데이터·AI 스킬

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

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

Configure and validate CUDA graphs in Megatron Bridge training. Choose whole-iteration or module-level capture, which records GPU operations for replay, then check compatibility constraints and compare runtime and memory use against ordinary execution.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Choose whole-iteration or module-level CUDA graph capture
• Configure graph scopes for dense or MoE training
• Investigate crashes after CUDA graph configuration changes

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

Search documents through NeMo Retriever MCP and use retrieved passages to support answers. Add missing documents when server access is confirmed, then search them while preserving available document identifiers and page numbers for citations.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Finding evidence in indexed documents
• Adding missing documents and searching their contents
• Preparing answers with document identifiers and page references

Deploy, run, and test NVIDIA speech recognition, speech synthesis, and translation services in the cloud or on your own infrastructure. Guidance covers model selection, environment checks, custom models, and adjustments to speech processing and pronunciation.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Deploy and test a speech recognition model
• Generate speech with Magpie
• Adjust pronunciation for specific words

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

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

Guide repeated NeMo-RL or NeMo-gym experiments to test hypotheses and improve training settings. Establish a baseline, compare relevant metrics, and preserve reproducible records in git and TSV logs while keeping users involved in launch decisions.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Testing hypotheses about training settings
• Comparing accuracy, reward, or throughput
• Recording reproducible experiment histories

Create and modify NeMo AutoModel training and evaluation recipes. Get guidance on YAML configuration, functions that build training components, execution flow, validation, and checkpoint saving and resuming, with minimal examples and local checks.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Create a new fine-tuning recipe
• Add or change YAML configuration fields
• Configure validation, checkpoint saving, and resuming

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

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

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

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

Configure and validate sequence packing, which combines training examples into token sequences, and long-context training in Megatron-Bridge. Compare offline, in-batch, and Energon online approaches while checking batch sizes, parallelism requirements, and model compatibility.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Configure offline packing for text model fine-tuning
• Choose packing methods for vision-language models
• Check parallelism constraints for long-context training

Find starting configurations for training MoE models, which route work among specialized expert networks, on H100, B200, GB200, and GB300. Provides model-specific parallel layouts, communication options, and approximate performance bands to guide measured comparisons.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Choose initial MoE training settings for specific hardware
• Compare communication backends and parallel layouts
• Estimate approximate training throughput for planning

Search document collections with the NeMo Retriever CLI. Ingest files into a local search index or use a deployed service, then retrieve evidence for questions across PDFs, Office files, images, audio, and video.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Find evidence across internal documents
• Build a local search index from files
• Query an existing Retriever service

Run training with mock or real data in Megatron-LM and Megatron Bridge, match their settings, and compare loss curves, which show training error over time. Includes recipes and multi-GPU examples for checking training behavior.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Compare Megatron-LM and Bridge loss curves
• Translate MLM command-line arguments into Bridge settings
• Check training on multiple GPUs

Train and evaluate robot policies, the models that choose actions, using RSL-RL or RLinf in healthcare simulations. Export trained artifacts and check that they work through the normal Workflow runner.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Train an ultrasound probe reach policy
• Post-train a Trocar policy from a local GR00T checkpoint
• Evaluate and export an existing training checkpoint

Launch distributed Megatron-LM training on a SLURM cluster with guidance on job scripts, environment variables, GPU settings, and containers. Monitor jobs and investigate failures across worker processes to identify the initial cause.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Write an sbatch script for distributed training
• Configure training across multiple GPU nodes
• Check CUDA settings for the parallelism mode

Choose software for serving language or vision-language models on NVIDIA Jetson using a current memory audit. Get specific launch flags for memory limits, context length, and concurrency to address memory shortages or compare lower-memory alternatives.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Choosing serving software for a Jetson memory budget
• Adjusting vLLM flags after out-of-memory errors
• Considering a switch from vLLM to llama.cpp

Measure language model response times and generation speed on Jetson using wrappers for vLLM, llama.cpp, or Ollama. Produces structured JSON results to assess models, power modes, runtime settings, and performance changes after tuning.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Measure response times for a deployed model
• Compare performance before and after tuning
• Evaluate models and power modes under controlled conditions

Provides instructions for serving language and vision-language models on Jetson with vLLM or SGLang. Selects a runtime for the device and JetPack version, supplies launch commands, and explains how to verify the OpenAI-compatible API.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Starting a language-model API on Jetson
• Connecting Open WebUI or an app to a model server
• Serving a vision-language model with vLLM

Configure speculative decoding on Jetson vLLM servers, letting a smaller model propose tokens for the main model to verify. Adjust memory allocation and benchmark before and after to assess generation speed under low concurrency.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reducing slow token generation on Thor or AGX Orin
• Choosing EAGLE-3 or a small draft model
• Comparing generation speed before and after configuration