https://devotd.net/en/news/huggingface-transformers-v5230-models-exporters/devotd.en2026-11-13T21:43:21ZTransformers v5.23.0: New model families, unified exporthttps://devotd.net/ja/news/huggingface-transformers-v5230-models-exporters/devotd.ja2026-11-13T21:43:21ZTransformers v5.23.0：新モデル群、統合エクスポートhttps://devotd.net/ko/news/huggingface-transformers-v5230-models-exporters/devotd.ko2026-11-13T21:43:21ZTransformers v5.23.0: 새 모델 계열, 통합 내보내기https://devotd.net/en/news/llm-post-training-weight-deletion-method/devotd.en2026-11-13T19:49:55ZDeletion Tuning: Gate weight deletion, smaller post-training adaptershttps://devotd.net/ja/news/llm-post-training-weight-deletion-method/devotd.ja2026-11-13T19:49:55ZDeletion Tuning：ゲート重みの削除、事後学習アダプターの縮小https://devotd.net/ko/news/llm-post-training-weight-deletion-method/devotd.ko2026-11-13T19:49:55ZDeletion Tuning: 게이트 가중치 삭제, 후학습 어댑터 축소https://devotd.net/en/news/nvidia-jetson-orin-cudla-python-inference/devotd.en2026-11-13T19:48:15Zcuda-bindings 13.3.0: Python DLA inference, GPU–DLA collaborationhttps://devotd.net/ja/news/nvidia-jetson-orin-cudla-python-inference/devotd.ja2026-11-13T19:48:15Zcuda-bindings 13.3.0：PythonでのDLA推論、GPU・DLAの連携https://devotd.net/ko/news/nvidia-jetson-orin-cudla-python-inference/devotd.ko2026-11-13T19:48:15Zcuda-bindings 13.3.0: 파이썬 DLA 추론, GPU·DLA 협업https://devotd.net/en/news/serving-glm-5-3-across-two-dgx-stations-vllm/devotd.en2026-11-13T17:52:50ZGLM-5.3: Distributed inference on two Stations, a single APIhttps://devotd.net/ja/news/serving-glm-5-3-across-two-dgx-stations-vllm/devotd.ja2026-11-13T17:52:50ZGLM-5.3：2台のStationで分散推論、単一APIhttps://devotd.net/ko/news/serving-glm-5-3-across-two-dgx-stations-vllm/devotd.ko2026-11-13T17:52:50ZGLM-5.3: 두 Station 분산 추론, 단일 APIhttps://devotd.net/en/news/github-ai-scan-repository-rest-api/devotd.en2026-11-13T17:45:26ZGitHub AI Scan: Organization settings, repository exceptionshttps://devotd.net/ja/news/github-ai-scan-repository-rest-api/devotd.ja2026-11-13T17:45:26ZGitHub AI Scan：組織別の設定、リポジトリ別の例外https://devotd.net/ko/news/github-ai-scan-repository-rest-api/devotd.ko2026-11-13T17:45:26ZGitHub AI Scan: 조직별 설정, 저장소별 예외https://devotd.net/en/news/llama-cpp-v1-release-runtime-changes/devotd.en2026-11-13T15:47:08Zllama.cpp v1.0.0: Compatibility rules, expanded server featureshttps://devotd.net/ja/news/llama-cpp-v1-release-runtime-changes/devotd.ja2026-11-13T15:47:08Zllama.cpp v1.0.0：互換性の基準、サーバー機能の拡張https://devotd.net/ko/news/llama-cpp-v1-release-runtime-changes/devotd.ko2026-11-13T15:47:08Zllama.cpp v1.0.0: 호환성 기준, 서버 기능 확장https://devotd.net/en/news/deepspeed-0200-training-inference-updates/devotd.en2026-11-13T11:42:21ZDeepSpeed v0.20.0: Deterministic execution, heterogeneous accelerator supporthttps://devotd.net/ja/news/deepspeed-0200-training-inference-updates/devotd.ja2026-11-13T11:42:21ZDeepSpeed v0.20.0：決定論的実行、異種アクセラレーター対応https://devotd.net/ko/news/deepspeed-0200-training-inference-updates/devotd.ko2026-11-13T11:42:21ZDeepSpeed v0.20.0: 결정론적 실행, 이종 가속기 지원https://devotd.net/en/news/mimo-code-v0117-provider-defaults-config/devotd.en2026-11-13T09:43:46ZMiMo-Code v0.1.17: Per-model settings, optional searchhttps://devotd.net/ja/news/mimo-code-v0117-provider-defaults-config/devotd.ja2026-11-13T09:43:46ZMiMo-Code v0.1.17：モデル別設定、選択式検索https://devotd.net/ko/news/mimo-code-v0117-provider-defaults-config/devotd.ko2026-11-13T09:43:46ZMiMo-Code v0.1.17: 모델별 설정, 선택형 검색https://devotd.net/en/news/edge0-v020-unified-inference-engine/devotd.en2026-11-13T09:40:55ZEdge0 v0.2.0: Unified engine, large models on one GPUhttps://devotd.net/ja/news/edge0-v020-unified-inference-engine/devotd.ja2026-11-13T09:40:55ZEdge0 v0.2.0：エンジン統合、単一GPUでの大規模モデル実行https://devotd.net/ko/news/edge0-v020-unified-inference-engine/devotd.ko2026-11-13T09:40:55ZEdge0 v0.2.0: 엔진 통합, 단일 GPU 대형 모델https://devotd.net/en/news/openbot-desktop-1-0-2-rc-3-podman-updates-local-data-safety/devotd.en2026-11-13T03:47:19ZOpenBot Desktop 1.0.2-rc.3: Isolated Podman runtime, local data protectionhttps://devotd.net/ja/news/openbot-desktop-1-0-2-rc-3-podman-updates-local-data-safety/devotd.ja2026-11-13T03:47:19ZOpenBot Desktop 1.0.2-rc.3：Podmanの分離実行、ローカルデータの保護https://devotd.net/ko/news/openbot-desktop-1-0-2-rc-3-podman-updates-local-data-safety/devotd.ko2026-11-13T03:47:19ZOpenBot Desktop 1.0.2-rc.3: Podman 분리 실행, 로컬 데이터 보호https://devotd.net/en/news/gemini-4-hydrogen-agentic-workflows/devotd.en2026-11-13T01:46:09ZGemini 4 Hydrogen: Fast execution, model routinghttps://devotd.net/ja/news/gemini-4-hydrogen-agentic-workflows/devotd.ja2026-11-13T01:46:09ZGemini 4 Hydrogen：高速実行、モデルルーティングhttps://devotd.net/ko/news/gemini-4-hydrogen-agentic-workflows/devotd.ko2026-11-13T01:46:09ZGemini 4 Hydrogen: 빠른 실행, 모델 라우팅https://devotd.net/en/news/executorch-1-6-0-runtime-backends-llm/devotd.en2026-11-12T23:42:00ZExecuTorch 1.6.0: Torch-free runtime, batched LLM servinghttps://devotd.net/ja/news/executorch-1-6-0-runtime-backends-llm/devotd.ja2026-11-12T23:42:00ZExecuTorch 1.6.0：PyTorch不要のランタイム、バッチLLMサービングhttps://devotd.net/ko/news/executorch-1-6-0-runtime-backends-llm/devotd.ko2026-11-12T23:42:00ZExecuTorch 1.6.0: 파이토치 없는 런타임, 묶음 LLM 서빙https://devotd.net/en/news/ltx-2-1-6-2-out-of-clip-guides/devotd.en2026-11-12T23:34:17ZLTX-2 1.6.2: Out-of-clip image guides, per-chunk conditioninghttps://devotd.net/ja/news/ltx-2-1-6-2-out-of-clip-guides/devotd.ja2026-11-12T23:34:17ZLTX-2 1.6.2：クリップ外の画像ガイド、チャンク別の条件指定https://devotd.net/ko/news/ltx-2-1-6-2-out-of-clip-guides/devotd.ko2026-11-12T23:34:17ZLTX-2 1.6.2: 클립 밖 이미지 가이드, 조각별 조건 지정https://devotd.net/en/news/unsloth-v0-1-917-beta-qwen-gguf-training-inference/devotd.en2026-11-12T15:45:27ZUnsloth v0.1.917-beta: Qwen3.8-27B GGUF, T4 image and generation improvementshttps://devotd.net/ja/news/unsloth-v0-1-917-beta-qwen-gguf-training-inference/devotd.ja2026-11-12T15:45:27ZUnsloth v0.1.917-beta：Qwen3.8-27B GGUF、T4の画像・生成を改善https://devotd.net/ko/news/unsloth-v0-1-917-beta-qwen-gguf-training-inference/devotd.ko2026-11-12T15:45:27ZUnsloth v0.1.917-beta: Qwen3.8-27B GGUF, T4 이미지·생성 개선https://devotd.net/en/news/nvidia-openshell-v0-1-8-ai-agent-sandbox-security/devotd.en2026-11-12T15:41:34ZNVIDIA OpenShell v0.1.8: Credential placement, WebSocket policieshttps://devotd.net/ja/news/nvidia-openshell-v0-1-8-ai-agent-sandbox-security/devotd.ja2026-11-12T15:41:34ZNVIDIA OpenShell v0.1.8：認証情報の配置指定、WebSocketポリシーhttps://devotd.net/ko/news/nvidia-openshell-v0-1-8-ai-agent-sandbox-security/devotd.ko2026-11-12T15:41:34ZNVIDIA OpenShell v0.1.8: 자격 증명 위치 지정, 웹소켓 정책https://devotd.net/en/news/codex-0-167-2-vertex-ai-incremental-tool-declarations/devotd.en2026-11-12T13:41:25ZCodex 0.167.2: Incremental tool updates enabled by default on Vertex AI, independent provider settingshttps://devotd.net/ja/news/codex-0-167-2-vertex-ai-incremental-tool-declarations/devotd.ja2026-11-12T13:41:25ZCodex 0.167.2：Vertex AIのツール一覧の増分更新を標準で有効化、プロバイダー別の独立設定https://devotd.net/ko/news/codex-0-167-2-vertex-ai-incremental-tool-declarations/devotd.ko2026-11-12T13:41:25ZCodex 0.167.2: Vertex AI 도구 목록 증분 갱신 기본 적용, 공급자별 독립 설정https://devotd.net/en/news/diffusers-0-43-modular-pipelines/devotd.en2026-11-12T11:40:33ZDiffusers 0.43.0: Modular integration, inference improvementshttps://devotd.net/ja/news/diffusers-0-43-modular-pipelines/devotd.ja2026-11-12T11:40:33ZDiffusers 0.43.0：モジュラー統合、推論の改善https://devotd.net/ko/news/diffusers-0-43-modular-pipelines/devotd.ko2026-11-12T11:40:33ZDiffusers 0.43.0: 모듈러 통합, 추론 개선https://devotd.net/en/news/google-labs-tabletop-collaborative-ai-planning/devotd.en2026-11-12T09:31:40ZGoogle Labs Tabletop: Prompt-based collaboration, group decisionshttps://devotd.net/ja/news/google-labs-tabletop-collaborative-ai-planning/devotd.ja2026-11-12T09:31:40ZGoogle Labs Tabletop：プロンプトの共同作業スペース、グループの意思決定https://devotd.net/ko/news/google-labs-tabletop-collaborative-ai-planning/devotd.ko2026-11-12T09:31:40ZGoogle Labs Tabletop: 프롬프트 협업 공간, 그룹 의사결정https://devotd.net/en/news/transformerlens-4-4-whisper-ssm-bridge/devotd.en2026-11-12T07:31:48ZTransformerLens 4.4.0: Whisper bridge, SSM and hybrid interpretationhttps://devotd.net/ja/news/transformerlens-4-4-whisper-ssm-bridge/devotd.ja2026-11-12T07:31:48ZTransformerLens 4.4.0：Whisperブリッジ、SSM・ハイブリッド解析https://devotd.net/ko/news/transformerlens-4-4-whisper-ssm-bridge/devotd.ko2026-11-12T07:31:48ZTransformerLens 4.4.0: Whisper 브리지, SSM·하이브리드 해석https://devotd.net/en/news/nvidia-hyperflare-inference-v0-1-0-multihost-prefix-cache/devotd.en2026-11-12T05:37:26ZNVIDIA Hyperflare Inference v0.1.0: Multi-host cache sharing, speculative decodinghttps://devotd.net/ja/news/nvidia-hyperflare-inference-v0-1-0-multihost-prefix-cache/devotd.ja2026-11-12T05:37:26ZNVIDIA Hyperflare Inference v0.1.0：マルチホストのキャッシュ共有、投機的デコーディングhttps://devotd.net/ko/news/nvidia-hyperflare-inference-v0-1-0-multihost-prefix-cache/devotd.ko2026-11-12T05:37:26ZNVIDIA Hyperflare Inference v0.1.0: 다중 호스트 캐시 공유, speculative decodinghttps://devotd.net/en/news/nvidia-dsx-runtime-ai-inventory/devotd.en2026-11-12T02:09:40ZNVIDIA DSX AI Factory AICR v1.2.0: Reproducible GPU configurations, AI asset trackinghttps://devotd.net/ja/news/nvidia-dsx-runtime-ai-inventory/devotd.ja2026-11-12T02:09:40ZNVIDIA DSX AI FactoryのAICR v1.2.0：GPU構成の再現、AI資産の追跡https://devotd.net/ko/news/nvidia-dsx-runtime-ai-inventory/devotd.ko2026-11-12T02:09:40ZNVIDIA DSX AI Factory의 AICR v1.2.0: GPU 구성 재현, AI 자산 추적https://devotd.net/en/news/agentscope-2-harness/devotd.en2026-11-12T02:05:48ZAgentScope 2.0: Harness as the default entry point, separate orchestration and sandbox executionhttps://devotd.net/ja/news/agentscope-2-harness/devotd.ja2026-11-12T02:05:48ZAgentScope 2.0：ハーネスを標準の入口に、サーバーの調整とサンドボックスの分離https://devotd.net/ko/news/agentscope-2-harness/devotd.ko2026-11-12T02:05:48ZAgentScope 2.0: 하네스 기본 진입점, 서버 조율과 샌드박스 분리https://devotd.net/en/news/zyphra-zonos3-mini-tts-webgpu-runtimes/devotd.en2026-11-11T23:39:08ZZONOS3-Mini: Real-time voice cloning, multiple runtime supporthttps://devotd.net/ja/news/zyphra-zonos3-mini-tts-webgpu-runtimes/devotd.ja2026-11-11T23:39:08ZZONOS3-Mini：リアルタイム音声クローン、複数ランタイム対応https://devotd.net/ko/news/zyphra-zonos3-mini-tts-webgpu-runtimes/devotd.ko2026-11-11T23:39:08ZZONOS3-Mini: 실시간 음성 복제, 다중 런타임 지원