Manage Model Entities for Customization

Before running a customization job, you need to set up a Model Entity that points to your base model checkpoint. This section covers creating the required FileSet and Model Entity.

Task Guides

Create a Model FileSet

Create a FileSet containing your base model checkpoint from HuggingFace, NGC, or local storage.

Create a Model Entity

Create a Model Entity that references your FileSet and enables customization.

Key Concepts

What is a FileSet?

A FileSet is a collection of files managed by the platform. For customization, you create a FileSet containing:

Model weights (.safetensors, .bin, or .nemo files)
Model configuration (config.json)
Tokenizer files (tokenizer.json, tokenizer_config.json, and so on)

FileSets can be populated from:

HuggingFace Hub - Download directly from HF repositories
NGC - Download from NVIDIA NGC catalogs
Local upload - Upload files from your local machine

What is a Model Entity?

A Model Entity is the platform’s representation of a model. It contains:

FileSet reference - Points to where the model files are stored
Model Spec - Auto-populated metadata about the model architecture
Adapters - LoRA or other adapters attached to this model (populated after training)
Base model link - For fine-tuned models, links back to the parent

Quick Start Example

Complete example of setting up a model for customization:

HuggingFace Token: If downloading from a gated HuggingFace repository (like Llama models), you will need to create a secret containing your HuggingFace API token first. Refer to Manage Secrets for instructions.

1 import os
2 import time
3 from nemo_platform import NeMoPlatform
4 from nemo_platform._exceptions import ConflictError
5 from nemo_platform.types.files import HuggingfaceStorageConfigParam
6 
7 client = NeMoPlatform(
8     base_url=os.environ.get("NMP_BASE_URL", "http://localhost:8080"),
9     workspace="default",
10 )
11 
12 # Step 1: Create FileSet from HuggingFace
13 try:
14     fileset = client.files.filesets.create(
15         workspace="default",
16         name="qwen3-1.7b",
17         description="Qwen3 1.7B base model from HuggingFace",
18         storage=HuggingfaceStorageConfigParam(
19             type="huggingface",
20             repo_id="Qwen/Qwen3-1.7B",
21             repo_type="model",
22         ),
23     )
24     print(f"Created FileSet: {fileset.name}")
25 except ConflictError:
26     print("FileSet already exists, retrieving...")
27     fileset = client.files.filesets.retrieve(workspace="default", name="qwen3-1.7b")
28 
29 # Step 2: Create Model Entity
30 try:
31     model = client.models.create(
32         workspace="default",
33         name="qwen3-1.7b",
34         fileset="default/qwen3-1.7b",  # Reference to the FileSet
35         description="Qwen3 1.7B base model for customization",
36     )
37     print(f"Created Model Entity: {model.name}")
38 except ConflictError:
39     print("Model Entity already exists, retrieving...")
40     model = client.models.retrieve(workspace="default", name="qwen3-1.7b")
41 
42 # Step 3: Wait for ModelSpec to be auto-populated
43 print("Waiting for model spec to be populated...")
44 while not model.spec:
45     time.sleep(5)
46     model = client.models.retrieve(workspace="default", name="qwen3-1.7b")
47 
48 print(f"Model ready!")
49 print(f" Architecture: {model.spec.family}")
50 print(f" Parameters: {model.spec.base_num_parameters:,}")
51 print(f" Layers: {model.spec.num_layers}")

After the Model Entity is ready (has a spec), you can use it in customization jobs with model: "default/qwen3-1.7b".