mirrors/pytorch-image-models
1
0
Fork 0
mirror of https://github.com/huggingface/pytorch-image-models.git synced 2025-06-03 15:01:08 +08:00
Code Issues Projects Releases Wiki Activity

Allow group_size override for more efficientnet and mobilenetv3 based models

Browse Source
This commit is contained in:
Ross Wightman 2024-08-21 16:51:38 -07:00
parent 00c5be7656
commit b9f020a509
2 changed files with 38 additions and 18 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

44
timm/models/efficientnet.py
View File

@ -488,7 +488,7 @@ def _gen_mnasnet_small(variant, channel_multiplier=1.0, pretrained=False, **kwar
def _gen_mobilenet_v1( def _gen_mobilenet_v1(
variant, channel_multiplier=1.0, depth_multiplier=1.0, variant, channel_multiplier=1.0, depth_multiplier=1.0,
fix_stem_head=False, head_conv=False, pretrained=False, **kwargs): group_size=None, fix_stem_head=False, head_conv=False, pretrained=False, **kwargs):
""" """
Ref impl: https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet_v2.py Ref impl: https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet_v2.py
Paper: https://arxiv.org/abs/1801.04381 Paper: https://arxiv.org/abs/1801.04381
@ -503,7 +503,12 @@ def _gen_mobilenet_v1(
round_chs_fn = partial(round_channels, multiplier=channel_multiplier) round_chs_fn = partial(round_channels, multiplier=channel_multiplier)
head_features = (1024 if fix_stem_head else max(1024, round_chs_fn(1024))) if head_conv else 0 head_features = (1024 if fix_stem_head else max(1024, round_chs_fn(1024))) if head_conv else 0
model_kwargs = dict( model_kwargs = dict(
block_args=decode_arch_def(arch_def, depth_multiplier=depth_multiplier, fix_first_last=fix_stem_head), block_args=decode_arch_def(
arch_def,
depth_multiplier=depth_multiplier,
fix_first_last=fix_stem_head,
group_size=group_size,
),
num_features=head_features, num_features=head_features,
stem_size=32, stem_size=32,
fix_stem=fix_stem_head, fix_stem=fix_stem_head,
@ -517,7 +522,9 @@ def _gen_mobilenet_v1(
def _gen_mobilenet_v2( def _gen_mobilenet_v2(
variant, channel_multiplier=1.0, depth_multiplier=1.0, fix_stem_head=False, pretrained=False, **kwargs): variant, channel_multiplier=1.0, depth_multiplier=1.0,
group_size=None, fix_stem_head=False, pretrained=False, **kwargs
):
""" Generate MobileNet-V2 network """ Generate MobileNet-V2 network
Ref impl: https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet_v2.py Ref impl: https://github.com/tensorflow/models/blob/master/research/slim/nets/mobilenet/mobilenet_v2.py
Paper: https://arxiv.org/abs/1801.04381 Paper: https://arxiv.org/abs/1801.04381
@ -533,7 +540,12 @@ def _gen_mobilenet_v2(
] ]
round_chs_fn = partial(round_channels, multiplier=channel_multiplier) round_chs_fn = partial(round_channels, multiplier=channel_multiplier)
model_kwargs = dict( model_kwargs = dict(
block_args=decode_arch_def(arch_def, depth_multiplier=depth_multiplier, fix_first_last=fix_stem_head), block_args=decode_arch_def(
arch_def,
depth_multiplier=depth_multiplier,
fix_first_last=fix_stem_head,
group_size=group_size,
),
num_features=1280 if fix_stem_head else max(1280, round_chs_fn(1280)), num_features=1280 if fix_stem_head else max(1280, round_chs_fn(1280)),
stem_size=32, stem_size=32,
fix_stem=fix_stem_head, fix_stem=fix_stem_head,
@ -764,7 +776,7 @@ def _gen_efficientnet_lite(variant, channel_multiplier=1.0, depth_multiplier=1.0
def _gen_efficientnetv2_base( def _gen_efficientnetv2_base(
variant, channel_multiplier=1.0, depth_multiplier=1.0, pretrained=False, **kwargs): variant, channel_multiplier=1.0, depth_multiplier=1.0, group_size=None, pretrained=False, **kwargs):
""" Creates an EfficientNet-V2 base model """ Creates an EfficientNet-V2 base model
Ref impl: https://github.com/google/automl/tree/master/efficientnetv2 Ref impl: https://github.com/google/automl/tree/master/efficientnetv2
@ -780,7 +792,7 @@ def _gen_efficientnetv2_base(
] ]
round_chs_fn = partial(round_channels, multiplier=channel_multiplier, round_limit=0.) round_chs_fn = partial(round_channels, multiplier=channel_multiplier, round_limit=0.)
model_kwargs = dict( model_kwargs = dict(
block_args=decode_arch_def(arch_def, depth_multiplier), block_args=decode_arch_def(arch_def, depth_multiplier, group_size=group_size),
num_features=round_chs_fn(1280), num_features=round_chs_fn(1280),
stem_size=32, stem_size=32,
round_chs_fn=round_chs_fn, round_chs_fn=round_chs_fn,
@ -831,7 +843,8 @@ def _gen_efficientnetv2_s(
return model return model
def _gen_efficientnetv2_m(variant, channel_multiplier=1.0, depth_multiplier=1.0, pretrained=False, **kwargs): def _gen_efficientnetv2_m(
variant, channel_multiplier=1.0, depth_multiplier=1.0, group_size=None, pretrained=False, **kwargs):
""" Creates an EfficientNet-V2 Medium model """ Creates an EfficientNet-V2 Medium model
Ref impl: https://github.com/google/automl/tree/master/efficientnetv2 Ref impl: https://github.com/google/automl/tree/master/efficientnetv2
@ -849,7 +862,7 @@ def _gen_efficientnetv2_m(variant, channel_multiplier=1.0, depth_multiplier=1.0,
] ]
model_kwargs = dict( model_kwargs = dict(
block_args=decode_arch_def(arch_def, depth_multiplier), block_args=decode_arch_def(arch_def, depth_multiplier, group_size=group_size),
num_features=1280, num_features=1280,
stem_size=24, stem_size=24,
round_chs_fn=partial(round_channels, multiplier=channel_multiplier), round_chs_fn=partial(round_channels, multiplier=channel_multiplier),
@ -861,7 +874,8 @@ def _gen_efficientnetv2_m(variant, channel_multiplier=1.0, depth_multiplier=1.0,
return model return model
def _gen_efficientnetv2_l(variant, channel_multiplier=1.0, depth_multiplier=1.0, pretrained=False, **kwargs): def _gen_efficientnetv2_l(
variant, channel_multiplier=1.0, depth_multiplier=1.0, group_size=None, pretrained=False, **kwargs):
""" Creates an EfficientNet-V2 Large model """ Creates an EfficientNet-V2 Large model
Ref impl: https://github.com/google/automl/tree/master/efficientnetv2 Ref impl: https://github.com/google/automl/tree/master/efficientnetv2
@ -879,7 +893,7 @@ def _gen_efficientnetv2_l(variant, channel_multiplier=1.0, depth_multiplier=1.0,
] ]
model_kwargs = dict( model_kwargs = dict(
block_args=decode_arch_def(arch_def, depth_multiplier), block_args=decode_arch_def(arch_def, depth_multiplier, group_size=group_size),
num_features=1280, num_features=1280,
stem_size=32, stem_size=32,
round_chs_fn=partial(round_channels, multiplier=channel_multiplier), round_chs_fn=partial(round_channels, multiplier=channel_multiplier),
@ -891,7 +905,8 @@ def _gen_efficientnetv2_l(variant, channel_multiplier=1.0, depth_multiplier=1.0,
return model return model
def _gen_efficientnetv2_xl(variant, channel_multiplier=1.0, depth_multiplier=1.0, pretrained=False, **kwargs): def _gen_efficientnetv2_xl(
variant, channel_multiplier=1.0, depth_multiplier=1.0, group_size=None, pretrained=False, **kwargs):
""" Creates an EfficientNet-V2 Xtra-Large model """ Creates an EfficientNet-V2 Xtra-Large model
Ref impl: https://github.com/google/automl/tree/master/efficientnetv2 Ref impl: https://github.com/google/automl/tree/master/efficientnetv2
@ -909,7 +924,7 @@ def _gen_efficientnetv2_xl(variant, channel_multiplier=1.0, depth_multiplier=1.0
] ]
model_kwargs = dict( model_kwargs = dict(
block_args=decode_arch_def(arch_def, depth_multiplier), block_args=decode_arch_def(arch_def, depth_multiplier, group_size=group_size),
num_features=1280, num_features=1280,
stem_size=32, stem_size=32,
round_chs_fn=partial(round_channels, multiplier=channel_multiplier), round_chs_fn=partial(round_channels, multiplier=channel_multiplier),
@ -1094,7 +1109,8 @@ def _gen_tinynet(
return model return model
def _gen_mobilenet_edgetpu(variant, channel_multiplier=1.0, depth_multiplier=1.0, pretrained=False, **kwargs): def _gen_mobilenet_edgetpu(
variant, channel_multiplier=1.0, depth_multiplier=1.0, group_size=None, pretrained=False, **kwargs):
""" """
Based on definitions in: https://github.com/tensorflow/models/tree/d2427a562f401c9af118e47af2f030a0a5599f55/official/projects/edgetpu/vision Based on definitions in: https://github.com/tensorflow/models/tree/d2427a562f401c9af118e47af2f030a0a5599f55/official/projects/edgetpu/vision
""" """
@ -1170,7 +1186,7 @@ def _gen_mobilenet_edgetpu(variant, channel_multiplier=1.0, depth_multiplier=1.0
] ]
model_kwargs = dict( model_kwargs = dict(
block_args=decode_arch_def(arch_def, depth_multiplier), block_args=decode_arch_def(arch_def, depth_multiplier, group_size=group_size),
num_features=num_features, num_features=num_features,
stem_size=stem_size, stem_size=stem_size,
stem_kernel_size=stem_kernel_size, stem_kernel_size=stem_kernel_size,

12
timm/models/mobilenetv3.py
View File

@ -450,7 +450,9 @@ def _gen_mobilenet_v3_rw(variant: str, channel_multiplier: float = 1.0, pretrain
return model return model
def _gen_mobilenet_v3(variant: str, channel_multiplier: float = 1.0, pretrained: bool = False, **kwargs) -> MobileNetV3: def _gen_mobilenet_v3(
variant: str, channel_multiplier: float = 1.0, group_size=None, pretrained: bool = False, **kwargs
) -> MobileNetV3:
"""Creates a MobileNet-V3 model. """Creates a MobileNet-V3 model.
Ref impl: ? Ref impl: ?
@ -533,7 +535,7 @@ def _gen_mobilenet_v3(variant: str, channel_multiplier: float = 1.0, pretrained:
] ]
se_layer = partial(SqueezeExcite, gate_layer='hard_sigmoid', force_act_layer=nn.ReLU, rd_round_fn=round_channels) se_layer = partial(SqueezeExcite, gate_layer='hard_sigmoid', force_act_layer=nn.ReLU, rd_round_fn=round_channels)
model_kwargs = dict( model_kwargs = dict(
block_args=decode_arch_def(arch_def), block_args=decode_arch_def(arch_def, group_size=group_size),
num_features=num_features, num_features=num_features,
stem_size=16, stem_size=16,
fix_stem=channel_multiplier < 0.75, fix_stem=channel_multiplier < 0.75,
@ -646,7 +648,9 @@ def _gen_lcnet(variant: str, channel_multiplier: float = 1.0, pretrained: bool =
return model return model
def _gen_mobilenet_v4(variant: str, channel_multiplier: float = 1.0, pretrained: bool = False, **kwargs) -> MobileNetV3: def _gen_mobilenet_v4(
variant: str, channel_multiplier: float = 1.0, group_size=None, pretrained: bool = False, **kwargs,
) -> MobileNetV3:
"""Creates a MobileNet-V4 model. """Creates a MobileNet-V4 model.
Ref impl: ? Ref impl: ?
@ -877,7 +881,7 @@ def _gen_mobilenet_v4(variant: str, channel_multiplier: float = 1.0, pretrained:
assert False, f'Unknown variant {variant}.' assert False, f'Unknown variant {variant}.'
model_kwargs = dict( model_kwargs = dict(
block_args=decode_arch_def(arch_def), block_args=decode_arch_def(arch_def, group_size=group_size),
head_bias=False, head_bias=False,
head_norm=True, head_norm=True,
num_features=num_features, num_features=num_features,