Ultralytics YOLO26
개요#
Ultralytics YOLO26은 Ultralytics YOLO26 논문에 소개된 통합 실시간 비전 모델 제품군입니다. 네이티브 종단 간 추론, 경량 감지 헤드, 업데이트된 학습 레시피, 감지·분할·분류·포즈 추정 및 방향성 감지를 위한 작업별 헤드를 도입했습니다.
YOLO26은 다섯 가지 탐지 스케일에서 COCO 기준 40.9-57.5 mAP와 T4 TensorRT 지연 시간 1.7-11.8 ms를 달성합니다. 논문은 또한 Intel Xeon CPU @ 2.00 GHz에서 YOLO26n이 YOLO11n보다 CPU ONNX 추론이 최대 43% 빠르다고 보고합니다.
계획된 아키텍처, 작업 및 예비 벤치마크는 미출시 YOLO27 미리 보기를 참조하세요.

from ultralytics import YOLO
model = YOLO("yolo26n.pt") # 사전 학습된 YOLO26n 모델 로드
results = model("path/to/bus.jpg") # 추론 실행Ultralytics Platform에서 YOLO26 모델을 직접 탐색하고 실행하세요.
YOLO26 모델 제품군은 네 가지 설계 영역을 기반으로 합니다:
- 네이티브 종단 간 추론: 선택적 일대일 감지 헤드는 비최대 억제(NMS) 없이 예측을 생성하여 배포를 간소화하고 후처리를 줄입니다.
- 경량 박스 회귀: YOLO26은 Distribution Focal Loss(DFL)를 제거하여 제약 없는 회귀 범위를 유지하면서 탐지 헤드의 복잡도를 낮춥니다.
- 학습 레시피 업데이트: 학습 파이프라인은 MuSGD(Muon과 SGD를 결합한 하이브리드 옵티마이저), Progressive Loss, STAL(Small-Target-Aware Label Assignment)을 결합해 최적화를 개선하고, 추론 시점의 헤드로 감독을 이동하며, 작은 객체의 양성 레이블 범위를 유지합니다. 공개된 체크포인트에 사용된 전체 하이퍼파라미터는 YOLO26 Training Recipe 가이드에 문서화되어 있습니다.
- 작업별 헤드 및 손실: YOLO26은 작업 전반에서 단일 모델 파이프라인을 유지하면서 인스턴스 세그멘테이션, 시맨틱 세그멘테이션 변형, 포즈 추정 및 방향 탐지를 위한 맞춤형 설계를 추가합니다.
이러한 업데이트를 통해 모델 스케일 및 배포 대상 전반에서 정확도와 지연 시간 간의 균형이 개선됩니다.
주요 기능#
-
DFL 없는 회귀 YOLO26은 Distribution Focal Loss(DFL)를 제거하여 탐지 헤드의 복잡도를 낮추고 내보내기를 간소화합니다.
-
엔드투엔드 NMS-free 추론 YOLO26은
nms=False을 사용한 네이티브 엔드투엔드 추론을 지원하며, 별도의 NMS 단계 없이 예측을 생성합니다. 기본 일대다 경로는 정확도를 위해 NMS를 사용합니다. -
Progressive Loss + STAL Progressive Loss는 학습의 초점을 추론 시점의 헤드로 옮기며, STAL은 소형 객체의 양성 레이블 커버리지를 개선합니다.
-
MuSGD Optimizer 대규모 언어 모델 학습의 최적화 아이디어를 컴퓨터 비전에 적용한 하이브리드 optimizer로, SGD와 Muon을 결합합니다.
-
효율적인 배포 간소화된 헤드와 선택적 NMS-free 경로는 내보내기 대상과 하드웨어 프로파일 전반에서 추론 오버헤드를 줄입니다. 여기에는 논문에서 보고한 YOLO26n의 CPU ONNX 속도 향상(YOLO11n 대비)도 포함됩니다.
-
인스턴스 분할 개선 사항 모델 수렴을 개선하기 위해 시맨틱 분할 손실을 도입하고, 더 나은 마스크 품질을 위해 다중 스케일 정보를 사용하는 프로토 모듈을 업그레이드했습니다. 논문은 COCO 인스턴스 분할에서 YOLO11 대비 박스 AP 최대 +2.5, 마스크 AP 최대 +3.7의 향상을 보고합니다.
-
정밀한 포즈 추정 보다 정확한 키포인트 위치 추정을 위해 잔차 로그 가능도 추정(RLE)을 통합하고, 추론 속도를 높이도록 디코딩 프로세스를 최적화합니다. 논문은 COCO 포즈 추정에서 YOLO11 대비 AP가 최대 +7.2 향상되었다고 보고합니다.
-
개선된 OBB 디코딩 정사각형 객체의 탐지 정확도를 개선하는 전용 각도 손실을 도입하고, 경계 불연속 문제를 해결하도록 OBB 디코딩을 최적화합니다. 논문은 DOTA-v1.0 방향 탐지에서 YOLO11 대비 mAP가 최대 +3.4 향상되었다고 보고합니다.

지원 작업 및 모드#
YOLO26은 다섯 가지 모델 스케일에 걸쳐 표준 Ultralytics 작업 세트를 지원합니다:
| 모델 | 파일 이름 | 작업 | 학습 | 검증 | 추론 | 내보내기 |
|---|---|---|---|---|---|---|
| YOLO26 | yolo26n.pt yolo26s.pt yolo26m.pt yolo26l.pt yolo26x.pt | 감지 | ✅ | ✅ | ✅ | ✅ |
| YOLO26-seg | yolo26n-seg.pt yolo26s-seg.pt yolo26m-seg.pt yolo26l-seg.pt yolo26x-seg.pt | 인스턴스 분할 | ✅ | ✅ | ✅ | ✅ |
| YOLO26-sem | yolo26n-sem.pt yolo26s-sem.pt yolo26m-sem.pt yolo26l-sem.pt yolo26x-sem.pt | 시맨틱 분할 | ✅ | ✅ | ✅ | ✅ |
| YOLO26-depth | yolo26n-depth.pt yolo26s-depth.pt yolo26m-depth.pt yolo26l-depth.pt yolo26x-depth.pt | 깊이 추정 | ✅ | ✅ | ✅ | ✅ |
| YOLO26-cls | yolo26n-cls.pt yolo26s-cls.pt yolo26m-cls.pt yolo26l-cls.pt yolo26x-cls.pt | 분류 | ✅ | ✅ | ✅ | ✅ |
| YOLO26-pose | yolo26n-pose.pt yolo26s-pose.pt yolo26m-pose.pt yolo26l-pose.pt yolo26x-pose.pt | 포즈/키포인트 | ✅ | ✅ | ✅ | ✅ |
| YOLO26-obb | yolo26n-obb.pt yolo26s-obb.pt yolo26m-obb.pt yolo26l-obb.pt yolo26x-obb.pt | 방향성 객체 감지 | ✅ | ✅ | ✅ | ✅ |
이 통합 프레임워크는 학습, 검증, 추론 및 내보내기를 지원하며 실시간 탐지, 인스턴스 세그멘테이션, 시맨틱 세그멘테이션, 단안 깊이 추정, 분류, 포즈 추정 및 방향 객체 탐지를 처리합니다.
yolo26-p2.yaml 및 yolo26-p6.yaml은 P2(소형 객체) 또는 P6(대형 입력) 탐지 헤드를 추가하며 YAML 아키텍처로만 제공됩니다. 스케일별 yolo26*-p2.pt 또는 yolo26*-p6.pt 가중치는 출시되지 않았습니다. YAML에서 스케일이 적용된 구성을 생성한 다음(예: YOLO("yolo26n-p6.yaml")), 필요에 따라 학습하거나 fine-tuning하세요.
성능 지표#
COCO에서 학습되고 사전 학습된 클래스 80개를 포함하는 이 모델의 사용 예제는 탐지 문서를 참조하세요. COCO를 참조하세요.
| 모델 | 크기 (픽셀) | mAPval50-95 | mAP검증50-95 (e2e) | 속도 CPU ONNX (ms) | 속도 T4 TensorRT10 (ms) | 파라미터 (M) | FLOPs (B) |
|---|---|---|---|---|---|---|---|
| YOLO26n | 640 | 40.9 | 40.1 | 38.9 ± 0.7 | 1.7 ± 0.0 | 2.4 | 5.5 |
| YOLO26s | 640 | 48.6 | 47.8 | 87.2 ± 0.9 | 2.5 ± 0.0 | 9.5 | 20.9 |
| YOLO26m | 640 | 53.1 | 52.5 | 220.0 ± 1.4 | 4.7 ± 0.1 | 20.4 | 68.4 |
| YOLO26l | 640 | 55.0 | 54.4 | 286.2 ± 2.0 | 6.2 ± 0.2 | 24.8 | 86.8 |
| YOLO26x | 640 | 57.5 | 56.9 | 525.8 ± 4.0 | 11.8 ± 0.2 | 55.7 | 194.4 |
Params 및 FLOPs 값은 Conv/BatchNorm 폴딩을 적용하고 사용되지 않는 감지 분기를 제거한 후의 융합 모델 기준입니다. 속도 측정에서는 nms=False을 사용해 NMS 없는 헤드를 선택합니다. 사전 학습된 체크포인트에는 전체 학습 아키텍처가 유지되므로 수치가 더 높게 나타날 수 있습니다.
사용 예시#
이 섹션에서는 YOLO26 학습 및 추론의 간단한 예제를 제공합니다.
아래 예제에서는 COCO8 예제 데이터셋의 객체 탐지를 위해 YOLO26 Detect 모델을 학습합니다.
PyTorch 사전 학습된 *.pt 모델과 구성 *.yaml 파일을 YOLO() 클래스에 전달하여 Python에서 모델 인스턴스를 생성할 수 있습니다:
from ultralytics import YOLO
# COCO로 사전 학습된 YOLO26n 모델을 불러옵니다
model = YOLO("yolo26n.pt")
# 'bus.jpg' 이미지에서 YOLO26n 모델로 추론을 실행합니다
results = model("path/to/bus.jpg")
# COCO8 예제 데이터셋으로 100 에포크 동안 모델 학습
results = model.train(data="coco8.yaml", epochs=100, imgsz=640)YOLO26 detection 모델은 다양한 배포 시나리오에 유연하게 대응하는 듀얼 헤드 아키텍처를 사용합니다:
- One-to-Many Head(기본값): NMS 후처리가 필요한 기존 YOLO 출력을 생성하며,
(N, nc + 4, 8400)을 출력합니다. 여기서nc은 클래스 수입니다. 이 헤드는 일반적으로 약간 더 높은 정확도를 제공하지만 추가 처리가 필요합니다. - One-to-One Head(
nms=False): NMS 없이 엔드투엔드 예측을 수행하며, 이미지당 최대 300개의 detection이 포함된(N, 300, 6)을 출력합니다. 이 헤드는 빠른 추론과 간소화된 배포에 최적화되어 있습니다.
Export, prediction 또는 validation 중에 헤드를 전환할 수 있습니다:
from ultralytics import YOLO
model = YOLO("yolo26n.pt")
# one-to-many head 사용(기본값, Ultralytics가 NMS 적용)
results = model.predict("image.jpg") # inference
metrics = model.val(data="coco.yaml") # validation
model.export(format="onnx") # export
# NMS-free one-to-one head 사용
results = model.predict("image.jpg", nms=False) # inference
metrics = model.val(data="coco.yaml", nms=False) # validation
model.export(format="onnx", nms=False) # exportPrediction과 validation은 정확도를 위해 기본적으로 one-to-many head를 사용합니다. NMS-free 추론에는 nms=False을 사용하고, 지원되는 export에 NMS를 포함하려면 nms=True을 사용하세요. 선택에 관계없이 두 헤드 모두 학습됩니다. 출력 형식 및 export 호환성은 End-to-End Detection 가이드를 참조하세요.
YOLOE-26: 개방형 어휘 탐지 및 분할#
YOLO26은 YOLOE-26에도 사용됩니다. YOLOE-26은 학습 시 학습된 고정 클래스 목록 대신 텍스트 프롬프트, 시각적 프롬프트 또는 프롬프트 없는 모드를 사용해 객체 범주를 감지하고 분할하는 오픈 보캐뷸러리 변형입니다. YOLOE-26은 YOLO26의 선택적 NMS 없는 종단 간(e2e) 헤드를 유지하므로, 대상 범주가 시간에 따라 바뀌는 동적 환경에서도 오픈 보캐뷸러리 추론을 충분히 빠르게 수행할 수 있습니다. YOLOE-26x는 텍스트 프롬프트에서 LVIS minival 기준 40.6 AP, 시각적 프롬프트에서 38.5 AP, 프롬프트 없이 31.1 AP를 달성합니다. 이는 논문의 Non-E2E 수치이며, 종단 간 헤드의 수치는 각각 1.1, 2.3, 1.2 AP 낮습니다.
스케일별 성능 표, 프롬프트 없는 변형 모델 및 전체 사용 예제는 YOLOE 문서를 참조하세요.
인용 및 감사의 글#
YOLO26 아키텍처, 학습 레시피, task head 및 YOLOE-26 개방형 어휘 확장에 대한 전체 기술 설명은 Ultralytics YOLO26: 통합 실시간 엔드투엔드 비전 모델을 참조하세요. 연구에 YOLO26을 사용하는 경우 다음을 인용해 주세요:
@misc{jocher2026ultralyticsyolo26unifiedrealtime,
title = {Ultralytics YOLO26: Unified Real-Time End-to-End Vision Models},
author = {Glenn Jocher and Jing Qiu and Mengyu Liu and Shuai Lyu and Fatih Cagatay Akyon and Muhammet Esat Kalfaoglu},
year = {2026},
eprint = {2606.03748},
archivePrefix = {arXiv},
primaryClass = {cs.CV},
doi = {10.48550/arXiv.2606.03748},
url = {https://arxiv.org/abs/2606.03748},
}YOLO26 코드, 모델 및 문서는 AGPL-3.0 및 기업용 라이선스에 따라 Ultralytics GitHub 저장소와 Ultralytics Docs에서 사용할 수 있습니다.
자주 묻는 질문#
- DFL-free 회귀: detection head와 export 경로를 간소화합니다.
- 엔드투엔드 NMS-free 추론:
nms=False으로 NMS-free 추론을 지원합니다. - Progressive Loss + STAL: 학습 정렬과 소형 객체의 라벨 커버리지를 개선합니다.
- MuSGD optimizer: SGD와 Muon에서 영감을 얻은 최적화를 결합하여 안정적인 학습을 지원합니다.
- 작업별 head 및 loss: segmentation, pose 및 oriented detection 지원을 개선합니다.
YOLO26은 다음과 같은 방식으로 배포 효율성을 높입니다:
- NMS 없이 실행하는 선택적 네이티브 엔드투엔드 추론(
nms=False) - DFL-free 회귀 및 더 경량화된 detection head
- 학습 전용 보조 구성 요소를 제거하는 융합 모델 export
- Intel Xeon CPU @ 2.00 GHz에서 YOLO26n은 YOLO11n보다 CPU ONNX 추론이 최대 43% 빠릅니다.
- TensorRT, ONNX, CoreML, LiteRT 및 OpenVINO를 포함하는 유연한 export 형식
- NMS 없이 실행하는 선택적 네이티브 엔드투엔드 추론(
YOLO26 모델은
ultralytics패키지를 통해 다운로드할 수 있습니다. 패키지를 설치하거나 업데이트한 다음 모델을 불러오세요:from ultralytics import YOLO # 사전 학습된 YOLO26 나노 모델을 불러옵니다 model = YOLO("yolo26n.pt") # 이미지에서 추론 실행 results = model("image.jpg")학습, validation 및 export 지침은 사용 예제 섹션을 참조하세요.
댓글117
질문을 하거나 결과를 공유하거나 다른 사람을 도와주세요.
hongzhuoxx
除了监督模型,我还期待着ultralytics可以推出高效准确易用的非监督瑕疵检测模型!
4Syedshajib55
How do I modify the architectural components (Head and Neck) of YOLOv26?
What is the process for inserting a Dropout layer at the end of the backbone in the YOLOv26 configuration?
2
Glenn JocherUltralyticsYOLO26 (no “v”) architecture edits are done in the model YAML, so start by copying
ultralytics/cfg/models/26/yolo26.yamlinto your project, modify thebackbone:/head:layer lists (the “neck” is part ofhead:), and train from that YAML following the Model YAML Configuration Guide.# yolo26_custom.yaml backbone: # ... existing backbone layers ... - [-1, 1, nn.Dropout, [0.2]] # add dropout at end of backbone (train-time only) head: # ... existing head/neck layers ...yolo train model=yolo26_custom.yaml data=your_dataset.yaml imgsz=640If you share which task model you’re editing (
detect/seg/pose/obb) and which variant (n/s/m/l/x), we can point you to the exact YAML to start from and where to insert the change in that graph.1Syedshajib55
Hi Glenn,
Thank you for the quick response and guidance! To answer your question: I am focusing on the detect task, and I will be experimenting with both the s (small) and m (medium) variants.
For context, I am working on a paper targeting a Q1 journal. My project focuses on Emotion Detection (5 classes: Angry, Happy, Sad, Neutral, Surprised) using the RAVDESS dataset. I have extracted peak emotion frames from the videos and augmented the dataset to roughly 21k images (640x640 resolution). My core research objective is to demonstrate that a customized YOLO26 architecture can outperform the standard baseline for detecting subtle facial micro-expressions. To achieve this, I want to modify the backbone—specifically by adding Batch Normalization and Dropout layers (potentially after the C2f blocks or may be after c3k2 blocks of that yolo26 architechtural images) to improve regularization and prevent overfitting on facial features.
Could you help me with the following?
-
Could you point me to the exact base YAML files for the yolo26s and yolo26m detection models to use as my starting point?
-
From an architectural standpoint, where exactly in the backbone would you theoretically recommend inserting these Dropout/BN layers? (e.g., after which specific layer index or block) to maximize accuracy without degrading the spatial feature maps necessary for facial detection?
Any insights on how to best structure this custom YAML would be incredibly helpful. Thanks again for your time!
-
Syedshajib55
- Baseline Reference: Finally, could you share the raw YAML configuration code that details the standard YOLO26 backbone and head? I would like to include this exact baseline structure in my paper’s methodology section to clearly contrast it with my proposed modifications.
hbzylilin
延迟了10天了,还没动静
1lwplw
真不错,期待开源
1
Glenn JocherUltralytics(수정됨)@etemical 感谢建议!目前未原生支持行为识别/多目标行为检测;建议先用 YOLO26 组合
detect+track多目标跟踪与可选pose关键点作为上游信号,并在 GitHub 提交详细需求(数据标注格式与评测指标)为 Feature Request 模板 便于评估;参考 Track 模式文档 与 Pose 任务文档。yolo track model=yolo26n.pt source=your_video.mp4 tracker=bytetrack.yaml show=True yolo pose predict model=yolo26n-pose.pt source=your_video.mp4 show=True@lwplw 谢谢关注!YOLO26 现已发布,可阅读我们的 YOLO26 介绍博客 了解功能。
Jiahao Chen
请问如果我想要引用本页面的YOLO26和其他模型的对比曲线图,已经在文中索引了来源。但是期刊要求需要拿到作者的授权,请问如何申请授权呢?
Paula DerrengerUltralytics您好,针对期刊要求的图片书面授权,建议联系 Ultralytics 法务团队,说明申请转载 YOLO26 对比曲线图,并附上页面链接、具体图片、论文题目、期刊名称、是否修改图片,以及期刊的授权模板(如有),请求确认版权归属并提供书面许可。目前未找到该图专门的公开授权流程,因此这里不能代为授予许可;请以收到的正式书面回复为准。
cyberpunk-edgerunner
maybe
CIFAR-10-LTdataset can help. you do not need background images because our initial conversation started without that, for an image, there can be one to 15 object in one image The cause of the shared-memory increase will remain unconfirmed unless you runpython train.pyin your computer with an rtx gpuOn Sat, Oct 10, 2026 at 6:15 AM Paula Derrenger @.***> wrote:
Understood—you don’t need to interrupt your work for the batch=16 comparison. Could you name a public long-tailed dataset you consider representative and share your approximate class counts, background-image proportion, and typical/maximum objects per image? That would help define a reproduction using the configuration you already shared, without requiring your company’s data or another production run. The cause of the shared-memory increase remains unconfirmed.
— Reply to this email directly, view it on GitHub https://github.com/orgs/ultralytics/discussions/26446?email_source=notifications&email_token=BSIM7DUZ6A5KFG7YKF2H6KT5TGBBRA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBYGQ2DONBZUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-18844749, or unsubscribe https://github.com/notifications/unsubscribe-auth/BSIM7DTBK4XV3HXIPMFHVXT5TGBBRAVCNFSNUABIKJSXA33TNF2G64TZHM2TGNJTGYYDINBVHNCGS43DOVZXG2LPNY5TCMBZGE3TMNRUUF3AE . Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS https://github.com/notifications/mobile/ios/BSIM7DST66YNYHAKPWPY3PT5TGBBRA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBYGQ2DONBZUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSVGM33PORSXEX3JN5ZQ and Android https://github.com/notifications/mobile/android/BSIM7DRDX2BJPHP7PHHKRG35TGBBRA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBYGQ2DONBZUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSXGM33PORSXEX3BNZSHE33JMQ. Download it today! You are receiving this because you commented.Message ID: @.*** com>
Paula DerrengerUltralyticsThanks for clarifying the 1–15 objects per image; we can keep background-only images out of the reproduction.
CIFAR-10-LTis a long-tailed classification benchmark, so it would not directly reproduce your multi-object detection setup. I can’t executepython train.pyon an RTX GPU here, and I haven’t reproduced the shared-memory increase. Could you share a public-data version of the script, including dataset preparation, so someone with suitable hardware can run the same test without your company’s data?
cyberpunk-edgerunner
I cannot run the same code with batch=16, reasons:
- I work in a company and have my own KPI to meet and improve
- I have added background image data in the dataset so it is no longer same.
- I am planning to add more images in future.
I suggest you download a long tailed distributed dataset to mimic my experiment.
thank you
On Fri, Oct 9, 2026 at 5:15 AM Paula Derrenger @.***> wrote:
Thanks for clarifying that it was 15 GB of shared GPU memory and sharing the configuration; sorry for requesting the script again. To test whether batch-related memory pressure contributes, could you compare your batch=64 run with batch=16, keeping all other settings unchanged? See the batch-size guidance https://docs.ultralytics.com/guides/model-training-tips#batch-size-and-gpu-utilization. Please share epoch 2–3 text logs and dedicated/shared GPU memory readings from the comparison; this will help narrow down the cause without assuming the memory change is a bug.
— Reply to this email directly, view it on GitHub https://github.com/orgs/ultralytics/discussions/26446?email_source=notifications&email_token=BSIM7DWLJ3KRNTXM7Y6Q4O35TARKLA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBYGI2TINRTUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-18825463, or unsubscribe https://github.com/notifications/unsubscribe-auth/BSIM7DQLZ4KWCGA6TEUDIS35TARKLAVCNFSNUABIKJSXA33TNF2G64TZHM2TGNJTGYYDINBVHNCGS43DOVZXG2LPNY5TCMBZGE3TMNRUUF3AE . Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS https://github.com/notifications/mobile/ios/BSIM7DV53PCZWPENVMYYNYT5TARKLA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBYGI2TINRTUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSVGM33PORSXEX3JN5ZQ and Android https://github.com/notifications/mobile/android/BSIM7DX2XUXERTKYLIDJ2C35TARKLA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBYGI2TINRTUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSXGM33PORSXEX3BNZSHE33JMQ. Download it today! You are receiving this because you commented.Message ID: @.*** com>
Paula DerrengerUltralyticsUnderstood—you don’t need to interrupt your work for the
batch=16comparison. Could you name a public long-tailed dataset you consider representative and share your approximate class counts, background-image proportion, and typical/maximum objects per image? That would help define a reproduction using the configuration you already shared, without requiring your company’s data or another production run. The cause of the shared-memory increase remains unconfirmed.
cyberpunk-edgerunner
- I have shared you exact code in this email thread(check my second reply) and the command to run it was `python .\det_yoloM_5-10.py' .
- yes, it was 15GB of shared GPU
- sorry, I do not have
separate dedicated/shared memory readings at comparable batch positions in each epoch, but there was a steady consumption of about 15GB of shared memory throughout the training. however I can provide you argos.yaml
task: detect mode: train model: yolo26m.pt data: yoloDet_clsAgnostic_split/data.yaml epochs: 200 time: null patience: 20 batch: 64 imgsz: 640 save: true save_period: -1 cache: false device: '0' workers: 6 project: null name: train-2 exist_ok: false pretrained: true cls_remap: true optimizer: auto verbose: false seed: 0 deterministic: true single_cls: false rect: false cos_lr: false close_mosaic: 10 resume: false amp: true fraction: 1.0 profile: false freeze: null multi_scale: 0.0 compile: false channels_last: null overlap_mask: true mask_ratio: 4 dropout: 0.0 val: true split: val save_json: false conf: null iou: 0.7 max_det: 300 quantize: null dnn: false plots: true nms: null source: null vid_stride: 1 stream_buffer: false visualize: false augment: false agnostic_nms: false classes: null retina_masks: false embed: null show: false save_frames: false save_txt: false save_conf: false save_crop: false show_labels: true show_conf: true show_boxes: true line_width: null format: torchscript keras: false optimize: false dynamic: false simplify: true opset: null workspace: null lr0: 0.01 lrf: 0.01 momentum: 0.937 weight_decay: 0.0005 warmup_epochs: 3.0 warmup_momentum: 0.8 warmup_bias_lr: 0.1 distill_model: null dis: 6.0 box: 7.5 cls: 0.5 cls_pw: 0.0 dfl: 1.5 pose: 12.0 kobj: 1.0 rle: 1.0 angle: 1.0 dlog: 1.0 dgrad: 0.5 dlam: 1.0 nbs: 64 hsv_h: 0.015 hsv_s: 0.7 hsv_v: 0.4 degrees: 0.0 translate: 0.1 scale: 0.5 shear: 0.0 perspective: 0.0 flipud: 0.0 fliplr: 0.5 bgr: 0.0 mosaic: 1.0 mixup: 0.0 cutmix: 0.0 copy_paste: 0.0 copy_paste_mode: flip auto_augment: randaugment erasing: 0.4 cfg: null tracker: tracktrack.yaml save_dir: C:\Training9\runs\detect\train-2On Thu, Oct 8, 2026 at 6:07 AM Paula Derrenger @.***> wrote:
Thanks for the complete checks and text logs. Epoch 2 is faster (14:27 versus 1:02:39), despite GPU_mem dropping from 28.6G to 17.4G, so these readings alone don’t establish GPU underutilization. Could you paste your exact training command or script, including the model and training settings? Also, does “around 15 epoch” mean 15 GB of shared GPU memory in both epochs? Please provide separate dedicated/shared memory readings at comparable batch positions in each epoch so we can investigate the discrepancy.
— Reply to this email directly, view it on GitHub https://github.com/orgs/ultralytics/discussions/26446?email_source=notifications&email_token=BSIM7DQKA3GIYQSCDVEWEJ35S3OWHA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBYGAZTKNJWUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-18803556, or unsubscribe https://github.com/notifications/unsubscribe-auth/BSIM7DX34MSUWFFX5WDRXW35S3OWHAVCNFSNUABIKJSXA33TNF2G64TZHM2TGNJTGYYDINBVHNCGS43DOVZXG2LPNY5TCMBZGE3TMNRUUF3AE . Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS https://github.com/notifications/mobile/ios/BSIM7DSVPQ3LAONGDZOZWWD5S3OWHA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBYGAZTKNJWUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSVGM33PORSXEX3JN5ZQ and Android https://github.com/notifications/mobile/android/BSIM7DQHEDECXNLNQEWJLC35S3OWHA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBYGAZTKNJWUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSXGM33PORSXEX3BNZSHE33JMQ. Download it today! You are receiving this because you commented.Message ID: @.*** com>
Paula DerrengerUltralyticsThanks for clarifying that it was 15 GB of shared GPU memory and sharing the configuration; sorry for requesting the script again. To test whether batch-related memory pressure contributes, could you compare your
batch=64run withbatch=16, keeping all other settings unchanged? See the batch-size guidance. Please share epoch 2–3 text logs and dedicated/shared GPU memory readings from the comparison; this will help narrow down the cause without assuming the memory change is a bug.
cyberpunk-edgerunner
here are my yolo checks yolo checks Ultralytics 8.4.155 Python-3.13.12 torch-2.12.0.dev20260408+cu128 CUDA:0 (NVIDIA GeForce RTX 5090, 32607MiB) Setup complete (24 CPUs, 63.5 GB RAM, 1738.9/1881.5 GB disk)
OS Windows-11-10.0.26200-SP0 Environment Windows Python 3.13.12 Install pip Path C:\Users\devth\miniconda3\Lib\site-packages\ultralytics RAM 63.46 GB Disk 1738.9/1881.5 GB CPU Intel Core Ultra 9 285K CPU count 24 GPU NVIDIA GeForce RTX 5090, 32607MiB GPU count 1 CUDA 12.8
cloudpickle 3.1.2>=3.1.1 filelock 3.29.0>=3.16.1 numpy 2.4.4>=1.23.0; platform_system != "Darwin" numpy 2.4.4!=2.0.,!=2.1.,!=2.2.*,!=2.3.0,!=2.3.1,!=2.3.2,!=2.3.3,!=2.3.4,>=1.23.0; platform_system == "Darwin" matplotlib 3.10.8>=3.3.0 opencv-python 4.13.0.92!=4.13.0.90,>=4.7.0 pillow 12.2.0>=7.1.2 pyyaml 6.0.3>=5.3.1 requests 2.32.5>=2.23.0 torch 2.12.0.dev20260408+cu128>=1.8.0 torch 2.12.0.dev20260408+cu128!=2.4.0,>=1.8.0; sys_platform == "win32" torchvision 0.27.0.dev20260407+cu128>=0.9.0 psutil 7.2.2>=5.8.0 polars 1.40.1>=0.20.0 nvidia-ml-py 13.610.43>=12.0.0 ultralytics-thop 2.1.6>=2.1.6 ultralytics-platform 0.1.48>=0.1.45; python_version >= "3.11"
(shared memory used in first and second epoch was around 15 epoch) these are the first two epoch details in Text
Epoch GPU_mem box_loss cls_loss l1_loss Instances Size 1/200 28.6G 1.174 1.891 0.01363 90640: 33% ━━━╸──────── 1561/4702 1.1it/s 23:55<49:25libpng warning: iCCP: known incorrect sRGB profile 1/200 28.6G 1.15 1.582 0.01327 91 640: 56% ━━━━━━╸───── 2674/4702 1.3it/s 40:45<26:43libpng warning: iCCP: known incorrect sRGB profile 1/200 28.6G 1.144 1.475 0.01325 77 640: 72% ━━━━━━━━╸─── 3419/4702 1.3it/s 49:44<16:45libpng warning: iCCP: known incorrect sRGB profile 1/200 28.6G 1.144 1.471 0.01325 84 640: 73% ━━━━━━━━╸─── 3451/4702 1.4it/s 50:07<14:41libpng warning: iCCP: known incorrect sRGB profile 1/200 28.6G 1.146 1.395 0.01332 84 640: 92% ━━━━━━━━━━━─ 4370/4702 5.0it/s 1:01:16<1:07libpng warning: iCCP: known incorrect sRGB profile 1/200 28.6G 1.146 1.373 0.01335 74 640: 100% ━━━━━━━━━━━━ 4702/4702 1.3it/s 1:02:39 Class Images Instances Box(P R mAP50 mAP50-95): 22% ━━╸───────── 57/256 2.5it/s 26.7s<1:19libpng warning: iCCP: known incorrect sRGB profile Class Images Instances Box(P R mAP50 mAP50-95): 47% ━━━━━╸────── 122/256 2.3it/s 1:10<59.4slibpng warning: iCCP: known incorrect sRGB profile Class Images Instances Box(P R mAP50 mAP50-95): 100% ━━━━━━━━━━━━ 256/256 1.9it/s 2:17 all 32677 47931 0.863 0.726 0.818 0.584
Epoch GPU_mem box_loss cls_loss l1_loss InstancesSize 2/200 17.4G 1.183 1.123 0.01413 74 640: 29% ━━━╸──────── 1380/4702 5.1it/s 4:16<10:45libpng warning: iCCP: known incorrect sRGB profile 2/200 17.4G 1.196 1.133 0.01441 107 640: 55% ━━━━━━╸───── 2615/4702 5.8it/s 8:09<6:02libpng warning: iCCP: known incorrect sRGB profile 2/200 17.4G 1.193 1.126 0.01434 99 640: 64% ━━━━━━━╸──── 3020/4702 5.5it/s 9:20<5:06libpng warning: iCCP: known incorrect sRGB profile 2/200 17.4G 1.193 1.125 0.01435 106 640: 67% ━━━━━━━━──── 3179/4702 5.6it/s 9:50<4:33libpng warning: iCCP: known incorrect sRGB profile 2/200 17.4G 1.192 1.121 0.01434 109 640: 81% ━━━━━━━━━╸── 3847/4702 5.5it/s 11:52<2:36libpng warning: iCCP: known incorrect sRGB profile 2/200 17.4G 1.195 1.12 0.01437 86 640: 100% ━━━━━━━━━━━━ 4702/4702 5.4it/s 14:27 Class Images Instances Box(P R mAP50 mAP50-95): 22% ━━╸───────── 57/256 2.5it/s 20.6s<1:21libpng warning: iCCP: known incorrect sRGB profile Class Images Instances Box(P R mAP50 mAP50-95): 47% ━━━━━╸────── 122/256 2.1it/s 50.3s<1:04libpng warning: iCCP: known incorrect sRGB profile Class Images Instances Box(P R mAP50 mAP50-95): 100% ━━━━━━━━━━━━ 256/256 2.3it/s 1:51 all 32677 47931 0.889 0.778 0.87 0.638
On Tue, Oct 6, 2026 at 10:29 AM Paula Derrenger @.***> wrote:
No—by “yolo checks output,” I mean the terminal’s software/hardware summary, not results.csv. Please activate the same Python environment used for training, run yolo checks in PowerShell, and paste its complete text output; it reports your Ultralytics, Python, PyTorch, and GPU details (environment checks reference https://docs.ultralytics.com/reference/utils/checks#ultralytics.utils.checks.check_yolo). (docs.ultralytics.com https://docs.ultralytics.com/reference/utils/checks) Thanks for sharing the driver information. Please also provide separate dedicated/shared GPU-memory readings at matching points in epochs 1 and 2, plus your training command or configuration, so we can investigate without assuming the memory drop indicates hardware underutilization.
— Reply to this email directly, view it on GitHub https://github.com/orgs/ultralytics/discussions/26446?email_source=notifications&email_token=BSIM7DSK7MKPVE56RJJGI435SR33BA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBXG4YDKNJSUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-18770552, or unsubscribe https://github.com/notifications/unsubscribe-auth/BSIM7DRBVYTDNBPWAZRUQXL5SR33BAVCNFSNUABIKJSXA33TNF2G64TZHM2TGNJTGYYDINBVHNCGS43DOVZXG2LPNY5TCMBZGE3TMNRUUF3AE . Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS https://github.com/notifications/mobile/ios/BSIM7DXMRZI3YZR75VFDAGT5SR33BA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBXG4YDKNJSUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSVGM33PORSXEX3JN5ZQ and Android https://github.com/notifications/mobile/android/BSIM7DRXQCNW5JUZNGVDHU35SR33BA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBXG4YDKNJSUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSXGM33PORSXEX3BNZSHE33JMQ. Download it today! You are receiving this because you commented.Message ID: @.*** com>
Paula DerrengerUltralyticsThanks for the complete checks and text logs. Epoch 2 is faster (14:27 versus 1:02:39), despite
GPU_memdropping from 28.6G to 17.4G, so these readings alone don’t establish GPU underutilization. Could you paste your exact training command or script, including the model and training settings? Also, does “around 15 epoch” mean 15 GB of shared GPU memory in both epochs? Please provide separate dedicated/shared memory readings at comparable batch positions in each epoch so we can investigate the discrepancy.
cyberpunk-edgerunner
pls explain what is yolo checks output? what do you mean by that ? are you referring to results.csv from runs folder??
here is my driver detail PS C:\Training9> nvidia-smi Tue Oct 6 09:22:05 2026 +-----------------------------------------------------------------------------------------+ | NVIDIA-SMI 610.60 KMD Version: 610.60 CUDA UMD Version: 13.3 | +-----------------------------------------+------------------------+----------------------+ | GPU Name Driver-Model | Bus-Id Disp.A | Volatile Uncorr. ECC | | Fan Temp Perf Pwr:Usage/Cap | Memory-Usage | GPU-Util Compute M. | | | | MIG M. | |=========================================+========================+======================| | 0 NVIDIA GeForce RTX 5090 WDDM | 00000000:01:00.0 On | N/A | | 57% 73C P1 479W / 575W | 19362MiB / 32607MiB | 80% Default | | | | N/A | +-----------------------------------------+------------------------+----------------------+
PS C:\Training9> nvcc --version nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2025 NVIDIA Corporation Built on Wed_Jan_15_19:38:46_Pacific_Standard_Time_2025 Cuda compilation tools, release 12.8, V12.8.61 Build cuda_12.8.r12.8/compiler.35404655_0
epoch 1 vram usage(in terminal) = 28.6 GB and in task manager ~29.1 GB epoch 2 to epoch 73 vram usage (in terminal) = 17.4 GB and in task manager vram usage is 18.1 GB
73rd epoch
Epoch GPU_mem box_loss cls_loss l1_loss InstancesSize 73/200 17.4G 0.9935 0.7903 0.01002 86 640: 0% ──────────── 4/4702 4.4it/s 1.0s<17:51libpng warning: iCCP: known incorrect sRGB profile 73/200 17.4G 1.023 0.7368 0.01159 84 640: 9% ━─────────── 446/4702 5.8it/s 1:17<12:17libpng warning: iCCP: known incorrect sRGB profile 73/200 17.4G 1.024 0.7397 0.01152 93 640: 63% ━━━━━━━╸──── 2990/4702 5.6it/s 8:41<5:06libpng warning: iCCP: known incorrect sRGB profile 73/200 17.4G 1.024 0.7402 0.01151 119 640: 65% ━━━━━━━╸──── 3080/4702 5.3it/s 8:57<5:06libpng warning: iCCP: known incorrect sRGB profile 73/200 17.4G 1.026 0.7422 0.01152 113 640: 76% ━━━━━━━━━─── 3586/4702 5.6it/s 10:28<3:20libpng warning: iCCP: known incorrect sRGB profile 73/200 17.4G 1.029 0.7459 0.01156 78 640: 100% ━━━━━━━━━━━━ 4702/4702 5.7it/s 13:46 Class Images Instances Box(P R mAP50 mAP50-95): 22% ━━╸───────── 57/256 2.7it/s 20.9s<1:13libpng warning: iCCP: known incorrect sRGB profile Class Images Instances Box(P R mAP50 mAP50-95): 47% ━━━━━╸────── 122/256 2.1it/s 48.9s<1:02libpng warning: iCCP: known incorrect sRGB profile Class Images Instances Box(P R mAP50 mAP50-95): 100% ━━━━━━━━━━━━ 256/256 2.4it/s 1:47 all 32677 47931 0.927 0.855 0.93 0.724
since you are trying to replicate my detection code , I must tell you in my train folder I have around 150000 images with txt files, 32000 in val and 31000 in test
while you are at it, I need a change in the ultralytics package, since it can detect and restore corrupt JPG(and many more errors), would it possible for you to make this change permanent? since it a type of fix and everyone wants that and no one wants to ultralytics package to just internally fix the images i know it won't be implemented by of security reasons by just keep in future todo list
to be honest , in my view, the ultralytics package does not fully utilise the hardware. I have run barebone code that using just the pytorch and few more basic codes and they utilise
thank you in advance
On Mon, Oct 5, 2026 at 10:42 PM Paula Derrenger @.***> wrote:
Thanks for sharing the script. A plausible explanation for the drop from 28.6G to 17.4G is cache cleanup between epochs: the training loop https://docs.ultralytics.com/reference/engine/trainer#ultralytics.engine.trainer.BaseTrainer._do_train conditionally clears reserved memory before validation and at epoch end when usage exceeds 50% of device capacity. This could explain the drop, but does not establish why Windows reports shared-memory usage. ( docs.ultralytics.com https://docs.ultralytics.com/reference/engine/trainer) Please share yolo checks output, your NVIDIA driver version, and dedicated/shared GPU-memory readings as text at matching points in both epochs so we can check your installed version and distinguish the two observations.
— Reply to this email directly, view it on GitHub https://github.com/orgs/ultralytics/discussions/26446?email_source=notifications&email_token=BSIM7DSOVWTY24AP4M6GU5D5SPI7FA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBXGYZTCOJVUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-18763195, or unsubscribe https://github.com/notifications/unsubscribe-auth/BSIM7DUUEOVQC4A4Q7SCGUL5SPI7FAVCNFSNUABIKJSXA33TNF2G64TZHM2TGNJTGYYDINBVHNCGS43DOVZXG2LPNY5TCMBZGE3TMNRUUF3AE . Triage notifications, keep track of coding agent tasks and review pull requests on the go with GitHub Mobile for iOS https://github.com/notifications/mobile/ios/BSIM7DVWK73UY4PTSMIRTTL5SPI7FA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBXGYZTCOJVUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSVGM33PORSXEX3JN5ZQ and Android https://github.com/notifications/mobile/android/BSIM7DW6Q4F7VJMCOTRDMPD5SPI7FA5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCOBXGYZTCOJVUZZGKYLTN5XKOY3PNVWWK3TUUVSXMZLOOSXGM33PORSXEX3BNZSHE33JMQ. Download it today! You are receiving this because you commented.Message ID: @.*** com>
Paula DerrengerUltralyticsNo—by “
yolo checksoutput,” I mean the terminal’s software/hardware summary, notresults.csv. Please activate the same Python environment used for training, runyolo checksin PowerShell, and paste its complete text output; it reports your Ultralytics, Python, PyTorch, and GPU details (environment checks reference). (docs.ultralytics.com) Thanks for sharing the driver information. Please also provide separate dedicated/shared GPU-memory readings at matching points in epochs 1 and 2, plus your training command or configuration, so we can investigate without assuming the memory drop indicates hardware underutilization.