Add Uniformer (segmentation) models

Derived from https://github.com/Sense-X/UniFormer/

uniformer_base/README.md

@@ -0,0 +1 @@
+Originally from https://github.com/Sense-X/UniFormer, converted to safetensors and flattend the config for inference, used under Apache-2.0

uniformer_base/test_config_g.py

@@ -0,0 +1,77 @@
+norm_cfg = dict(type='SyncBN', requires_grad=True)
+model = dict(
+    type='EncoderDecoder',
+    pretrained=None,
+    backbone=dict(
+        type='UniFormer',
+        embed_dim=[64, 128, 320, 512],
+        layers=[5, 8, 20, 7],
+        head_dim=64,
+        mlp_ratio=4.,
+        qkv_bias=True,
+        drop_rate=0.,
+        attn_drop_rate=0.,
+        drop_path_rate=0.4,
+        windows=False,
+        hybrid=False),
+    decode_head=dict(
+        type='UPerHead',
+        in_channels=[64, 128, 320, 512],
+        in_index=[0, 1, 2, 3],
+        pool_scales=(1, 2, 3, 6),
+        channels=512,
+        dropout_ratio=0.1,
+        num_classes=150,
+        norm_cfg=norm_cfg,
+        align_corners=False,
+        loss_decode=dict(
+            type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0)),
+    auxiliary_head=dict(
+        type='FCNHead',
+        in_channels=320,
+        in_index=2,
+        channels=256,
+        num_convs=1,
+        concat_input=False,
+        dropout_ratio=0.1,
+        num_classes=150,
+        norm_cfg=norm_cfg,
+        align_corners=False,
+        loss_decode=dict(
+            type='CrossEntropyLoss', use_sigmoid=False, loss_weight=0.4)),
+    # model training and testing settings
+    train_cfg=dict(),
+    test_cfg=dict(mode='whole'))
+
+# dataset settings
+dataset_type = 'ADE20KDataset'
+data_root = 'data/ade/ADEChallengeData2016'
+img_norm_cfg = dict(
+    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+crop_size = (512, 512)
+
+test_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(
+        type='MultiScaleFlipAug',
+        img_scale=(2048, 512),
+        # img_ratios=[0.5, 0.75, 1.0, 1.25, 1.5, 1.75],
+        flip=False,
+        transforms=[
+            dict(type='Resize', keep_ratio=True),
+            dict(type='RandomFlip'),
+            dict(type='Normalize', **img_norm_cfg),
+            dict(type='ImageToTensor', keys=['img']),
+            dict(type='Collect', keys=['img']),
+        ])
+]
+
+data = dict(
+    samples_per_gpu=2,
+    workers_per_gpu=4,
+    test=dict(
+        type=dataset_type,
+        data_root=data_root,
+        img_dir='images/validation',
+        ann_dir='annotations/validation',
+        pipeline=test_pipeline))

uniformer_base/upernet_global_base.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b29aa3cb2a5765d61a56b573f36e2fbd59ac24cc78e1662b49ce898bd5e40758
+size 319165833

uniformer_small/README.md

@@ -0,0 +1 @@
+Originally from https://github.com/Sense-X/UniFormer, converted to safetensors and flattend the config for inference, used under Apache-2.0

uniformer_small/test_config_g.py

@@ -0,0 +1,78 @@
+# model settings
+norm_cfg = dict(type='SyncBN', requires_grad=True)
+model = dict(
+    type='EncoderDecoder',
+    pretrained=None,
+    backbone=dict(
+        type='UniFormer',
+        embed_dim=[64, 128, 320, 512],
+        layers=[3, 4, 8, 3],
+        head_dim=64,
+        mlp_ratio=4.,
+        qkv_bias=True,
+        drop_rate=0.,
+        attn_drop_rate=0.,
+        drop_path_rate=0.25,
+        windows=False,
+        hybrid=False),
+    decode_head=dict(
+        type='UPerHead',
+        in_channels=[64, 128, 320, 512],
+        in_index=[0, 1, 2, 3],
+        pool_scales=(1, 2, 3, 6),
+        channels=512,
+        dropout_ratio=0.1,
+        num_classes=150,
+        norm_cfg=norm_cfg,
+        align_corners=False,
+        loss_decode=dict(
+            type='CrossEntropyLoss', use_sigmoid=False, loss_weight=1.0)),
+    auxiliary_head=dict(
+        type='FCNHead',
+        in_channels=320,
+        in_index=2,
+        channels=256,
+        num_convs=1,
+        concat_input=False,
+        dropout_ratio=0.1,
+        num_classes=150,
+        norm_cfg=norm_cfg,
+        align_corners=False,
+        loss_decode=dict(
+            type='CrossEntropyLoss', use_sigmoid=False, loss_weight=0.4)),
+    # model training and testing settings
+    train_cfg=dict(),
+    test_cfg=dict(mode='whole'))
+
+# dataset settings
+dataset_type = 'ADE20KDataset'
+data_root = 'data/ade/ADEChallengeData2016'
+img_norm_cfg = dict(
+    mean=[123.675, 116.28, 103.53], std=[58.395, 57.12, 57.375], to_rgb=True)
+crop_size = (512, 512)
+
+test_pipeline = [
+    dict(type='LoadImageFromFile'),
+    dict(
+        type='MultiScaleFlipAug',
+        img_scale=(2048, 512),
+        # img_ratios=[0.5, 0.75, 1.0, 1.25, 1.5, 1.75],
+        flip=False,
+        transforms=[
+            dict(type='Resize', keep_ratio=True),
+            dict(type='RandomFlip'),
+            dict(type='Normalize', **img_norm_cfg),
+            dict(type='ImageToTensor', keys=['img']),
+            dict(type='Collect', keys=['img']),
+        ])
+]
+
+data = dict(
+    samples_per_gpu=2,
+    workers_per_gpu=4,
+    test=dict(
+        type=dataset_type,
+        data_root=data_root,
+        img_dir='images/validation',
+        ann_dir='annotations/validation',
+        pipeline=test_pipeline))

uniformer_small/upernet_global_small.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bae1398435fd2ef9da3bdec5031211cb71b515a97651ccfbac83987f6f575d5d
+size 206193755