Unable to reproduce the accuracy on the COCO dataset as reported in the paper

[1826133674]

I am trying to reproduce the accuracy of FastSAM on the COCO dataset as reported in the Chinese paper, but I have encountered the following issues: How can I fix these issues?

1. Unable to generate plots. The plot_mc_curve function in ap_per_class within the file path /FastSAM/ultralytics/yolo/utils/metrics.py raises an error. The error is as follows:

Traceback (most recent call last):
  File "/github/FastSAM/val_sa.py", line 4, in <module>
    model.val(data="sa_0.yaml", \
  File "/miniforge3/envs/FastSAM/lib/python3.9/site-packages/torch/utils/_contextlib.py", line 120, in decorate_context
    return func(*args, **kwargs)
  File "/github/FastSAM/ultralytics/yolo/engine/model.py", line 303, in val
    validator(model=self.model)
  File "/miniforge3/envs/FastSAM/lib/python3.9/site-packages/torch/utils/_contextlib.py", line 120, in decorate_context
    return func(*args, **kwargs)
  File "/github/FastSAM/ultralytics/yolo/engine/validator.py", line 148, in __call__
    desc = self.get_desc()
  File "/github/FastSAM/ultralytics/yolo/v8/detect/val.py", line 128, in get_stats
    self.metrics.process(*stats)
  File "/github/FastSAM/ultralytics/yolo/utils/metrics.py", line 769, in process
    results_mask = ap_per_class(tp_m,
  File "/github/FastSAM/ultralytics/yolo/utils/metrics.py", line 488, in ap_per_class
    plot_mc_curve(px, f1, save_dir / f'{prefix}F1_curve.png', names, ylabel='F1', on_plot=on_plot)
  File "/github/FastSAM/ultralytics/yolo/utils/__init__.py", line 196, in wrapper
    result = func(*args, **kwargs)
  File "/github/FastSAM/ultralytics/yolo/utils/metrics.py", line 358, in plot_mc_curve
    ax.plot(px, y, linewidth=1, label=f'{names[i]}')  # plot(confidence, metric)
KeyError: 1

2. The actual data involved in the test is very small:
   After commenting out the plotting code, it was found that there were only 76 images in the entire COCO test dataset; the remaining 4,000+ images all generated such warnings and were filtered out.

coco_segs/images/val2017/000000581781.jpg: ignoring corrupt image/label: Label class 46 exceeds dataset class count 1. Possible class labels are 0-0

The final test results are output as follows:

Evaluating pycocotools mAP using fastsam/val83/predictions.json and dataset/coco_segs/annotations/instances_val2017.json...
loading annotations into memory...
Done (t=0.39s)
creating index...
index created!
Loading and preparing results...
DONE (t=0.02s)
creating index...
index created!
Running per image evaluation...
Evaluate annotation type *bbox*
DONE (t=0.09s).
Accumulating evaluation results...
DONE (t=0.04s).
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.011
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.017
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.011
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.003
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.013
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.020
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.000
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.168
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.366
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.189
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.354
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.448
Running per image evaluation...
Evaluate annotation type *segm*
DONE (t=0.10s).
Accumulating evaluation results...
DONE (t=0.04s).
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.009
 Average Precision  (AP) @[ IoU=0.50      | area=   all | maxDets=100 ] = 0.015
 Average Precision  (AP) @[ IoU=0.75      | area=   all | maxDets=100 ] = 0.011
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.003
 Average Precision  (AP) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.010
 Average Precision  (AP) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.018
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=  1 ] = 0.004
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets= 10 ] = 0.146
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=   all | maxDets=100 ] = 0.319
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= small | maxDets=100 ] = 0.172
 Average Recall     (AR) @[ IoU=0.50:0.95 | area=medium | maxDets=100 ] = 0.315
 Average Recall     (AR) @[ IoU=0.50:0.95 | area= large | maxDets=100 ] = 0.380

The sa.yaml settings I adopted are as follows:

path: '/dataset/coco_segs/'

train: 'train2017.txt'

val: 'val2017.txt'
test: 'val2017.txt'

# Classes
names:
   0: object