Data analysis, machine learning, LLM apps, and prompt design (9.1K skills)
Find the right next step for an unclear NVFLARE request. Uses source or data inspection evidence to recommend a specific workflow skill for conversion, validation, deployment, or diagnosis, while keeping the project unchanged.
Investigate failures in NVIDIA TAO Visual ChangeNet classification experiments using prediction scores and actual images. Produces an illustrated root cause report covering data, reference images, training settings, and simulated changes to assess possible fixes.
Convert existing Hugging Face Trainer or TRL SFTTrainer code into an NVFLARE job for federated learning, where multiple sites train a shared model. Preserve training settings, validate locally, and export the resulting job.
Run video augmentation, which creates variations of source videos, and automatic labeling on OSMO. Select a workflow, check credentials and resources, monitor execution, download outputs, and compare augmented videos with their originals.
Integrate PhysicsNeMo domain parallelism, splitting one input across GPUs, into training or inference scripts. Guides DDP or FSDP2 integration, adds support for new operations, and checks distributed outputs and gradients against single-GPU results.
Convert existing PyTorch training code into an NVFLARE federated learning job, where sites train with their own data. Adapt training and evaluation, exchange model updates through the Client API, validate locally, and export the job.
Compute statistics across distributed NVFLARE sites without sharing raw data. Generate and validate jobs for numeric tables or image datasets, then report site-level and global summaries, missing data rates, and applied privacy settings.
Find PhysicsNeMo models, data loading pipelines, and examples for machine learning tasks involving physical systems. The skill checks current repository files, presents relevant alternatives, and suggests a reading order without writing training code.
Plan and troubleshoot Nemotron retrieval recipes for embeddings, which turn text into searchable vectors, and reranking, which reorders search results. Review configurations, prepare training and evaluation commands, and check export or deployment requirements.
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.
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.
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.
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.
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.
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.
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.
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.
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.
Plan and tune training for vision-language models that use MoE, a mixture of expert networks, in Megatron Bridge. Compares FSDP and 3D parallel training, with guidance on memory use, batch sizes, and performance checks using realistic image inputs.
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.
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.
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.
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.
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.