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

Fix some mistakes in mnv4 model defs

Browse Source
This commit is contained in:
Ross Wightman 2024-05-23 14:24:32 -07:00
parent 70176a2dae
commit cb33956b20
1 changed files with 24 additions and 13 deletions

37
timm/models/mobilenetv3.py

@ -640,7 +640,10 @@ def _gen_mobilenet_v4(variant: str, channel_multiplier: float = 1.0, pretrained:
# stage 0, 112x112 in # stage 0, 112x112 in
['er_r1_k3_s2_e4_c48'], ['er_r1_k3_s2_e4_c48'],
# stage 1, 56x56 in # stage 1, 56x56 in
['uir_r1_a3_k5_s2_e4_c80', 'uir_r1_a3_k3_s1_e2_c80'], [
'uir_r1_a3_k5_s2_e4_c80',
'uir_r1_a3_k3_s1_e2_c80',
],
# stage 2, 28x28 in # stage 2, 28x28 in
[ [
'uir_r1_a3_k5_s2_e6_c160', 'uir_r1_a3_k5_s2_e6_c160',
@ -685,7 +688,10 @@ def _gen_mobilenet_v4(variant: str, channel_multiplier: float = 1.0, pretrained:
# stage 0, 112x112 in # stage 0, 112x112 in
['er_r1_k3_s2_e4_c48'], ['er_r1_k3_s2_e4_c48'],
# stage 1, 56x56 in # stage 1, 56x56 in
['uir_r1_a3_k5_s2_e4_c96', 'uir_r1_a3_k3_s1_e4_c96'], [
'uir_r1_a3_k5_s2_e4_c96',
'uir_r1_a3_k3_s1_e4_c96',
],
# stage 2, 28x28 in # stage 2, 28x28 in
[ [
'uir_r1_a3_k5_s2_e4_c192', 'uir_r1_a3_k5_s2_e4_c192',
@ -710,13 +716,13 @@ def _gen_mobilenet_v4(variant: str, channel_multiplier: float = 1.0, pretrained:
'uir_r1_a5_k3_s1_e4_c512', 'uir_r1_a5_k3_s1_e4_c512',
'uir_r1_a5_k5_s1_e4_c512', 'uir_r1_a5_k5_s1_e4_c512',
'mqa_r1_k3_h8_s1_d64_c512', 'mqa_r1_k3_h8_s1_d64_c512',
'uir_r3_a5_k0_s1_e4_c512', # convnext 'uir_r1_a5_k0_s1_e4_c512', # convnext
'mqa_r1_k3_h8_s1_d64_c512', 'mqa_r1_k3_h8_s1_d64_c512',
'uir_r3_a5_k0_s1_e4_c512', # convnext 'uir_r1_a5_k0_s1_e4_c512', # convnext
'mqa_r1_k3_h8_s1_d64_c512', 'mqa_r1_k3_h8_s1_d64_c512',
'uir_r3_a5_k0_s1_e4_c512', # convnext 'uir_r1_a5_k0_s1_e4_c512', # convnext
'mqa_r1_k3_h8_s1_d64_c512', 'mqa_r1_k3_h8_s1_d64_c512',
'uir_r3_a5_k0_s1_e4_c512', # convnext 'uir_r1_a5_k0_s1_e4_c512', # convnext
], ],
# stage 4, 7x7 in # stage 4, 7x7 in
['cn_r1_k1_s1_c960'], ['cn_r1_k1_s1_c960'],
@ -758,15 +764,18 @@ def _gen_mobilenet_v4(variant: str, channel_multiplier: float = 1.0, pretrained:
# stage 0, 112x112 in # stage 0, 112x112 in
['er_r1_k3_s2_e4_c48'], ['er_r1_k3_s2_e4_c48'],
# stage 1, 56x56 in # stage 1, 56x56 in
['uir_r1_a3_k5_s2_e4_c80', 'uir_r1_a3_k3_s1_e2_c80'], [
'uir_r1_a3_k5_s2_e4_c80',
'uir_r1_a3_k3_s1_e2_c80',
],
# stage 2, 28x28 in # stage 2, 28x28 in
[ [
'uir_r1_a5_k3_s2_e6_c160', 'uir_r1_a3_k5_s2_e6_c160',
'uir_r2_a3_k3_s1_e4_c160', 'uir_r2_a3_k3_s1_e4_c160',
'uir_r1_a3_k3_s1_e4_c160', 'uir_r1_a3_k5_s1_e4_c160',
'uir_r1_a3_k3_s1_e4_c160', 'uir_r1_a3_k3_s1_e4_c160',
'uir_r1_a3_k0_s1_e4_c160', # convnext 'uir_r1_a3_k0_s1_e4_c160', # convnext
'uir_r2_a0_k0_s1_e2_c160', 'uir_r1_a0_k0_s1_e2_c160',
'uir_r1_a3_k0_s1_e4_c160', # convnext 'uir_r1_a3_k0_s1_e4_c160', # convnext
], ],
# stage 3, 14x14in # stage 3, 14x14in
@ -776,7 +785,6 @@ def _gen_mobilenet_v4(variant: str, channel_multiplier: float = 1.0, pretrained:
'uir_r2_a3_k5_s1_e4_c256', 'uir_r2_a3_k5_s1_e4_c256',
'uir_r1_a0_k0_s1_e4_c256', 'uir_r1_a0_k0_s1_e4_c256',
'uir_r1_a3_k0_s1_e4_c256', # convnext 'uir_r1_a3_k0_s1_e4_c256', # convnext
'uir_r1_a3_k0_s1_e4_c256', # convnext
'uir_r1_a3_k5_s1_e2_c256', 'uir_r1_a3_k5_s1_e2_c256',
'uir_r1_a5_k5_s1_e4_c256', 'uir_r1_a5_k5_s1_e4_c256',
'uir_r2_a0_k0_s1_e4_c256', 'uir_r2_a0_k0_s1_e4_c256',
@ -793,7 +801,10 @@ def _gen_mobilenet_v4(variant: str, channel_multiplier: float = 1.0, pretrained:
# stage 0, 112x112 in # stage 0, 112x112 in
['er_r1_k3_s2_e4_c48'], ['er_r1_k3_s2_e4_c48'],
# stage 1, 56x56 in # stage 1, 56x56 in
['uir_r1_a3_k5_s2_e4_c96', 'uir_r1_a3_k3_s1_e4_c96'], [
'uir_r1_a3_k5_s2_e4_c96',
'uir_r1_a3_k3_s1_e4_c96',
],
# stage 2, 28x28 in # stage 2, 28x28 in
[ [
'uir_r1_a3_k5_s2_e4_c192', 'uir_r1_a3_k5_s2_e4_c192',
@ -986,7 +997,7 @@ def mobilenetv3_large_100(pretrained: bool = False, **kwargs) -> MobileNetV3:
@register_model @register_model
def mobilenetv3_large_150(pretrained: bool = False, **kwargs) -> MobileNetV3: def mobilenetv3_large_150(pretrained: bool = False, **kwargs) -> MobileNetV3:
""" MobileNet V3 """ """ MobileNet V3 """
model = _gen_mobilenet_v3('mobilenetv3_large_100', 1.5, pretrained=pretrained, **kwargs) model = _gen_mobilenet_v3('mobilenetv3_large_150', 1.5, pretrained=pretrained, **kwargs)
return model return model