Initial commit

conditioning_images.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:74a6dde1abd9534efb52b289b36286cc8ad998df00a3930da2853a5087e197f0
+size 260754

images.zip

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:0aa05cfd16da3809b4f6d97a853108370f32aa27a42cb9728ecdc54c327dc4eb
+size 7920242

testdataset.py

@@ -0,0 +1,101 @@
+import pandas as pd
+from huggingface_hub import hf_hub_url
+import datasets
+import os
+
+_VERSION = datasets.Version("1.0.0")
+
+_DESCRIPTION = "TODO"
+_HOMEPAGE = "TODO"
+_LICENSE = "TODO"
+_CITATION = "TODO"
+
+_FEATURES = datasets.Features(
+    {
+        "image": datasets.Image(),
+        "conditioning_image": datasets.Image(),
+        "text": datasets.Value("string"),
+    },
+)
+
+METADATA_URL = hf_hub_url(
+    "defjam/testdataset",
+    filename="train.jsonl",
+    repo_type="dataset",
+)
+
+IMAGES_URL = hf_hub_url(
+    "defjam/testdataset",
+    filename="images.zip",
+    repo_type="dataset",
+)
+
+CONDITIONING_IMAGES_URL = hf_hub_url(
+    "defjam/testdataset",
+    filename="conditioning_images.zip",
+    repo_type="dataset",
+)
+
+_DEFAULT_CONFIG = datasets.BuilderConfig(name="default", version=_VERSION)
+
+
+class Fill50k(datasets.GeneratorBasedBuilder):
+    BUILDER_CONFIGS = [_DEFAULT_CONFIG]
+    DEFAULT_CONFIG_NAME = "default"
+
+    def _info(self):
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=_FEATURES,
+            supervised_keys=None,
+            homepage=_HOMEPAGE,
+            license=_LICENSE,
+            citation=_CITATION,
+        )
+
+    def _split_generators(self, dl_manager):
+        metadata_path = dl_manager.download(METADATA_URL)
+        images_dir = dl_manager.download_and_extract(IMAGES_URL)
+        conditioning_images_dir = dl_manager.download_and_extract(
+            CONDITIONING_IMAGES_URL
+        )
+
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                # These kwargs will be passed to _generate_examples
+                gen_kwargs={
+                    "metadata_path": metadata_path,
+                    "images_dir": images_dir,
+                    "conditioning_images_dir": conditioning_images_dir,
+                },
+            ),
+        ]
+
+    def _generate_examples(self, metadata_path, images_dir, conditioning_images_dir):
+        metadata = pd.read_json(metadata_path, lines=True)
+
+        for _, row in metadata.iterrows():
+            text = row["text"]
+
+            image_path = row["image"]
+            image_path = os.path.join(images_dir, image_path)
+            image = open(image_path, "rb").read()
+
+            conditioning_image_path = row["conditioning_image"]
+            conditioning_image_path = os.path.join(
+                conditioning_images_dir, row["conditioning_image"]
+            )
+            conditioning_image = open(conditioning_image_path, "rb").read()
+
+            yield row["image"], {
+                "text": text,
+                "image": {
+                    "path": image_path,
+                    "bytes": image,
+                },
+                "conditioning_image": {
+                    "path": conditioning_image_path,
+                    "bytes": conditioning_image,
+                },
+            }

train.jsonl

@@ -0,0 +1,19 @@
+{"text": "a rcn coat", "image": "images/1.png", "conditioning_image": "conditioning_images/1.png"}
+{"text": "a rcn coat", "image": "images/2.png", "conditioning_image": "conditioning_images/2.png"}
+{"text": "a rcn coat", "image": "images/3.png", "conditioning_image": "conditioning_images/3.png"}
+{"text": "a rcn coat", "image": "images/4.png", "conditioning_image": "conditioning_images/4.png"}
+{"text": "a rcn coat", "image": "images/5.png", "conditioning_image": "conditioning_images/5.png"}
+{"text": "a rcn coat", "image": "images/6.png", "conditioning_image": "conditioning_images/6.png"}
+{"text": "a rcn coat", "image": "images/7.png", "conditioning_image": "conditioning_images/7.png"}
+{"text": "a rcn coat", "image": "images/8.png", "conditioning_image": "conditioning_images/8.png"}
+{"text": "a rcn coat", "image": "images/9.png", "conditioning_image": "conditioning_images/9.png"}
+{"text": "a rcn coat", "image": "images/10.png", "conditioning_image": "conditioning_images/10.png"}
+{"text": "a rcn coat", "image": "images/11.png", "conditioning_image": "conditioning_images/11.png"}
+{"text": "a rcn coat", "image": "images/12.png", "conditioning_image": "conditioning_images/12.png"}
+{"text": "a rcn coat", "image": "images/13.png", "conditioning_image": "conditioning_images/13.png"}
+{"text": "a rcn coat", "image": "images/14.png", "conditioning_image": "conditioning_images/14.png"}
+{"text": "a rcn coat", "image": "images/15.png", "conditioning_image": "conditioning_images/15.png"}
+{"text": "a rcn coat", "image": "images/16.png", "conditioning_image": "conditioning_images/16.png"}
+{"text": "a rcn coat", "image": "images/17.png", "conditioning_image": "conditioning_images/17.png"}
+{"text": "a rcn coat", "image": "images/18.png", "conditioning_image": "conditioning_images/18.png"}
+{"text": "a rcn coat", "image": "images/19.png", "conditioning_image": "conditioning_images/19.png"}