Upload folder using huggingface_hub

.gitattributes

@@ -47,3 +47,4 @@ scannet200/distill-ptv3_scannet200_dino-L/train.log filter=lfs diff=lfs merge=lf
 scannet200/semseg-ptv3_weight=distill-ptv3_scannet200+structured3d_dino-L/train.log filter=lfs diff=lfs merge=lfs -text
 nuscenes/semseg-ptv3_weight=distill-ptv3_nuscenes_dino-L/test.log filter=lfs diff=lfs merge=lfs -text
 nuscenes/semseg-ptv3_weight=distill-ptv3_nuscenes_dino-L/train.log filter=lfs diff=lfs merge=lfs -text
+nuscenes/distill-ptv3_nuscenes_dino-L/train.log filter=lfs diff=lfs merge=lfs -text

nuscenes/distill-ptv3_nuscenes_dino-L/config.py

@@ -0,0 +1,249 @@
+weight = None
+resume = False
+evaluate = True
+test_only = False
+seed = 14047740
+save_path = 'exp/nuscenes/distill-ptv3dino-Llevels=0'
+wandb_project = 'pointcept_distill_nuscenes'
+num_worker = 16
+batch_size = 12
+batch_size_val = None
+batch_size_test = None
+epoch = 50
+eval_epoch = 50
+clip_grad = None
+sync_bn = False
+enable_amp = True
+empty_cache = False
+empty_cache_per_epoch = False
+find_unused_parameters = False
+mix_prob = 0.8
+param_dicts = [dict(keyword='^(?!.*img_enc).*block', lr=0.0002)]
+hooks = [
+    dict(type='CheckpointLoader'),
+    dict(type='IterationTimer', warmup_iter=2),
+    dict(type='InformationWriter'),
+    dict(type='DistillEvaluator'),
+    dict(type='CheckpointSaver', save_freq=None)
+]
+train = dict(type='DefaultTrainer')
+test = dict()
+model = dict(
+    type='DefaultDistiller',
+    backbone_out_channels=(64, 64, 128, 256, 512),
+    levels=(0, ),
+    image_feature_channels=(1024, 1024, 1024, 1024, 1024),
+    backbone=dict(
+        type='PT-v3m1-distill',
+        in_channels=4,
+        order=['z', 'z-trans', 'hilbert', 'hilbert-trans'],
+        stride=(2, 2, 2, 2),
+        enc_depths=(2, 2, 2, 6, 2),
+        enc_channels=(32, 64, 128, 256, 512),
+        enc_num_head=(2, 4, 8, 16, 32),
+        enc_patch_size=(1024, 1024, 1024, 1024, 1024),
+        dec_depths=(2, 2, 2, 2),
+        dec_channels=(64, 64, 128, 256),
+        dec_num_head=(4, 4, 8, 16),
+        dec_patch_size=(1024, 1024, 1024, 1024),
+        mlp_ratio=4,
+        qkv_bias=True,
+        qk_scale=None,
+        attn_drop=0.0,
+        proj_drop=0.0,
+        drop_path=0.3,
+        shuffle_orders=True,
+        pre_norm=True,
+        enable_rpe=False,
+        enable_flash=True,
+        upcast_attention=False,
+        upcast_softmax=False,
+        cls_mode=False,
+        pdnorm_bn=False,
+        pdnorm_ln=False,
+        pdnorm_decouple=True,
+        pdnorm_adaptive=False,
+        pdnorm_affine=True,
+        pdnorm_conditions=('nuScenes', 'SemanticKITTI', 'Waymo'),
+        dinov2='large'),
+    criteria=[dict(type='MSELoss', loss_weight=1.0)])
+optimizer = dict(type='AdamW', lr=0.002, weight_decay=0.005)
+scheduler = dict(
+    type='OneCycleLR',
+    max_lr=[0.002, 0.0002],
+    pct_start=0.04,
+    anneal_strategy='cos',
+    div_factor=10.0,
+    final_div_factor=100.0)
+dataset_type = 'NuScenesDataset'
+data_root = 'data/nuscenes'
+ignore_index = -1
+names = [
+    'barrier', 'bicycle', 'bus', 'car', 'construction_vehicle', 'motorcycle',
+    'pedestrian', 'traffic_cone', 'trailer', 'truck', 'driveable_surface',
+    'other_flat', 'sidewalk', 'terrain', 'manmade', 'vegetation'
+]
+data = dict(
+    num_classes=16,
+    ignore_index=-1,
+    names=[
+        'barrier', 'bicycle', 'bus', 'car', 'construction_vehicle',
+        'motorcycle', 'pedestrian', 'traffic_cone', 'trailer', 'truck',
+        'driveable_surface', 'other_flat', 'sidewalk', 'terrain', 'manmade',
+        'vegetation'
+    ],
+    train=dict(
+        type='NuScenesDataset',
+        split='train',
+        data_root='data/nuscenes',
+        with_images=True,
+        transform=[
+            dict(type='ImageResize', size=[378, 672]),
+            dict(
+                type='ImageColorJitter',
+                brightness=0.4,
+                contrast=0.4,
+                saturation=0.2,
+                hue=0.1),
+            dict(type='ImageRandomHorizontalFlip'),
+            dict(type='ImageNormalize'),
+            dict(
+                type='RandomRotate',
+                angle=[-1, 1],
+                axis='z',
+                center=[0, 0, 0],
+                p=0.5),
+            dict(type='RandomScale', scale=[0.9, 1.1]),
+            dict(type='RandomFlip', p=0.5),
+            dict(type='RandomJitter', sigma=0.005, clip=0.02),
+            dict(
+                type='GridSample',
+                grid_size=0.05,
+                hash_type='fnv',
+                mode='train',
+                keys=('coord', 'strength', 'segment', 'image_coord',
+                      'image_mask'),
+                return_grid_coord=True),
+            dict(type='ToTensor'),
+            dict(
+                type='Collect',
+                keys=('coord', 'grid_coord', 'segment', 'image', 'image_coord',
+                      'image_mask'),
+                feat_keys=('coord', 'strength'))
+        ],
+        test_mode=False,
+        ignore_index=-1,
+        loop=1),
+    val=dict(
+        type='NuScenesDataset',
+        split='val',
+        data_root='data/nuscenes',
+        with_images=True,
+        transform=[
+            dict(type='ImageResize', size=[378, 672]),
+            dict(type='ImageNormalize'),
+            dict(
+                type='GridSample',
+                grid_size=0.05,
+                hash_type='fnv',
+                mode='train',
+                keys=('coord', 'strength', 'segment', 'image_coord',
+                      'image_mask'),
+                return_grid_coord=True),
+            dict(type='ToTensor'),
+            dict(
+                type='Collect',
+                keys=('coord', 'grid_coord', 'segment', 'image', 'image_coord',
+                      'image_mask'),
+                feat_keys=('coord', 'strength'))
+        ],
+        test_mode=False,
+        ignore_index=-1),
+    test=dict(
+        type='NuScenesDataset',
+        split='val',
+        data_root='data/nuscenes',
+        with_images=True,
+        transform=[
+            dict(type='ImageResize', size=[378, 672]),
+            dict(type='Copy', keys_dict=dict(segment='origin_segment')),
+            dict(
+                type='GridSample',
+                grid_size=0.025,
+                hash_type='fnv',
+                mode='train',
+                keys=('coord', 'strength', 'segment', 'image_coord',
+                      'image_mask'),
+                return_inverse=True)
+        ],
+        test_mode=True,
+        test_cfg=dict(
+            voxelize=dict(
+                type='GridSample',
+                grid_size=0.05,
+                hash_type='fnv',
+                mode='test',
+                return_grid_coord=True,
+                keys=('coord', 'strength', 'image_coord', 'image_mask')),
+            crop=None,
+            post_transform=[
+                dict(type='ImageNormalize'),
+                dict(type='ToTensor'),
+                dict(
+                    type='Collect',
+                    keys=('coord', 'grid_coord', 'index', 'image',
+                          'image_coord', 'image_mask'),
+                    feat_keys=('coord', 'strength'))
+            ],
+            aug_transform=[[{
+                'type': 'RandomScale',
+                'scale': [0.9, 0.9]
+            }], [{
+                'type': 'RandomScale',
+                'scale': [0.95, 0.95]
+            }], [{
+                'type': 'RandomScale',
+                'scale': [1, 1]
+            }], [{
+                'type': 'RandomScale',
+                'scale': [1.05, 1.05]
+            }], [{
+                'type': 'RandomScale',
+                'scale': [1.1, 1.1]
+            }],
+                           [{
+                               'type': 'RandomScale',
+                               'scale': [0.9, 0.9]
+                           }, {
+                               'type': 'RandomFlip',
+                               'p': 1
+                           }],
+                           [{
+                               'type': 'RandomScale',
+                               'scale': [0.95, 0.95]
+                           }, {
+                               'type': 'RandomFlip',
+                               'p': 1
+                           }],
+                           [{
+                               'type': 'RandomScale',
+                               'scale': [1, 1]
+                           }, {
+                               'type': 'RandomFlip',
+                               'p': 1
+                           }],
+                           [{
+                               'type': 'RandomScale',
+                               'scale': [1.05, 1.05]
+                           }, {
+                               'type': 'RandomFlip',
+                               'p': 1
+                           }],
+                           [{
+                               'type': 'RandomScale',
+                               'scale': [1.1, 1.1]
+                           }, {
+                               'type': 'RandomFlip',
+                               'p': 1
+                           }]]),
+        ignore_index=-1))

nuscenes/distill-ptv3_nuscenes_dino-L/events.out.tfevents.1729984113.n23g0002.hpc.itc.rwth-aachen.de

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:ecc40c1006dc0ab3c779b4ea965e14b109bd4454118221aa601950972622be65
+size 18709910

nuscenes/distill-ptv3_nuscenes_dino-L/model/model_best.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bcffbc01f76d1ff6772d10c19c00edeaa4f5a5aa304b7f28ecf121650bf657a7
+size 555305064

nuscenes/distill-ptv3_nuscenes_dino-L/model/model_last.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:bcffbc01f76d1ff6772d10c19c00edeaa4f5a5aa304b7f28ecf121650bf657a7
+size 555305064

nuscenes/distill-ptv3_nuscenes_dino-L/train.log

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c673338ccf14ecc86b0e76af9bc5c48685b527b312ccae8c4587dab2c89e72f8
+size 34723713