Missing __init__.py and relative import

main.py

@@ -12,15 +12,7 @@
 from pytorch_lightning.utilities.distributed import rank_zero_only
 
 from taming.data.utils import custom_collate
-
-
-def get_obj_from_str(string, reload=False):
-    module, cls = string.rsplit(".", 1)
-    if reload:
-        module_imp = importlib.import_module(module)
-        importlib.reload(module_imp)
-    return getattr(importlib.import_module(module, package=None), cls)
-
+from taming.util import get_obj_from_str, instantiate_from_config
 
 def get_parser(**parser_kwargs):
     def str2bool(v):
@@ -113,10 +105,7 @@ def nondefault_trainer_args(opt):
     return sorted(k for k in vars(args) if getattr(opt, k) != getattr(args, k))
 
 
-def instantiate_from_config(config):
-    if not "target" in config:
-        raise KeyError("Expected key `target` to instantiate.")
-    return get_obj_from_str(config["target"])(**config.get("params", dict()))
+
 
 
 class WrappedDataset(Dataset):

scripts/make_samples.py

@@ -3,11 +3,13 @@
 import numpy as np
 from omegaconf import OmegaConf
 from PIL import Image
-from main import instantiate_from_config, DataModuleFromConfig
+from main import DataModuleFromConfig
 from torch.utils.data import DataLoader
 from torch.utils.data.dataloader import default_collate
 from tqdm import trange
 
+from taming.util import instantiate_from_config
+
 
 def save_image(x, path):
     c,h,w = x.shape

scripts/sample_conditional.py

@@ -5,10 +5,12 @@
 import streamlit as st
 from streamlit import caching
 from PIL import Image
-from main import instantiate_from_config, DataModuleFromConfig
+from main import DataModuleFromConfig
 from torch.utils.data import DataLoader
 from torch.utils.data.dataloader import default_collate
 
+from taming.util import instantiate_from_config
+
 
 rescale = lambda x: (x + 1.) / 2.
 

scripts/sample_fast.py

@@ -7,9 +7,8 @@
 from tqdm import tqdm, trange
 from einops import repeat
 
-from main import instantiate_from_config
 from taming.modules.transformer.mingpt import sample_with_past
-
+from taming.util import instantiate_from_config
 
 rescale = lambda x: (x + 1.) / 2.
 

taming/data/ade20k.py

@@ -5,7 +5,7 @@
 from PIL import Image
 from torch.utils.data import Dataset
 
-from taming.data.sflckr import SegmentationBase # for examples included in repo
+from .sflckr import SegmentationBase # for examples included in repo
 
 
 class Examples(SegmentationBase):

taming/data/annotated_objects_coco.py

@@ -6,8 +6,8 @@
 
 from tqdm import tqdm
 
-from taming.data.annotated_objects_dataset import AnnotatedObjectsDataset
-from taming.data.helper_types import Annotation, ImageDescription, Category
+from .annotated_objects_dataset import AnnotatedObjectsDataset
+from .helper_types import Annotation, ImageDescription, Category
 
 COCO_PATH_STRUCTURE = {
     'train': {

taming/data/annotated_objects_dataset.py

@@ -7,11 +7,11 @@
 from torch.utils.data import Dataset
 from torchvision import transforms
 
-from taming.data.conditional_builder.objects_bbox import ObjectsBoundingBoxConditionalBuilder
-from taming.data.conditional_builder.objects_center_points import ObjectsCenterPointsConditionalBuilder
-from taming.data.conditional_builder.utils import load_object_from_string
-from taming.data.helper_types import BoundingBox, CropMethodType, Image, Annotation, SplitType
-from taming.data.image_transforms import CenterCropReturnCoordinates, RandomCrop1dReturnCoordinates, \
+from .conditional_builder.objects_bbox import ObjectsBoundingBoxConditionalBuilder
+from .conditional_builder.objects_center_points import ObjectsCenterPointsConditionalBuilder
+from .conditional_builder.utils import load_object_from_string
+from .helper_types import BoundingBox, CropMethodType, Image, Annotation, SplitType
+from .image_transforms import CenterCropReturnCoordinates, RandomCrop1dReturnCoordinates, \
     Random2dCropReturnCoordinates, RandomHorizontalFlipReturn, convert_pil_to_tensor
 
 

taming/data/annotated_objects_open_images.py

@@ -4,10 +4,11 @@
 from typing import Dict, List, Any
 import warnings
 
-from taming.data.annotated_objects_dataset import AnnotatedObjectsDataset
-from taming.data.helper_types import Annotation, Category
 from tqdm import tqdm
 
+from .annotated_objects_dataset import AnnotatedObjectsDataset
+from .helper_types import Annotation, Category
+
 OPEN_IMAGES_STRUCTURE = {
     'train': {
         'top_level': '',

taming/data/coco.py

@@ -6,7 +6,7 @@
 from tqdm import tqdm
 from torch.utils.data import Dataset
 
-from taming.data.sflckr import SegmentationBase # for examples included in repo
+from .sflckr import SegmentationBase # for examples included in repo
 
 
 class Examples(SegmentationBase):

taming/data/conditional_builder/objects_bbox.py

@@ -3,12 +3,12 @@
 
 from PIL import Image as pil_image, ImageDraw as pil_img_draw, ImageFont
 from more_itertools.recipes import grouper
-from taming.data.image_transforms import convert_pil_to_tensor
 from torch import LongTensor, Tensor
 
-from taming.data.helper_types import BoundingBox, Annotation
-from taming.data.conditional_builder.objects_center_points import ObjectsCenterPointsConditionalBuilder
-from taming.data.conditional_builder.utils import COLOR_PALETTE, WHITE, GRAY_75, BLACK, additional_parameters_string, \
+from ..image_transforms import convert_pil_to_tensor
+from ..helper_types import BoundingBox, Annotation
+from ..conditional_builder.objects_center_points import ObjectsCenterPointsConditionalBuilder
+from ..conditional_builder.utils import COLOR_PALETTE, WHITE, GRAY_75, BLACK, additional_parameters_string, \
     pad_list, get_plot_font_size, absolute_bbox
 
 

taming/data/conditional_builder/objects_center_points.py

@@ -6,11 +6,11 @@
 
 from PIL import Image as pil_image, ImageDraw as pil_img_draw, ImageFont
 from more_itertools.recipes import grouper
-from taming.data.conditional_builder.utils import COLOR_PALETTE, WHITE, GRAY_75, BLACK, FULL_CROP, filter_annotations, \
+from ..conditional_builder.utils import COLOR_PALETTE, WHITE, GRAY_75, BLACK, FULL_CROP, filter_annotations, \
     additional_parameters_string, horizontally_flip_bbox, pad_list, get_circle_size, get_plot_font_size, \
     absolute_bbox, rescale_annotations
-from taming.data.helper_types import BoundingBox, Annotation
-from taming.data.image_transforms import convert_pil_to_tensor
+from ..helper_types import BoundingBox, Annotation
+from ..image_transforms import convert_pil_to_tensor
 from torch import LongTensor, Tensor
 
 

taming/data/conditional_builder/utils.py

@@ -1,7 +1,7 @@
 import importlib
 from typing import List, Any, Tuple, Optional
 
-from taming.data.helper_types import BoundingBox, Annotation
+from ..helper_types import BoundingBox, Annotation
 
 # source: seaborn, color palette tab10
 COLOR_PALETTE = [(30, 118, 179), (255, 126, 13), (43, 159, 43), (213, 38, 39), (147, 102, 188),

taming/data/custom.py

@@ -3,7 +3,7 @@
 import albumentations
 from torch.utils.data import Dataset
 
-from taming.data.base import ImagePaths, NumpyPaths, ConcatDatasetWithIndex
+from .base import ImagePaths, NumpyPaths, ConcatDatasetWithIndex
 
 
 class CustomBase(Dataset):

taming/data/faceshq.py

@@ -3,7 +3,7 @@
 import albumentations
 from torch.utils.data import Dataset
 
-from taming.data.base import ImagePaths, NumpyPaths, ConcatDatasetWithIndex
+from .base import ImagePaths, NumpyPaths, ConcatDatasetWithIndex
 
 
 class FacesBase(Dataset):

taming/data/image_transforms.py

@@ -7,7 +7,7 @@
 from torchvision.transforms import RandomCrop, functional as F, CenterCrop, RandomHorizontalFlip, PILToTensor
 from torchvision.transforms.functional import _get_image_size as get_image_size
 
-from taming.data.helper_types import BoundingBox, Image
+from .helper_types import BoundingBox, Image
 
 pil_to_tensor = PILToTensor()
 

taming/data/imagenet.py

@@ -7,9 +7,9 @@
 from omegaconf import OmegaConf
 from torch.utils.data import Dataset
 
-from taming.data.base import ImagePaths
-from taming.util import download, retrieve
-import taming.data.utils as bdu
+from .base import ImagePaths
+from ..util import download, retrieve
+import .utils as bdu
 
 
 def give_synsets_from_indices(indices, path_to_yaml="data/imagenet_idx_to_synset.yaml"):

taming/data/utils.py

@@ -7,11 +7,12 @@
 
 import numpy as np
 import torch
-from taming.data.helper_types import Annotation
 from torch._six import string_classes
 from torch.utils.data._utils.collate import np_str_obj_array_pattern, default_collate_err_msg_format
 from tqdm import tqdm
 
+from .helper_types import Annotation
+
 
 def unpack(path):
     if path.endswith("tar.gz"):

taming/models/cond_transformer.py

@@ -3,8 +3,8 @@
 import torch.nn.functional as F
 import pytorch_lightning as pl
 
-from main import instantiate_from_config
-from taming.modules.util import SOSProvider
+from ..util import instantiate_from_config
+from ..modules.util import SOSProvider
 
 
 def disabled_train(self, mode=True):

taming/models/vqgan.py

@@ -2,12 +2,11 @@
 import torch.nn.functional as F
 import pytorch_lightning as pl
 
-from main import instantiate_from_config
-
-from taming.modules.diffusionmodules.model import Encoder, Decoder
-from taming.modules.vqvae.quantize import VectorQuantizer2 as VectorQuantizer
-from taming.modules.vqvae.quantize import GumbelQuantize
-from taming.modules.vqvae.quantize import EMAVectorQuantizer
+from ..util import instantiate_from_config
+from ..modules.diffusionmodules.model import Encoder, Decoder
+from ..modules.vqvae.quantize import VectorQuantizer2 as VectorQuantizer
+from ..modules.vqvae.quantize import GumbelQuantize
+from ..modules.vqvae.quantize import EMAVectorQuantizer
 
 class VQModel(pl.LightningModule):
     def __init__(self,

taming/modules/discriminator/model.py

@@ -2,7 +2,7 @@
 import torch.nn as nn
 
 
-from taming.modules.util import ActNorm
+from ...modules.util import ActNorm
 
 
 def weights_init(m):

taming/modules/losses/__init__.py

@@ -1,2 +1,2 @@
-from taming.modules.losses.vqperceptual import DummyLoss
+from .vqperceptual import DummyLoss
 

taming/modules/losses/lpips.py

@@ -5,7 +5,7 @@
 from torchvision import models
 from collections import namedtuple
 
-from taming.util import get_ckpt_path
+from ...util import get_ckpt_path
 
 
 class LPIPS(nn.Module):

taming/modules/losses/vqperceptual.py

@@ -2,8 +2,8 @@
 import torch.nn as nn
 import torch.nn.functional as F
 
-from taming.modules.losses.lpips import LPIPS
-from taming.modules.discriminator.model import NLayerDiscriminator, weights_init
+from ..losses.lpips import LPIPS
+from ..discriminator.model import NLayerDiscriminator, weights_init
 
 
 class DummyLoss(nn.Module):

taming/modules/util.py

@@ -1,7 +1,8 @@
+import importlib
+
 import torch
 import torch.nn as nn
 
-
 def count_params(model):
     total_params = sum(p.numel() for p in model.parameters())
     return total_params

taming/util.py

@@ -1,4 +1,5 @@
 import os, hashlib
+import importlib
 import requests
 from tqdm import tqdm
 
@@ -14,6 +15,17 @@
     "vgg_lpips": "d507d7349b931f0638a25a48a722f98a"
 }
 
+def instantiate_from_config(config):
+    if not "target" in config:
+        raise KeyError("Expected key `target` to instantiate.")
+    return get_obj_from_str(config["target"])(**config.get("params", dict()))
+
+def get_obj_from_str(string, reload=False):
+    module, cls = string.rsplit(".", 1)
+    if reload:
+        module_imp = importlib.import_module(module)
+        importlib.reload(module_imp)
+    return getattr(importlib.import_module(module, package=None), cls)
 
 def download(url, local_path, chunk_size=1024):
     os.makedirs(os.path.split(local_path)[0], exist_ok=True)