Skip to content
FLORA FLORA API

Discover LoRA trainers and compatible models

GET
/loras/trainers
 
import { FloraClient } from '@flora-ai/flora';


const client = new FloraClient({ apiKey: process.env['FLORA_API_KEY'] });


const lora = await client.loras.listTrainers({
  workspace_id: 'ws_abc123',
});
console.log(lora);

Returns available LoRA families, selected trainers, image limits and supported training parameters with the same effective defaults as the web Train dialog for a dataset without manual labels. Also returns compatible visible inference model IDs and their parameter defaults/ranges. Pass the base_model key to POST /loras/train; when its saved Style is ready, use params.lora_id with one of the compatible inference models. Trainer image inputs and fixed output formats are managed by the training service and are not caller-configurable.

Error responses use the standard error body.

Authorizations

Section titled “Authorizations ”

Parameters

Section titled “ Parameters ”

Query Parameters

Section titled “Query Parameters ”
workspace_id
required

Workspace for trainer discovery and subsequent training

string
/^ws_\S+$/

Workspace for trainer discovery and subsequent training

Responses

Section titled “ Responses ”

200

Section titled “200 ”

Available LoRA trainers returned.

Media type application/json
object
trainers
required

Available trainers using the same visibility and selection rules as the web Train dialog

Array<object>
object
base_model
required

Family key to pass as base_model when training

string
model_id
required

Selected trainer endpoint identifier; use the LoRA training endpoint to invoke it

string
name
required

Trainer display name

string
image_count
required

Image-count limits enforced by the shared training flow

object
min
required

Minimum number of training images

integer
max
required

Maximum number of training images

integer
image_extensions
required

Accepted training-image formats

Array<string>
params
required

Caller-configurable trainer parameters with effective web/API defaults for a dataset without manual labels. Pass overrides in trainer_params.

Array<object>
object
name
required

Parameter name to pass in generation params

string
type
required

Parameter value type

string
Allowed values: string string[] bool int int? seed float dict
required
required

Whether the model requires this parameter

boolean
default

Default parameter value

min

Minimum numeric value

number
max

Maximum numeric value

number
label

Human-readable parameter label

string
description

Parameter help text

string
options

Allowed values for enum-like parameters

Array<object>
object
label
required

Displayed option label

string
value
required

Option value to pass in generation params

string
description

Option description

string
properties

Nested numeric properties for object parameters

object
key
additional properties
object
min
required
number
max
required
number
default
required
number
inference_models
required

Visible inference models accepting this trainer’s family

Array<object>
object
model_id
required

Compatible inference endpoint identifier for POST /generate

string
name
required

Inference model display name

string
lora_id_parameter
required

Put the returned sty_ Style identifier in this params field

Allowed value: lora_id
params
required

Inference parameters, including the model’s strength default and range

Array<object>
object
name
required

Parameter name to pass in generation params

string
type
required

Parameter value type

string
Allowed values: string string[] bool int int? seed float dict
required
required

Whether the model requires this parameter

boolean
default

Default parameter value

min

Minimum numeric value

number
max

Maximum numeric value

number
label

Human-readable parameter label

string
description

Parameter help text

string
options

Allowed values for enum-like parameters

Array<object>
object
label
required

Displayed option label

string
value
required

Option value to pass in generation params

string
description

Option description

string
properties

Nested numeric properties for object parameters

object
key
additional properties
object
min
required
number
max
required
number
default
required
number
Example 
{
  "trainers": [
    {
      "params": [
        {
          "type": "string"
        }
      ],
      "inference_models": [
        {
          "lora_id_parameter": "lora_id",
          "params": [
            {
              "type": "string"
            }
          ]
        }
      ]
    }
  ]
}