'nvidia' 검색 결과

TRL로 언어 모델을 추가 학습해 지시와 사람의 선호에 맞는 답변을 만들도록 안내합니다. 데이터 준비, 학습 방식 선택, 보상 모델 학습과 평가를 다룹니다.

• 답변 예시로 지시 수행을 학습합니다
• DPO로 답변 선호를 학습합니다
• 보상 모델을 쓰는 RLHF를 구성합니다

선호 답변과 비선호 답변 쌍으로 별도 참조 모델 없이 대규모 언어 모델을 추가 학습합니다. 모델별 설정과 학습 불안정, 메모리 부족에 대한 대응을 안내합니다.

• 답변 선호도 데이터로 모델을 학습합니다
• 지시 응답 모델을 추가 조정합니다
• 수학과 코드용 학습 설정을 조정합니다

Reduce the memory needed to run large language models with GPTQ quantization, which stores model weights at lower precision. Covers loading compressed models, quantizing your own models, choosing execution backends, and connecting to memory-efficient fine-tuning.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Run a language model within limited GPU memory
• Load a pre-quantized model for text generation
• Quantize and save your own trained model

Explore training workflows that reduce harmful model responses using a written set of principles. The skill covers self-critique, revised training examples, and RLAIF, reinforcement learning from AI feedback, with implementation examples and hardware guidance.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Designing principles for model safety training
• Creating training examples through self-critique and revision
• Planning reinforcement learning with AI-rated responses

Plan and configure large language model training with OpenRLHF to better match human preferences. Covers training method selection, distributed GPU setup, inference acceleration with vLLM, and troubleshooting memory or training stability issues.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Choosing a method for training models on human preferences
• Setting up PPO training with a reward model
• Configuring GRPO training without a critic model

Train language models with SimPO using pairs of preferred and rejected answers. Configure preference training without a separate reference model, adapt settings for base or instruction models, and address unstable training, memory limits, or lost capabilities.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Training a base model with preferred and rejected answers
• Further tuning an instruction model with preference data
• Adjusting training settings for math and code tasks

Guide language model training with TRL, from learning to follow instructions to matching human preferences. Covers method selection, training data, reward models that score responses, evaluation, and adjustments for memory limits or unstable training.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Train a model on instruction and response examples
• Align responses using preferred and rejected answer pairs
• Build a training pipeline using human feedback

Guides the use of RWKV, a language model architecture that carries information forward in a compact state. Covers streaming text generation, long document processing, fine-tuning, and memory comparisons with Transformers.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Build streaming text generation
• Process long documents in chunks
• Compare memory needs with Transformers

Deploy language models with vLLM through an OpenAI-compatible API or offline batch processing. Tune response speed, throughput, and GPU memory use with quantization, which reduces numerical precision, and execution across multiple GPUs.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Deploy an OpenAI-compatible language model API
• Generate text from many prompts in a batch
• Serve quantized models with limited GPU memory

Reduce GPU memory use by loading language models in 8-bit or 4-bit formats with bitsandbytes and HuggingFace Transformers. Includes configuration examples, QLoRA fine-tuning of compressed models, and optimizers that reduce memory needed during training.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Load larger models within limited GPU memory
• Compare 8-bit and 4-bit model settings
• Fine-tune a model with QLoRA

Build and evaluate Mamba models for long sequences using state-space models, which summarize past inputs in an internal state. Covers setup, pretrained text generation, differences between Mamba-1 and Mamba-2, and speed comparisons with Transformers.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Generate text with pretrained Mamba models
• Compare Mamba-1 and Mamba-2 configurations
• Benchmark generation speed against Transformers

Use LitGPT to explore language model implementations and train models on custom data. Covers pretrained model loading, LoRA fine-tuning that updates a small set of parameters, pretraining, and preparing trained models for deployment.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Studying language model architecture through readable code
• Fine-tuning a model on custom instruction data
• Configuring LoRA training for a single GPU

Prepare training datasets for large language models with GPU-assisted cleaning. Covers duplicate removal, quality filtering, and personal information redaction, with workflows for text, images, video, and audio using NeMo Curator.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Clean web data for model training
• Remove exact and near duplicates from large text collections
• Redact personal information from training text

Helps configure and troubleshoot DeepSpeed for training large models across multiple devices. Covers memory optimization, parallel training, and DeepNVMe storage transfers, with examples for tuning data movement and configuring models that divide work among specialized components.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Configuring distributed training for large models
• Tuning tensor transfers to NVMe storage
• Setting up models with specialized expert components

Transcribe audio files through 9Router using models from providers such as OpenAI, Groq, and Deepgram. Receive plain text, structured results with timestamps, or SRT and VTT subtitles, depending on the selected model’s support.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Transcribing recorded meetings or interviews
• Creating SRT or VTT subtitles from audio
• Getting transcriptions with timestamps

Build visual AI systems that detect objects or divide images into labeled regions. This skill guides dataset preparation, model selection, training, evaluation, inference optimization, and deployment using frameworks such as PyTorch and Ultralytics.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Build an object detection pipeline
• Clean images and convert annotation formats
• Optimize a trained model for deployment

Analyze the full retained history of one NemoClaw pull request to identify contributor, review, and automation delays. Generate a summary and an interactive timeline file, then compare observed timing with a default ten-minute target or a requested alternative.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Find delays across a pull request's lifetime
• Compare processing time with a target
• Measure review and merge waiting time

Implement an accepted NemoClaw issue or repair a classified pull request finding. Keep changes within the agreed scope, run focused checks for affected behavior, and review the final changes before handing evidence to an authorized publication workflow.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Implement an accepted issue within its agreed scope
• Repair a classified pull request finding
• Validate failure, recovery, and cleanup behavior

Audit and upgrade NemoClaw dependencies, including Hermes and base images. Trace upstream changes to affected code, implement required migrations, and verify tests, distributed artifacts, and the versions actually selected for use.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Assess the impact of a dependency upgrade
• Plan migrations for a Hermes upgrade
• Verify base images and selected artifacts

Find the right NemoClaw skill for a task using a catalog organized for users, maintainers, and contributors. Understand which skill handles documentation, setup, implementation, reviews, or releases, and choose an appropriate starting point.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Finding a skill that matches a NemoClaw task
• Browsing skills by user, maintainer, or contributor role
• Choosing skills for setup, implementation, and pull request publication

Recheck older NemoClaw bug reports against the reported and newest releases, then prepare evidence-backed recommendations for maintainers. It documents reproduction results and proposed updates for approval without automatically closing issues.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Rechecking older bug reports
• Comparing reported and newest releases
• Preparing evidence-backed issue updates

Validate the staging NemoClaw Brev Launchable by checking its web deployment flow, environment access, deployed image, and preinstalled user journey. Report inference results and missing dependencies separately as advisory evidence outside automated end-to-end testing.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Check the staging Launchable web deployment flow
• Verify an existing environment's image and source commit
• Test onboarding and inference in the preinstalled environment

Draft a NemoClaw release announcement from verified GitHub tags and shipped pull requests. Produce a local Markdown draft with an opening narrative, categorized changes, visible pull request links, and thanks to external contributors.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Drafting an announcement after a release tag is created
• Explaining shipped changes and their impact on users
• Preparing acknowledgments for external contributors

Reorganize NemoClaw documentation into focused pages with clearer navigation and a single home for each topic. Plan or implement structural changes while preserving useful information and existing published URLs, then validate routes, links, and builds.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Splitting long documentation pages into focused topics
• Consolidating duplicated troubleshooting and reference content
• Reorganizing navigation while preserving published URLs