NVIDIA Corporation

Helps configure training across multiple GPUs or machines in NeMo AutoModel. Guides strategy selection, YAML settings, and parallelism choices that split model computation or data, with constraints for large models, long sequences, and expert models.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Choosing a strategy for multi-GPU training
• Configuring parallelism for large models
• Splitting long input sequences across GPUs

Improve MoE training throughput in Megatron Bridge. MoE models divide work among expert networks; this skill identifies memory, communication, and compute bottlenecks, guides controlled configuration comparisons, and validates speed alongside training health.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Investigate training slowdowns after code or configuration changes
• Find memory, communication, and compute bottlenecks
• Compare parallelism, precision, and dispatcher settings

Choose Megatron Bridge training configurations for a model, GPU allocation, and training goal. Get starting commands and parallel processing adjustments, with clear distinctions between changes that affect learning behavior, execution tuning, and benchmark-only settings.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Choose a starting configuration for pretraining or fine-tuning
• Adapt parallel processing settings to available GPUs
• Compare training and throughput benchmark configurations

Configure fault detection and recovery for distributed training in Megatron Bridge. The skill covers slow GPU detection, automatic restarts, job interruption handling, and checkpoint saving, with configuration examples and guidance for investigating hangs or failed recovery.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Configuring automatic recovery after training failures
• Detecting slow GPUs in distributed training
• Preparing for scheduled job interruptions

Choose how to split large-model training across GPUs in Megatron Bridge. This guide helps combine parallelism methods, estimate GPU requirements, match settings to hardware connections, and check memory use and performance after configuration changes.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Choosing parallelism settings for a model and GPU count
• Estimating GPU requirements for MoE models
• Adjusting parallelism for long input sequences

Guides typed data integration in NeMo Relay by choosing codecs, converters between application values and JSON. Helps preserve middleware edits and caller-visible responses while distinguishing application data conversion from provider-specific LLM request and response handling.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Connecting existing domain models to NeMo Relay
• Choosing codecs for Python or Node.js
• Making provider payloads inspectable by middleware

Export existing medical AI inference results to MLflow, a system for recording machine learning runs. Preview sanitized metadata, parameters, reproducibility details, and engineering quality metrics before export, with optional image artifacts under explicit upload policies.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Recording completed inference runs in MLflow
• Keeping generation parameters and reproducibility details
• Previewing export contents without contacting MLflow

Add support for new model architectures in NeMo AutoModel, from inspecting configuration files to implementing layers, converting saved weights, registering models, and testing numerical agreement with Hugging Face implementations.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Adding support for a new language model
• Mapping MoE router and expert weights
• Integrating a model that handles images and text

Package reusable NeMo Relay behavior as an embedded component or an independently distributed plugin. Design shared configuration, validate it before activation, and register behavior with rollback handling so failed setup does not leave partial changes active.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Share runtime behavior across Relay applications
• Validate plugin configuration before activation
• Package a native Rust plugin or a local gRPC worker

Configure CPU offloading in Megatron Bridge to reduce GPU memory use by moving intermediate training data or optimizer states to CPU memory. Check compatibility, adjust the offloaded fraction, and investigate memory errors or slowdowns.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reduce GPU memory use during model training
• Tune how much optimizer state moves to CPU memory
• Check compatibility with parallelism and CUDA graphs

Enable Megatron FSDP in Megatron-Bridge to distribute model training state across GPUs. Follow configuration and checkpoint requirements, investigate memory failures or performance regressions, and verify the setup with an existing two-GPU functional test.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Switch from DDP to Megatron FSDP
• Investigate memory failures after FSDP configuration changes
• Check checkpoint formats and incompatible settings

Write and review NeMo-RL documentation using its project conventions. Keep the documentation index aligned with added or renamed Markdown files, document new features in the appropriate location, and format function and class docstrings in Google style.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Add a new document and update the documentation index
• Update index entries after renaming a Markdown file
• Review function and class docstring formatting

Configure observability in NeMo Relay 0.6 or 0.7 to inspect recorded activity. Choose raw events, execution histories, or traces that show how calls unfold, then verify the selected output before adding more destinations.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Inspecting raw events during local debugging
• Exporting execution histories in ATIF format
• Configuring OpenTelemetry trace output

Configure Megatron-Bridge to overlap communication with computation during distributed training. Review settings for tensor, data, and pipeline parallelism, check activation requirements, and use included tests to investigate configuration issues or throughput regressions.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Enable communication overlap in distributed training
• Check why tensor parallel overlap is inactive
• Investigate slower training after configuration changes

Helps tune activation recomputation in Megatron Bridge, recalculating intermediate training values to reduce GPU memory use. Guides selective or full-layer settings, checks compatibility, and compares memory peaks with training speed to assess whether a configuration completes training steps.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Investigate training failures caused by activation memory
• Choose recomputation targets for attention or MoE models
• Compare GPU memory savings against training speed

Try NeMo Relay through a CLI wrapper, a supported integration, or direct Python, Node.js, or Rust instrumentation. Choose a minimal first run and verify recorded tool and model activity before considering one additional plugin.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Try Relay with Codex or Claude Code through the CLI
• Choose a supported integration for an existing application
• Check recorded tool and LLM calls in an application

Plan domain or language adaptation for Nemotron Speech recognition. Compare word boosting, language models, and fine-tuning based on available audio, accuracy targets, and budget, then route data preparation, training, evaluation, and deployment to the appropriate sub-skills.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Choose an approach for recognizing specialist terminology
• Plan adaptation for accents, noise, or another language
• Estimate audio data, GPU, time, and cost requirements

Tune Megatron Bridge training for MoE models, which route tokens to specialized experts, by overlapping expert communication with computation. Check dispatcher requirements, avoid conflicting settings, and compare steady training performance while changing one overlap setting at a time.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reduce exposed token dispatch and combine time
• Check requirements for expert communication overlap
• Configure flex dispatch and resolve overlap conflicts

Convert single-node training or inference commands into multi-node Slurm batch jobs. Choose an appropriate process launch method, configure shared storage and containers, and investigate communication timeouts, launch errors, and GPU memory shortages.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Converting a single-node command into a multi-node sbatch job
• Investigating NCCL communication timeouts
• Estimating GPU memory needs for MoE models

Guides hierarchical context parallelism in Megatron-Bridge, which splits input sequences across nested GPU groups. Covers configuration, implementation locations, constraints, and log checks to help enable the feature and investigate memory errors or performance regressions.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Configure hierarchical context parallelism
• Scale context parallelism beyond KV head counts
• Investigate memory errors after configuration changes

Configure and validate communication overlap in Megatron-Bridge for MoE models, which distribute work among expert networks. Check compatibility, compare training speed, and investigate regressions across standard dispatch, DeepEP, HybridEP, and delayed weight-gradient computation.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Enable expert communication overlap in MoE training
• Check compatibility before using DeepEP or HybridEP
• Compare plain overlap and delayed weight-gradient computation

Integrate NVIDIA NeMo Fabric into an application or evaluation system through its Python SDK. Map application settings into typed configuration, choose an execution lifecycle, validate prerequisites, and handle standardized results, artifacts, and telemetry.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Integrate agent execution into an application
• Build configurations for different evaluation conditions
• Manage single calls and conversations with retained state

Build, migrate, and review adapters that connect external agent runtimes to NVIDIA NeMo Fabric. It guides configuration translation, startup and shutdown, discovery metadata, and validation against the public contract, the rules defining how adapters interact with Fabric.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Building an adapter for an external agent runtime
• Migrating an adapter to the public contract
• Reviewing configuration mapping and runtime isolation

Diagnose GPU out-of-memory errors during Megatron Bridge training and choose memory reduction settings. Covers fragmentation, LoRA input storage, recomputation, parallelism changes, and CPU offloading, explaining compatibility limits and potential tradeoffs between memory use and training throughput.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Investigating GPU out-of-memory errors during training
• Reducing retained inputs in LoRA training
• Estimating memory needs before a large training job