Upload trainimgs.py

trainimgs.py

@@ -0,0 +1,133 @@
+import os
+import datasets
+import urllib.request
+import csv
+
+_CITATION = """\
+@InProceedings{huggingface:dataset,
+title = {diffusion train set},
+}
+"""
+
+_DESCRIPTION = """\
+This is a dataset that image data and caption txt
+"""
+_HOMEPAGE = ""
+
+_LICENSE = ""
+
+_URL = "./"
+
+_URLS = {        
+            "train": _URL + "train_dataset.csv",
+            "reg": _URL + "reg_dataset.csv",
+            }
+
+class imgdataset(datasets.GeneratorBasedBuilder):
+    """"""
+    VERSION = datasets.Version("1.1.0")
+
+    BUILDER_CONFIGS = [
+        datasets.BuilderConfig(name="train", version=VERSION),
+        datasets.BuilderConfig(name="reg", version=VERSION),
+    ]
+
+
+    def _info(self):
+        if self.config.name == "train":  
+            features = datasets.Features(
+                {
+                    "Class_name": datasets.Value("string"),
+                    "file_name": datasets.Value("string"),
+                    "file_id": datasets.Value("string")
+
+                }
+            )
+        else:  
+            features = datasets.Features(
+                {
+                    "Class_name": datasets.Value("string"),
+                    "file_name": datasets.Value("string"),
+                    "file_id": datasets.Value("string")
+
+                }
+            )
+        return datasets.DatasetInfo(
+            description=_DESCRIPTION,
+            features=features, 
+        )
+    
+    def _split_generators(self, dl_manager):
+        """This function returns the examples in the raw (text) form."""
+        
+        urls = _URLS[self.config.name]
+        data_dir = dl_manager.download_and_extract(urls)
+        return [
+            datasets.SplitGenerator(
+                name=datasets.Split.TRAIN,
+                # These kwargs will be passed to _generate_examples
+                gen_kwargs={
+                    "filepath": os.path.join(data_dir, "train.jsonl"),
+                    "split": "train",
+                },
+            ),
+            datasets.SplitGenerator(
+                name=datasets.Split.VALIDATION,
+                # These kwargs will be passed to _generate_examples
+                gen_kwargs={
+                    "filepath": os.path.join(data_dir, "reg.jsonl"),
+                    "split": "reg",
+                },
+            ),
+       ]
+    def _download_image(self, dl_manager):
+
+
+        train = open("/content/drive/MyDrive/train_dataset.csv", "r")
+        train_reader = csv.DictReader(train)
+        # if not os.path.isdir(outpath): #폴더가 존재하지 않는다면 폴더 생성
+        #     os.makedirs(outpath)
+        for row in train_reader:
+            class_name = f"{row['Class_name']}"
+            file_name = f"{row['file_name']}"
+            url = f"{row['file_id']}"
+            path = os.path.join('./img',class_name,file_name)
+            folder = os.path.join('./img',class_name)
+            if not os.path.isdir(folder): #폴더가 존재하지 않는다면 폴더 생성
+                os.makedirs(folder)
+            urllib.request.urlretrieve(url, path)
+
+        reg = open("/content/drive/MyDrive/reg_dataset.csv", "r")
+        reg_reader = csv.DictReader(reg)
+        # if not os.path.isdir(outpath): #폴더가 존재하지 않는다면 폴더 생성
+        #     os.makedirs(outpath)
+        for row in reg_reader:
+            class_name = f"{row['Class_name']}"
+            file_name = f"{row['file_name']}"
+            url = f"{row['file_id']}"
+            path = os.path.join('./reg',class_name,file_name)
+            folder = os.path.join('./reg',class_name)
+            if not os.path.isdir(folder): #폴더가 존재하지 않는다면 폴더 생성
+                os.makedirs(folder)
+            urllib.request.urlretrieve(url, path)
+            
+    # def _generate_examples(self, filepath, split):
+    #     # TODO: This method handles input defined in _split_generators to yield (key, example) tuples from the dataset.
+    #     # The `key` is for legacy reasons (tfds) and is not important in itself, but must be unique for each example.
+        
+    #     with open(filepath, encoding="utf-8") as f:
+    #         for key, row in enumerate(f):
+    #             data = csv.loads(row)
+    #             if self.config.name == "train":  
+    #                 yield key, {
+    #                     "Class_name": datasets.Value("string"),
+    #                     "file_name": datasets.Value("string"),
+    #                     "file_id": datasets.Value("string")
+    #                 }
+    #             else:  
+    #                 yield key, {
+    #                     "Class_name": datasets.Value("string"),
+    #                     "file_name": datasets.Value("string"),
+    #                     "file_id": datasets.Value("string")
+    #                 }
+