训练精度RayIoU较低

您好，我的config文件如下所示
`dataset_type = 'NuSceneOcc'
dataset_root = 'data/nuscenes/'
occ_gt_root = 'data/nuscenes/occ3d'

If point cloud range is changed, the models should also change their point

cloud range accordingly

point_cloud_range = [-40, -40, -1.0, 40, 40, 5.4]
occ_size = [200, 200, 16]

img_norm_cfg = dict(
mean=[123.675, 116.280, 103.530],
std=[58.395, 57.120, 57.375],
to_rgb=True
)

For nuScenes we usually do 10-class detection

det_class_names = [
'car', 'truck', 'construction_vehicle', 'bus', 'trailer', 'barrier',
'motorcycle', 'bicycle', 'pedestrian', 'traffic_cone'
]

occ_class_names = [
'others', 'barrier', 'bicycle', 'bus', 'car', 'construction_vehicle',
'motorcycle', 'pedestrian', 'traffic_cone', 'trailer', 'truck',
'driveable_surface', 'other_flat', 'sidewalk',
'terrain', 'manmade', 'vegetation', 'free'
]

input_modality = dict(
use_lidar=False,
use_camera=True,
use_radar=False,
use_map=False,
use_external=False
)

dim = 256
num_points = 4
num_groups = 4
num_layers = 2
num_frames = 2
num_queries = 100
topk_training = [4000, 16000, 64000]
topk_testing = [2000, 8000, 32000]

model = dict(
type='SparseOcc',
data_aug=dict(
img_color_aug=True, # Move some augmentations to GPU
img_norm_cfg=img_norm_cfg,
img_pad_cfg=dict(size_divisor=32)),
use_mask_camera=False,
img_backbone=dict(
type='ResNet',
depth=50,
num_stages=4,
out_indices=(0, 1, 2, 3),
frozen_stages=1,
norm_cfg=dict(type='BN2d', requires_grad=True),
norm_eval=True,
style='pytorch',
with_cp=True),
img_neck=dict(
type='FPN',
in_channels=[256, 512, 1024, 2048],
out_channels=dim,
num_outs=4),
pts_bbox_head=dict(
type='SparseOccHead',
class_names=occ_class_names,
embed_dims=dim,
occ_size=occ_size,
pc_range=point_cloud_range,
transformer=dict(
type='SparseOccTransformer',
embed_dims=dim,
num_layers=num_layers,
num_frames=num_frames,
num_points=num_points,
num_groups=num_groups,
num_queries=num_queries,
num_levels=4,
num_classes=len(occ_class_names),
pc_range=point_cloud_range,
occ_size=occ_size,
topk_training=topk_training,
topk_testing=topk_testing),
loss_cfgs=dict(
loss_mask2former=dict(
type='Mask2FormerLoss',
num_classes=len(occ_class_names),
no_class_weight=0.1,
loss_cls_weight=2.0,
loss_mask_weight=5.0,
loss_dice_weight=5.0,
),
loss_geo_scal=dict(
type='GeoScalLoss',
num_classes=len(occ_class_names),
loss_weight=1.0
),
loss_sem_scal=dict(
type='SemScalLoss',
num_classes=len(occ_class_names),
loss_weight=1.0
)
),
),
)

ida_aug_conf = {
'resize_lim': (0.38, 0.55),
'final_dim': (256, 704),
'bot_pct_lim': (0.0, 0.0),
'rot_lim': (0.0, 0.0),
'H': 900, 'W': 1600,
'rand_flip': True,
}

bda_aug_conf = dict(
rot_lim=(-22.5, 22.5),
scale_lim=(1., 1.),
flip_dx_ratio=0.5,
flip_dy_ratio=0.5
)

train_pipeline = [
dict(type='LoadMultiViewImageFromFiles', to_float32=False, color_type='color'),
dict(type='LoadMultiViewImageFromMultiSweeps', sweeps_num=num_frames - 1),
dict(type='BEVAug', bda_aug_conf=bda_aug_conf, classes=det_class_names, is_train=True),
dict(type='LoadOccGTFromFile', num_classes=len(occ_class_names)),
dict(type='RandomTransformImage', ida_aug_conf=ida_aug_conf, training=True),
dict(type='DefaultFormatBundle3D', class_names=det_class_names),
dict(type='Collect3D', keys=['img', 'voxel_semantics', 'voxel_instances', 'instance_class_ids'], # other keys: 'mask_camera'
meta_keys=('filename', 'ori_shape', 'img_shape', 'pad_shape', 'lidar2img', 'img_timestamp', 'ego2lidar'))
]

test_pipeline = [
dict(type='LoadMultiViewImageFromFiles', to_float32=False, color_type='color'),
dict(type='LoadMultiViewImageFromMultiSweeps', sweeps_num=num_frames - 1, test_mode=True),
dict(type='BEVAug', bda_aug_conf=bda_aug_conf, classes=det_class_names, is_train=False),
dict(type='LoadOccGTFromFile', num_classes=len(occ_class_names)),
dict(type='RandomTransformImage', ida_aug_conf=ida_aug_conf, training=False),
dict(type='DefaultFormatBundle3D', class_names=det_class_names),
dict(type='Collect3D', keys=['img', 'voxel_semantics', 'voxel_instances', 'instance_class_ids'],
meta_keys=('filename', 'ori_shape', 'img_shape', 'pad_shape', 'lidar2img', 'img_timestamp', 'ego2lidar'))
]

data = dict(
workers_per_gpu=8,
train=dict(
type=dataset_type,
data_root=dataset_root,
occ_gt_root=occ_gt_root,
ann_file=dataset_root + 'sparsebev_infos_train_sweep.pkl',
pipeline=train_pipeline,
classes=det_class_names,
modality=input_modality,
test_mode=False
),
val=dict(
type=dataset_type,
data_root=dataset_root,
occ_gt_root=occ_gt_root,
ann_file=dataset_root + 'sparsebev_infos_val_sweep.pkl',
pipeline=test_pipeline,
classes=det_class_names,
modality=input_modality,
test_mode=True
),
test=dict(
type=dataset_type,
data_root=dataset_root,
occ_gt_root=occ_gt_root,
ann_file=dataset_root + 'sparsebev_infos_test_sweep.pkl',
pipeline=test_pipeline,
classes=det_class_names,
modality=input_modality,
test_mode=True
),
)

optimizer = dict(
type='AdamW',
lr=5e-4,
paramwise_cfg=dict(
custom_keys={
'img_backbone': dict(lr_mult=0.1),
'sampling_offset': dict(lr_mult=0.1),
}),
weight_decay=0.01
)
optimizer_config = dict(grad_clip=dict(max_norm=35, norm_type=2))

lr_config = dict(
policy='step',
warmup='linear',
warmup_iters=500,
warmup_ratio=1.0 / 3,
by_epoch=True,
step=[22, 24],
gamma=0.2
)
total_epochs = 24
batch_size = 8

load pretrained weights

load_from = 'pretrain/cascade_mask_rcnn_r50_fpn_coco-20e_20e_nuim_20201009_124951-40963960.pth'
revise_keys = [('backbone', 'img_backbone')]

resume the last training

resume_from = None

checkpointing

checkpoint_config = dict(interval=1, max_keep_ckpts=1)

logging

log_config = dict(
interval=1,
hooks=[
dict(type='MyTextLoggerHook', interval=1, reset_flag=True),
dict(type='MyTensorboardLoggerHook', interval=500, reset_flag=True)
]
)

evaluation

eval_config = dict(interval=total_epochs)

other flags

debug = False`
但是训练24个epoch之后
[2025-11-05 17:54:25,594][INFO] - Saving checkpoint at 24 epochs
[2025-11-05 18:05:45,133][INFO] - --- Evaluation Results ---
[2025-11-05 18:05:45,133][INFO] - RayIoU: 0.3296
RayIoU只能到达0.33，达不到论文中的0.34，请问大致是什么原因呢

Start with the configuration and the final evaluation log in the issue, then compare the dataset paths, occupancy classes, augmentation settings, pretrained weights, epoch schedule, and RayIoU evaluation setup with the paper's reported configuration. Reproduce the 0.3296 result and identify which setting or evaluation condition accounts for the gap to 0.34; no source file or test is named in the report.