Model customization for NeMo microservices
NeMo Customizer Training Container - NeMo Automodel
This container provides the training functionality for the NeMo Customizer ecosystem. It is specifically designed to run customization jobs that do not use Megatron Bridge optimized recipes, and is the default SFT training solution.
The NeMo Customizer training image executes the model training tasks and creates models that integrate with NVIDIA NIM for LLMs. It makes the jobs scale with the available nodes you have, supporting fine-tuning workloads on a single node with multiple GPUs or in multi-node, multi-GPU configurations.
Resources
Note: Use, distribution or deployment of this microservice in production requires an NVIDIA AI Enterprise License.
Governing Terms
The software and materials are governed by the NVIDIA Software License Agreement and the Product-Specific Terms for NVIDIA AI Products.