: "syncing" or "completed".

## Get Models by Project

 - [GET /tailored-gen/projects/{project_id}/models](https://docs.bria.ai/tailored-generation/model/get-models-by-project.md): Retrieve all models for a project. If there are no models, an empty array is returned.

## Get Model by ID

 - [GET /tailored-gen/models/{model_id}](https://docs.bria.ai/tailored-generation/model/get-model.md): Retrieve full model information including name, description, status (Created/InProgress/Completed/Failed/Stopping/Stopped), training version (Light/Max), generation prefix, project ID, dataset ID, and timestamps.

## Update Model

 - [PUT /tailored-gen/models/{model_id}](https://docs.bria.ai/tailored-generation/model/update-model.md): Update a model's name and description. Other model attributes such as training version and dataset cannot be modified after creation.

## Delete Model

 - [DELETE /tailored-gen/models/{model_id}](https://docs.bria.ai/tailored-generation/model/delete-model.md): Delete a specific model. Changes status to Deleted.

## Start Training

 - [POST /tailored-gen/models/{model_id}/start_training](https://docs.bria.ai/tailored-generation/model/start-training.md): Start model training. Training duration is typically 2-4 hours.
The associated dataset must have a status of 'completed' before training can begin.

Defaults by Training Version:
If parameters are not provided, default values are automatically selected based on the model's training_version:

* FIBO (fibo):
    * rank: 128
    * learning_rate: 0.0001
    * lr_scheduler: "constant_with_warmup"
    * total_training_steps: 750
    * checkpoint_interval: 500

* Legacy (max, light, 3.2, 2.3):
    * rank: 128
    * learning_rate: 0.0001
    * lr_scheduler: "constant"
    * total_training_steps: 1500
    * checkpoint_interval: 500
    
Constraints:
* Advanced training parameters are only supported when the model's training_mode is set to 'expert'.

## Stop Training

 - [POST /tailored-gen/models/{model_id}/stop_training](https://docs.bria.ai/tailored-generation/model/stop-training.md): Stop an ongoing model training process. Once stopped, training cannot be resumed - a new model would need to be created and trained.

## List Checkpoints

 - [GET /tailored-gen/models/{model_id}/checkpoints](https://docs.bria.ai/tailored-generation/model/list-checkpoints.md): Retrieve a list of all available checkpoints for a model. This is only available for models trained in expert mode.

## Get Specific Checkpoint

 - [GET /tailored-gen/models/{model_id}/checkpoints/{checkpoint_step}](https://docs.bria.ai/tailored-generation/model/get-checkpoint.md): Retrieve details for a specific model checkpoint by its step number.

## Delete Checkpoint

 - [DELETE /tailored-gen/models/{model_id}/checkpoints/{checkpoint_step}](https://docs.bria.ai/tailored-generation/model/delete-checkpoint.md): Permanently delete a specific model checkpoint. Deletion is not allowed if the checkpoint is currently selected for inference.

## Download Tailored Model

 - [GET /tailored-gen/models/{model_id}/download](https://docs.bria.ai/tailored-generation/model/download-tailored-model.md): Enables users to download a trained tailored generation model after completing the training process. 
The response includes a pre-signed URL for downloading the model, details about the base model used, 
and the prompt prefix applied during training.

To use the tailored model source code, access to the base model source code is required. 
The base model source code is exclusively available through Bria's Source Code and Weights product. 
For more information or to gain access, contact us at info@bria.ai.