Skip to main content
POST
/
adaption
/
train
Trigger Training
curl --request POST \
  --url https://api.example.com/adaption/train \
  --header 'Authorization: <authorization>' \
  --header 'Content-Type: application/json' \
  --data '
{
  "memory_group": "<string>",
  "lora_name": "<string>",
  "force": true
}
'
{
  "status": "<string>",
  "message": "<string>",
  "lora_name": "<string>",
  "version": 123,
  "memories_count": 123,
  "job_id": "<string>"
}

​
POST /adaption/train

Trigger LoRA training for a memory group. This endpoint can be called automatically when the training threshold is met, or manually by the user.
Adaption is currently in Beta and may not be stable. Users may also encounter slower outputs due to vLLM cold-start times.

​
Authentication

​
Authorization
string
required
Bearer token with your API key. Format: Bearer YOUR_ASYMMETRIC_API_KEY

​
Request Body

​
memory_group
string
required
The memory group to train on
​
lora_name
string
required
Name for the LoRA adapter
​
force
boolean
default:"false"
Force override stale training locks

​
Response

​
status
string
Training status: training_started, already_training, or no_data
​
message
string
Status message
​
lora_name
string
Name of the LoRA adapter
​
version
integer
Training version number
​
memories_count
integer
Number of memories used for training
​
job_id
string
Async job identifier

​
Example

import requests

response = requests.post(
    "https://rkdune--symmetry.modal.run/adaption/train",
    headers={"Authorization": "Bearer YOUR_ASYMMETRIC_API_KEY"},
    json={
        "memory_group": "darwin_agent",
        "lora_name": "darwin_adapter",
        "force": True
    }
)

data = response.json()
print(f"Status: {data['status']}")
print(f"Job ID: {data.get('job_id')}")

​
Errors

StatusDescription
400memory_group and lora_name are required
401Invalid API key
500Training failed