#25570·ultralytics

Question About YOLOE-26 Result

Author: zixi01chenCreated Jul 31, 2026Updated Sep 14, 2026
LabelsquestionStale

Search before asking

  • I have searched the Ultralytics YOLO issues and discussions and found no similar questions.

Question

Thanks for the work YOLO26 https://arxiv.org/abs/2606.03748. I noticed that the result of YOLOE-26M result is 35.4 on LVIS minival dataset.

However, i use the code from https://docs.ultralytics.com/models/yolo26#supported-tasks-and-modes

from ultralytics import YOLOE

model = YOLOE("yoloe-26m-seg.pt")  # or yoloe-26s/m-seg.pt for different sizes

metrics = model.val(data="lvis.yaml")

But I get result:

DONE (t=28.35s).
Accumulating evaluation results...
COCOeval_opt.accumulate() finished...
DONE (t=0.00s).
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 catIds=all] = 0.328
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 catIds=all] = 0.429
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 catIds=all] = 0.356
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 catIds=all] = 0.223
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 catIds=all] = 0.419
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 catIds=all] = 0.540
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 catIds=  r] = 0.326
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 catIds=  c] = 0.320
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 catIds=  f] = 0.336
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 catIds=all] = 0.285
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 catIds=all] = 0.415
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 catIds=all] = 0.424
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 catIds=all] = 0.269
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 catIds=all] = 0.511
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 catIds=all] = 0.667
 Average Recall     (AR) @[ IoU=0.50      | area=   all | maxDets=100 catIds=all] = 0.536
 Average Recall     (AR) @[ IoU=0.75      | area=   all | maxDets=100 catIds=all] = 0.460
Evaluate annotation type *segm*
COCOeval_opt.evaluate() finished...
DONE (t=49.92s).
Accumulating evaluation results...
COCOeval_opt.accumulate() finished...
DONE (t=0.00s).
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 catIds=all] = 0.278
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 catIds=all] = 0.410
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 catIds=all] = 0.293
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 catIds=all] = 0.165
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 catIds=all] = 0.370
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 catIds=all] = 0.480
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 catIds=  r] = 0.283
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 catIds=  c] = 0.281
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 catIds=  f] = 0.273
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 catIds=all] = 0.247
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 catIds=all] = 0.352
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 catIds=all] = 0.359
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 catIds=all] = 0.198
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 catIds=all] = 0.455
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 catIds=all] = 0.596
 Average Recall     (AR) @[ IoU=0.50      | area=   all | maxDets=100 catIds=all] = 0.514
 Average Recall     (AR) @[ IoU=0.75      | area=   all | maxDets=100 catIds=all] = 0.383

It not equal to 35.4. I wonder how can I reproduce the YOLOE-26M Result? Thanks.

Additional

No response