Translate Hugging Face blog post: LeRobot v0.6.0: Imagine, Evaluate, Improve

[Jwaminju]

Source: https://huggingface.co/blog/lerobot-release-v060

This PR adds a Korean translation draft for lerobot-release-v060.

Downstream handoff:

• SEO review should use the translation-flow manifest.
• Quality review should use the translation-flow manifest.

[github-actions]

PR Preview Action v1.8.1
🚀 View preview at https://hugging-face-krew.github.io/pr-preview/pr-166/
Built to branch gh-pages at 2026-07-08 09:30 UTC. Preview will be ready when the GitHub Pages deployment is complete.

[github-actions]

HF Agent Review

Gate	Result
Quality	❌ Fail
SEO	✅ Pass

Head SHA: 7bb1b486917dedb66b445159099c91e88416f888

Quality report — ❌ Fail

Quality Report

• Status: reject
• Quality Score: 0.0
• Hard failures: 97
• Issues: 364
• Source available: True
• Source changed: False
• Source segments: 122
• Target segments: 102

Scorecard

Dimension	Score
adequacy	0.0
technical_accuracy	0.0
completeness	0.0
terminology	0.0
fluency	0.0
publishing_integrity	0.0
style_locale	60.0

Metrics

• qe_metric: heuristic
• qe_average: 0.6495
• qe_min: 0.0
• embedding_similarity_average: 0.581
• embedding_similarity_min: 0.0161
• cache_hits: 0
• cache_misses: 204

MQM Judge

• Enabled: True
• Provider: openai
• Model: gpt-5.6-luna
• Reasoning effort: none
• Prompt: /home/runner/work/hugging-face-krew.github.io/hugging-face-krew.github.io/workflow/skills/quality/judges/mqm_prompt.md
• Prompt hash: 887d2931aa289213f0bdce4a917a0ac8364dad8e470011069b8f8bf758d69a90
• Style guide hash: 937d8cd893578d30e716a3eb513cdf5f10d6fd3ad8f5e77068b57f96e160de12
• Requested segments: 102
• Evaluated segments: 102
• MQM errors: 119
• Cache hits: 0
• Cache misses: 102
• Severity counts: {'critical': 94, 'major': 20, 'minor': 5}
• adequacy_average: 0.1117
• technical_average: 0.3227
• fluency_average: 0.7773

Style Guide

• Enabled: True
• Guide: /home/runner/work/hugging-face-krew.github.io/hugging-face-krew.github.io/workflow/skills/quality/style/hf-blog-ko-translation-guide.md
• Policy: /home/runner/work/hugging-face-krew.github.io/hugging-face-krew.github.io/workflow/skills/quality/configs/style_policy.yml
• Style score: 60.0
• Rule hits: {'alt_text_caption': 7, 'first_mention_bilingual': 2, 'link_text_translation': 67, 'list_consistency': 1, 'modal_strength': 9, 'translationese': 1}

Style Guide Findings

Rule	Severity	Segment	Current	Suggested
translationese	minor		에 의해	Rewrite the sentence in natural Korean.
list_consistency	minor		phrase, phrase, phrase, phrase, phrase, phrase, phrase, phrase, phrase, sentence, sentence, sentence, sentence, phrase, sentence, phrase, sentence, sentence, sentence, sentence, sentence, phrase, sentence, sentence, sentence	Use either sentence-style endings or phrase-style endings consistently within one list.
modal_strength	major	h_027	GR00T N1.7	Preserve the strength of up to using: 최대.
modal_strength	major	p_042	성공 탐지와 진행 추정은 로봇 학습 루프에서 빠진 조각이며, v0.6.0이 이를 지원합니다. LeRobot은 이제 정책 API를 모방하는 단일 인터페이스 뒤에 네 가지 보상 모델을 제공하는 통합 보상 모델 API(lerobot.rewards)를 갖추었다. 여기에 HIL-SERL 보상 분류기, SARM, 그리고 두 가지 새로운 추가 항목이 포함됩니다:	Preserve the strength of can using: 수 있습니다.
modal_strength	major	p_048	두 모델 모두 각 프레임의 진행 곡선을 데이터셋에 쓰는 라벨링 스크립트를 함께 제공하므로 보상 인식 행동 복제(RA-BC), 데이터셋 품질 검사, 진행 오버레이 동영상에 대비됩니다. Robometer와 TOPReward 문서를 확인하세요.	Preserve the strength of can using: 수 있습니다.
modal_strength	major	p_052	전체 세부 정보는 video encoding documentation에 있습니다.	Preserve the strength of can using: 수 있습니다.
modal_strength	major	h_079	FSDP: GPU보다 큰 모델 학습	Preserve the strength of up to using: 최대.
modal_strength	major	p_080	로봇 파운데이션 모델은 단일 GPU를 넘어 커지고 있다. LeRobot 학습은 이제 Accelerate를 통해 FSDP(완전 분할 데이터 병렬) 학습을 지원합니다: 파라미터, 그래디언트, 옵티마이저 상태가 GPU 간에 샤딩되고, 체크포인트는 다른 정책처럼 로드되는 단일 파일 model.safetensors로 다시 모아진다. 다른 GPU 수에서 FSDP 실행을 재개할 수도 있다. multi-GPU training docs를 참조하세요.	Preserve the strength of up to using: 최대.
modal_strength	major	l_088	--display_mode=foxglove은 원격 조작, 녹화, 롤아웃을 Foxglove로 스트리밍하며, 이 시각화 도구는 로봇 공학 세계의 많은 사람들에 의해 이미 사용됩니다. 원격 설정에서도 작동하며, lerobot-dataset-viz은 정리 가능한 데이터셋 재생을 제공합니다.	Preserve the strength of can using: 수 있습니다.
modal_strength	major	b_091	[!WARNING]	Preserve the strength of up to using: 최대.

Issues

QL-001 formatting / critical

• Message: TODO/FIXME/TBD or unresolved placeholder marker remains.
• Target: TODO
• Suggested fix: Remove unresolved markers before publishing.

QL-002 technical / critical

• Message: inline code mismatch.
• Source: lerobot-rollout
• Suggested fix: Preserve source inline code exactly.
• Reason: Hard gate exact-match validator failed: missing=['lerobot-rollout']

QL-003 formatting / critical

• Message: link target mismatch.
• Source: #benchmarks-one-cli-to-evaluate-them-all, #cloud-training-with-hf-jobs, #codebase-leaner-and-cleaner, #community--ecosystem, #datasets-faster-loading-richer-data, #depth-support-end-to-end, #eo-1, #evo1
• Target: #section-10, #section-11, #section-3, #section-4, #section-5, #section-6, #section-7, #section-8
• Suggested fix: Preserve source link target exactly.
• Reason: Hard gate exact-match validator failed: missing=['#benchmarks-one-cli-to-evaluate-them-all', '#cloud-training-with-hf-jobs', '#codebase-leaner-and-cleaner', '#community--ecosystem', '#datasets-faster-loading-richer-data', '#depth-support-end-to-end', '#eo-1', '#evo1']; extra=['#section-10', '#section-11', '#section-3', '#section-4', '#section-5', '#section-6', '#section-7', '#section-8']

QL-004 technical / major

• Message: model or dataset id mismatch.
• Source: CloudXR/OpenXR, SO-100/101, lerobot/Robometer-4B, lerobot/fastwam_base
• Target: lerobot/Robometer-
• Suggested fix: Preserve source model or dataset id exactly.
• Reason: Review gate exact-match validator failed: missing=['CloudXR/OpenXR', 'SO-100/101', 'lerobot/Robometer-4B', 'lerobot/fastwam_base']; extra=['lerobot/Robometer-']

QL-005 technical / major

• Message: number/unit token mismatch.
• Source: 2, 450, 5B
• Target: 4.5, 50, 7, 9
• Suggested fix: Preserve source number/unit token exactly.
• Reason: Review gate exact-match validator failed: missing=['2', '450', '5B']; extra=['4.5', '50', '7', '9']

QL-006 accuracy / major

• Message: Source segment coverage is low.
• Source: source_segments=122
• Target: target_segments=102
• Suggested fix: Check for omitted paragraphs, headings, list items, or table cells.
• Reason: Segment count validator found fewer target text segments than source text segments.

QL-007 accuracy / major

• Message: Duplicate target segments detected.
• Target: vlas: 모델 모음이 계속 확장됩니다 | 데이터셋: 더 빠른 로딩, 더 풍부한 데이터 | 벤치마크: 모든 것을 평가하는 하나의 cli | 보상 모델: 로봇이 성공했는지 아는 방법
• Suggested fix: Remove repeated translated segments unless the source intentionally repeats them.
• Reason: Duplicate detector found repeated normalized target segments.

QL-008 accuracy / major

• Message: QE metric score is low.
• Target: 0.4875
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-009 accuracy / major

• Message: QE metric score is low.
• Target: 0.4875
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-010 accuracy / major

• Message: QE metric score is low.
• Target: 0.5461
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-011 accuracy / major

• Message: QE metric score is low.
• Target: 0.4875
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-012 accuracy / major

• Message: QE metric score is low.
• Target: 0.5439
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-013 accuracy / major

• Message: QE metric score is low.
• Target: 0.5458
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-014 accuracy / major

• Message: QE metric score is low.
• Target: 0.3557
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-015 accuracy / major

• Message: QE metric score is low.
• Target: 0.4446
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-016 accuracy / major

• Message: QE metric score is low.
• Target: 0.3708
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-017 accuracy / major

• Message: QE metric score is low.
• Target: 0.2683
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-018 accuracy / major

• Message: QE metric score is low.
• Target: 0.2570
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-019 accuracy / major

• Message: QE metric score is low.
• Target: 0.5489
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-020 accuracy / major

• Message: QE metric score is low.
• Target: 0.3965
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-021 accuracy / major

• Message: QE metric score is low.
• Target: 0.2468
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-022 accuracy / major

• Message: QE metric score is low.
• Target: 0.5393
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-023 accuracy / major

• Message: QE metric score is low.
• Target: 0.0000
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-024 accuracy / minor

• Message: Embedding similarity is an outlier.
• Target: 0.0384
• Suggested fix: Review whether the segment still corresponds to the source.
• Reason: Embedding similarity is below threshold 0.08.

QL-025 accuracy / major

• Message: QE metric score is low.
• Target: 0.3102
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-026 accuracy / major

• Message: QE metric score is low.
• Target: 0.4392
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-027 accuracy / major

• Message: QE metric score is low.
• Target: 0.5312
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-028 accuracy / major

• Message: QE metric score is low.
• Target: 0.2736
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-029 accuracy / major

• Message: QE metric score is low.
• Target: 0.3466
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-030 accuracy / minor

• Message: Embedding similarity is an outlier.
• Target: 0.0536
• Suggested fix: Review whether the segment still corresponds to the source.
• Reason: Embedding similarity is below threshold 0.08.

QL-031 accuracy / major

• Message: QE metric score is low.
• Target: 0.1995
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-032 accuracy / major

• Message: QE metric score is low.
• Target: 0.2789
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-033 accuracy / minor

• Message: Embedding similarity is an outlier.
• Target: 0.0161
• Suggested fix: Review whether the segment still corresponds to the source.
• Reason: Embedding similarity is below threshold 0.08.

QL-034 accuracy / major

• Message: QE metric score is low.
• Target: 0.2852
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-035 accuracy / major

• Message: QE metric score is low.
• Target: 0.4048
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-036 accuracy / major

• Message: QE metric score is low.
• Target: 0.2657
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-037 accuracy / minor

• Message: Embedding similarity is an outlier.
• Target: 0.0643
• Suggested fix: Review whether the segment still corresponds to the source.
• Reason: Embedding similarity is below threshold 0.08.

QL-038 accuracy / major

• Message: QE metric score is low.
• Target: 0.5428
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-039 accuracy / major

• Message: QE metric score is low.
• Target: 0.3423
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-040 accuracy / minor

• Message: Embedding similarity is an outlier.
• Target: 0.0513
• Suggested fix: Review whether the segment still corresponds to the source.
• Reason: Embedding similarity is below threshold 0.08.

QL-041 accuracy / major

• Message: QE metric score is low.
• Target: 0.0000
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-042 accuracy / minor

• Message: Embedding similarity is an outlier.
• Target: 0.0247
• Suggested fix: Review whether the segment still corresponds to the source.
• Reason: Embedding similarity is below threshold 0.08.

QL-043 accuracy / major

• Message: QE metric score is low.
• Target: 0.5057
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-044 accuracy / major

• Message: QE metric score is low.
• Target: 0.4311
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-045 accuracy / major

• Message: QE metric score is low.
• Target: 0.5310
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-046 accuracy / major

• Message: QE metric score is low.
• Target: 0.3965
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-047 accuracy / major

• Message: QE metric score is low.
• Target: 0.4982
• Suggested fix: Review this segment for omission, unrelated translation, or over-compression.
• Reason: QE score is below threshold 0.55.

QL-048 fluency / minor

• Message: MQM judge reported fluency issue.
• Source: This new release is about closing the robot learning loop: policies that imagine the future before acting, reward models that tell you when your robot succeeds, a deployment CLI that turns failures into training data, and six new simulation benchmarks to measure it all.
• Target: 이 새 릴리스의 핵심은 로봇 학습 루프를 닫는 것입니다: 행동하기 전에 미래를 상상하는 정책, 로봇이 성공했는지 알려주는 보상 모델, 실패를 학습 데이터로 바꿔주는 배포 CLI, 그리고 이를 모두 측정하기 위한 여섯 개의 새로운 시뮬레이션 벤치마크.
• Suggested fix: 이번 릴리스에서는 로봇 학습 루프를 완성하는 기능을 소개합니다. 행동하기 전에 미래를 예측하는 정책, 로봇의 성공 여부를 알려주는 보상 모델, 실패를 학습 데이터로 전환하는 배포 CLI, 그리고 이를 종합적으로 측정할 수 있는 6개의 새로운 시뮬레이션 벤치마크를 제공합니다.
• Reason: 의미는 전달되지만 콜론 뒤의 긴 나열과 ‘루프를 닫는 것입니다’가 영어 구조를 강하게 따라 기술 블로그 문장으로 다소 딱딱합니다.

QL-049 accuracy / major

• Message: MQM judge reported accuracy issue.
• Source: world model policies (VLA-JEPA, FastWAM, LingBot-VA) that learn to imagine the future, a wave of new VLAs (GR00T N1.7, MolmoAct2, EO-1, EVO1, Multitask DiT)
• Target: 미래를 상상하는 월드 모델 정책(VLA-JEPA, FastWAM, LingBot-VA)을 도입하고, 미래를 상상하는 VLAs의 물결(GR00T N1.7, MolmoAct2, EO-1, EVO1, Multitask DiT)
• Suggested fix: 미래를 상상하도록 학습하는 월드 모델 정책(VLA-JEPA, FastWAM, LingBot-VA)을 도입하고, 새로운 VLA(GR00T N1.7, MolmoAct2, EO-1, EVO1, Multitask DiT)가 대거 추가되었으며
• Reason: ‘that learn to imagine the future’는 월드 모델 정책을 수식하지만, 목표 문장에서는 ‘미래를 상상하는’이 VLAs에도 반복되어 원문에 없는 의미를 추가하고 두 기술의 설명 범위를 혼동하게 만듭니다.

QL-050 fluency / minor

• Message: MQM judge reported fluency issue.
• Source: all on top of a leaner installation
• Target: 더 간소화된 설치 위에 구축됩니다
• Suggested fix: 이 모든 기능은 더 간소화된 설치 환경에서 제공됩니다
• Reason: ‘설치 위에 구축됩니다’는 영어 구조를 직역한 표현으로 기술 블로그 한국어에서 부자연스럽습니다.

QL-051 accuracy / critical

• Message: MQM judge reported accuracy issue.
• Source: LeRobot v0.6.0: Imagine, Evaluate, Improve
• Target: 월드 모델: 미래를 상상하는 정책들
• Suggested fix: LeRobot v0.6.0: 상상하고, 평가하고, 개선하기
• Reason: 원문의 제품명과 버전인 LeRobot v0.6.0이 완전히 누락되었고, Imagine, Evaluate, Improve라는 제목이 ‘월드 모델: 미래를 상상하는 정책들’로 바뀌어 원문에 없는 ‘월드 모델’과 ‘정책들’이라는 내용을 추가했습니다. 해당 목록 항목이 다른 주제의 제목으로 대체되어 독자가 릴리스와 버전 정보를 확인할 수 없습니다.

QL-052 technical / major

• Message: MQM judge technical score is low.
• Target: 0.1000
• Suggested fix: Review this segment against the MQM judge feedback.
• Reason: MQM technical score is below threshold 0.75.

QL-053 accuracy / major

• Message: MQM judge reported accuracy issue.
• Source: TL;DR
• Target: VLAs: 모델 모음이 계속 확장됩니다
• Suggested fix: TL;DR
• Reason: 원문의 ‘TL;DR’(핵심 요약) 표기가 ‘VLAs’로 임의 변경되었고, 원문에 없는 ‘모델 모음이 계속 확장됩니다’라는 내용이 추가되었습니다. 독자가 해당 항목을 요약 섹션으로 인식하지 못할 수 있습니다.

QL-054 accuracy / critical

• Message: MQM judge reported accuracy issue.
• Source: Table of contents
• Target: 보상 모델: 로봇이 성공했는지 아는 방법
• Suggested fix: 목차
• Reason: 원문의 목차 항목이라는 의미가 완전히 누락되고, 존재하지 않는 ‘보상 모델: 로봇이 성공했는지 아는 방법’이라는 기술적 제목이 추가되었습니다. 문서 구조와 내용이 모두 왜곡됩니다.

QL-055 accuracy / critical

• Message: MQM judge reported accuracy issue.
• Source: World models: policies that imagine
• Target: 데이터셋: 더 빠른 로딩, 더 풍부한 데이터
• Suggested fix: 월드 모델: 상상하는 정책
• Reason: 원문의 ‘World models’와 ‘policies that imagine’가 모두 ‘데이터셋’ 및 로딩·데이터 관련 내용으로 바뀌어 핵심 주제와 설명이 완전히 다른 의미가 되었습니다.

QL-056 technical / major

• Message: MQM judge technical score is low.
• Target: 0.1000
• Suggested fix: Review this segment against the MQM judge feedback.
• Reason: MQM technical score is below threshold 0.75.

QL-057 accuracy / critical

• Message: MQM judge reported accuracy issue.
• Source: VLA-JEPA
• Target: 벤치마크: 모든 것을 평가하는 하나의 CLI
• Suggested fix: VLA-JEPA
• Reason: 원문의 모델명 또는 프로젝트명인 “VLA-JEPA”가 완전히 누락되고, 원문에 없는 “벤치마크: 모든 것을 평가하는 하나의 CLI”라는 내용으로 대체되었습니다. 이는 핵심 식별자와 의미를 모두 훼손해 독자가 다른 항목으로 오인할 수 있습니다.

QL-058 technical / major

• Message: MQM judge technical score is low.
• Target: 0.0000
• Suggested fix: Review this segment against the MQM judge feedback.
• Reason: MQM technical score is below threshold 0.75.

QL-059 technical / critical

• Message: MQM judge reported technical issue.
• Source: LingBot-VA
• Target: 훈련 및 추론
• Suggested fix: LingBot-VA
• Reason: 원문의 모델명 또는 프로젝트명이 완전히 다른 일반 명사구인 '훈련 및 추론'으로 대체되어 핵심 식별 정보가 누락되고 의미가 변경되었습니다.

QL-060 accuracy / major

• Message: MQM judge adequacy score is low.
• Target: 0.1000
• Suggested fix: Review this segment against the MQM judge feedback.
• Reason: MQM adequacy score is below threshold 0.75.

QL-061 accuracy / critical

• Message: MQM judge reported accuracy issue.
• Source: FastWAM
• Target: 코드베이스: 더 간결하고 깔끔하게
• Suggested fix: FastWAM
• Reason: 원문의 고유명사 또는 프로젝트명인 "FastWAM"이 완전히 누락되고, 원문에 없는 "코드베이스"와 평가적 설명인 "더 간결하고 깔끔하게"가 추가되었습니다. 독자가 FastWAM이라는 대상을 식별할 수 없고 핵심 정보가 다른 내용으로 대체되었습니다.

QL-062 technical / major

• Message: MQM judge technical score is low.
• Target: 0.1000
• Suggested fix: Review this segment against the MQM judge feedback.
• Reason: MQM technical score is below threshold 0.75.

QL-063 accuracy / major

• Message: MQM judge reported accuracy issue.
• Source: VLAs: the model zoo keeps growing
• Target: 커뮤니티 & 생태계
• Suggested fix: VLAs: 모델 동물원이 계속 확장되고 있습니다
• Reason: 원문의 핵심인 VLAs와 모델 동물원(model zoo)이 계속 확장된다는 내용을 번역하지 않고, 의미가 다른 ‘커뮤니티 & 생태계’로 대체했습니다. 독자가 해당 목록 항목의 주제와 내용을 잘못 이해할 수 있습니다.

QL-064 technical / major

• Message: MQM judge technical score is low.
• Target: 0.2000
• Suggested fix: Review this segment against the MQM judge feedback.
• Reason: MQM technical score is below threshold 0.75.

QL-065 accuracy / critical

• Message: MQM judge reported accuracy issue.
• Source: GR00T N1.7
• Target: Final thoughts
• Suggested fix: GR00T N1.7
• Reason: 고유한 모델명 또는 항목명인 “GR00T N1.7”이 전혀 다른 일반 문구인 “Final thoughts”로 대체되어 원문의 핵심 정보와 항목 의미가 완전히 누락되었습니다.

QL-066 accuracy / critical

• Message: MQM judge reported accuracy issue.
• Source: MolmoAct2
• Target: 월드 모델: 미래를 상상하는 정책들
• Suggested fix: MolmoAct2
• Reason: 원문의 모델명 ‘MolmoAct2’가 완전히 누락되고, 원문에 없는 ‘월드 모델: 미래를 상상하는 정책들’이라는 제목으로 대체되었습니다. 모델 식별이 불가능해지는 게시 차단 수준의 오류입니다.

QL-067 technical / major

• Message: MQM judge technical score is low.
• Target: 0.1000
• Suggested fix: Review this segment against the MQM judge feedback.
• Reason: MQM technical score is below threshold 0.75.

QL-068 accuracy / critical

• Message: MQM judge reported accuracy issue.
• Source: EO-1
• Target: 로봇 공학 세계는 큰 질문에 직면해 있다: 월드 모델이 실제로 로봇 정책에 도움이 되는가? v0.6.0은 이 질문에 답하기 위해 LeRobot에 세 가지 정책을 도입합니다. 각 정책은 학습의 일부로 미래를 상상하는 법을 배우고, 각각은 그 상상을 합리적으로 유지하기 위해 서로 다른 경로를 택합니다.
• Suggested fix: 원문 'EO-1'을 그대로 유지하고, 해당 항목에 대응하는 실제 번역문을 제공하세요.
• Reason: 원문 리스트 항목 'EO-1'이 번역문에서 누락되고, 제공된 번역문은 원문에 없는 별도 문단으로 대체되었습니다. 필수 콘텐츠와 구조가 보존되지 않아 게시를 진행할 수 없습니다.

QL-069 technical / major

• Message: MQM judge technical score is low.
• Target: 0.0500
• Suggested fix: Review this segment against the MQM judge feedback.
• Reason: MQM technical score is below threshold 0.75.

QL-070 fluency / major

• Message: MQM judge fluency score is low.
• Target: 0.0500
• Suggested fix: Review this segment against the MQM judge feedback.
• Reason: MQM fluency score is below threshold 0.75.

QL-071 accuracy / critical

• Message: MQM judge reported accuracy issue.
• Source: Multitask DiT
• Target: VLA-JEPA
• Suggested fix: Multitask DiT
• Reason: 원문의 기술 항목명인 Multitask DiT가 전혀 다른 항목명인 VLA-JEPA로 바뀌어 대상과 의미를 식별할 수 없습니다.

QL-072 technical / major

• Message: MQM judge technical score is low.
• Target: 0.0000
• Suggested fix: Review this segment against the MQM judge feedback.
• Reason: MQM technical score is below threshold 0.75.

QL-073 accuracy / critical

• Message: MQM judge reported accuracy issue.
• Source: EVO1
• Target: VLA-JEPA는 잠재 공간에서 미래를 예측하도록 컴팩트한 VLA(Qwen3-VL-2B 기반)를 학습하는 동안, 행동하는 방법을 배우게 합니다: 학습 중에 JEPA 월드 모델은 모델의 자체 행동으로 다가오는 프레임을 예측해야 합니다는 점입니다. 비밀은 추론 시 월드 모델이 사라져 추가 추론 비용 없이 월드-모델 감독을 얻을 수 있다는 것입니다. 허브에 준비된 세 가지 바로 사용 가능한 체크포인트가 있으며, 미세 조정을 위한 DROID 사전학습 기본 모델도 포함되어 있다:
• Suggested fix: EVO1
• Reason: 원문 세그먼트가 제목 또는 목록 항목인 단일 식별자 EVO1인데, 대상은 전혀 다른 VLA-JEPA 설명 문단으로 대체되었습니다. 원문의 필수 콘텐츠가 보존되지 않았고, 대상의 기술 정보는 이 세그먼트에 근거 없이 추가되었습니다.

QL-074 technical / major

• Message: MQM judge technical score is low.
• Target: 0.3500
• Suggested fix: Review this segment against the MQM judge feedback.
• Reason: MQM technical score is below threshold 0.75.

QL-075 fluency / major

• Message: MQM judge fluency score is low.
• Target: 0.1800
• Suggested fix: Review this segment against the MQM judge feedback.
• Reason: MQM fluency score is below threshold 0.75.

QL-076 accuracy / critical

• Message: MQM judge reported accuracy issue.
• Source: `Rewar

SEO report — ✅ Pass

SEO Eval Report

Gate: ✅ PASS — deterministic AND rubric

• File: ../target/_posts/2026-07-07-lerobot-release-v060.md
• Source: —
• Primary keyword: (none — D5 skipped)
• Mode: file

Gate

• Status: PASS
• Blockers: ✅ pass
• Deterministic REQUIRED (D1–D7): ✅ pass
• Rubric (R1–R6): ✅ pass (mean None, min None)

Blockers

✅ body_not_empty: Body is not empty
✅ robots_indexable: Robots is indexable
✅ internal_links_resolve: All internal links resolve
✅ local_images_resolve: All local images resolve

Required checks (gated)

✅ heading_hierarchy: Heading hierarchy: Valid
✅ alt_text_coverage: Alt text coverage: 7/7 images
✅ descriptive_alt_text: Descriptive alt text: 7/7 (≥80% recommend)
✅ image_files_exist: All 0 local image file(s) exist

OpenAI rubric checks

✅ semantic_metadata: PASS (required) — 타이틀, H1, 주요 헤딩 및 오프닝의 의미가 서로 일치함
✅ alt_semantics: PASS (review) — —

Advisory checks (not gated)

✅ opening_summary: Opening 3 paragraphs: 385 chars (recommend ≥150 for KO/GEO)
✅ h1_count: Markdown H1 count: 1 (review against rendered layout)
✅ citations: Citations/statistics: 40 (recommend ≥1 for GEO)
⚠️ question_headings: Scannable H2/H3 (question or keyword): 0 (0 question, 0 keyword)
✅ internal_links: Internal links: 9 (recommend 2-3)
✅ word_count: Word count: 2124 (recommend ≥300)
ℹ️ primary_keyword: No primary_keyword in manifest — keyword check skipped
⚠️ webp_format: WebP format: 0/7 images (≥50% recommend)
⚠️ lazy_loading: Lazy loading: 0 images (optional)

Signals (evidence — not directly gated)

• Frontmatter: title 32 chars, description 0 chars, author present True
• Title text: LeRobot v0.6.0: 상상하고, 평가하고, 개선하기
• Description text: —
• Opening text: * TOC
  {:toc}

이 글은 Hugging Face 블로그의 LeRobot v0.6.0: Imagine, Evaluate, Improve를 한국어로 번역한 글입니다.

• Opening: first paragraph 159 chars, first 3 paragraphs 385 chars
• Headings: markdown H1 1, rendered effective H1 2, layout title H1 True
• Links: total 48, external 39, internal 0, citation signals 40
• Images: total 7, empty alt 0, filename-like alt 0, missing local files 0

Semantic review packet

• Title: LeRobot v0.6.0: 상상하고, 평가하고, 개선하기
• Description: —
• Rendered H1 candidates: LeRobot v0.6.0: 상상하고, 평가하고, 개선하기, LeRobot v0.6.0: 상상하고, 평가하고, 개선하기
• Opening: * TOC
  {:toc}

이 글은 Hugging Face 블로그의 LeRobot v0.6.0: Imagine, Evaluate, Improve를 한국어로 번역한 글입니다.

• Canonical/permalink: —
• Instruction: Compare title, description, rendered H1, and opening text for meaning consistency. This packet is evidence only; it does not decide pass/fail.

Frontmatter (advisory — written by metadata step, not gated)

✅ title: Title: 32 chars (recommend ≤60)
❌ description: Description is missing
✅ image: OG image: assets/images/blog/posts/2026-07-07-lerobot-release-v060/thumbnail.png
✅ categories: Categories: 2 (recommend 2-3)
✅ author: Author: dailybot