Update an existing Customization Config#

Prerequisites#

Before you can update a customization configuration, make sure that you have:

  • Access to the NeMo Customizer service

  • Set the CUSTOMIZER_BASE_URL environment variable to your NeMo Customizer service endpoint

export CUSTOMIZER_BASE_URL="https://your-customizer-service-url"

To Update an Existing Customization Config#

Choose one of the following options to update an existing customization config.

import os
from nemo_microservices import NeMoMicroservices

# Initialize the client
client = NeMoMicroservices(
    base_url=os.environ['CUSTOMIZER_BASE_URL']
)

# Update customization config
updated_config = client.customization.configs.update(
    config_name="llama-3.1-8b-instruct@v1.0.0+A100",
    namespace="default",
    description="Updated description",
    max_seq_length=4096
)

print(f"Updated config: {updated_config.name}")
print(f"New description: {updated_config.description}")
print(f"Max sequence length: {updated_config.max_seq_length}")

curl -X PATCH "${CUSTOMIZER_BASE_URL}/v1/customization/configs/default/llama-3.1-8b-instruct@v1.0.0+A100" \
  -H 'accept: application/json' \
  -H 'Content-Type: application/json' \
  -d '{
   "description": "Updated description",
   "max_seq_length": 4096
  }' | jq

Note

The update endpoint supports many additional parameters beyond description and max_seq_length, including:

  • Training Options Management: training_options, add_training_options, remove_training_options

  • Templates: prompt_template, chat_prompt_template

  • Hardware: pod_spec, training_precision

  • Metadata: project, custom_fields, ownership

  • Data: dataset_schemas

Training options are identified by the combination of training_type and finetuning_type. You can update existing options, add new ones, or remove specific combinations as needed.

Example Response
{
 "created_at": "2024-11-26T02:58:55.339737",
 "updated_at": "2024-11-26T03:58:55.339737",
 "id": "customization_config-MedVscVbr4pgLhLgKTLbv9",
 "name": "llama-3.1-8b-instruct@v1.0.0+A100",
 "namespace": "default",
 "description": "Updated description",
 "target": {
     "id": "customization_target-A5bK7mNpR8qE9sL2fG3hJ6",
     "name": "meta/llama-3.1-8b-instruct@2.0",
     "namespace": "default",
     "base_model": "meta/llama-3.1-8b-instruct",
     "enabled": true,
     "num_parameters": 8000000000,
     "precision": "bf16",
     "status": "ready"
 },
 "training_options": [
     {
         "training_type": "sft",
         "finetuning_type": "lora",
         "num_gpus": 2,
         "num_nodes": 1,
         "tensor_parallel_size": 1,
         "pipeline_parallel_size": 1,
         "micro_batch_size": 1,
         "use_sequence_parallel": false
     }
 ],
 "training_precision": "bf16",
 "max_seq_length": 4096,
 "pod_spec": {
     "node_selectors": {
         "nvidia.com/gpu.product": "NVIDIA-A100-SXM4-80GB"
     },
     "annotations": {
         "nmp/job-type": "customization"
     },
     "tolerations": [
         {
             "key": "app",
             "operator": "Equal",
             "value": "customizer",
             "effect": "NoSchedule"
         }
     ]
 },
 "prompt_template": "{input} {output}",
 "chat_prompt_template": null,
 "dataset_schemas": [],
 "project": null,
 "ownership": {}
}