Datasets:
				
			
			
	
			
	
		
			
	
		Tasks:
	
	
	
	
	Image Classification
	
	
	Sub-tasks:
	
	
	
	
	multi-class-image-classification
	
	
	Languages:
	
	
	
		
	
	English
	
	
	Size:
	
	
	
	
	10M<n<100M
	
	
	ArXiv:
	
	
	
	
	
	
	
	
License:
	
	
	
	
	
	
	
Commit 
							
							·
						
						83b2798
	
1
								Parent(s):
							
							059cb70
								
Make script parallelizable (#3)
Browse files- Make script parallelizable (b89324d4754d89c8332d8b2999b22353946d5809)
- quickdraw.py +6 -4
    	
        quickdraw.py
    CHANGED
    
    | @@ -246,6 +246,7 @@ class Quickdraw(datasets.GeneratorBasedBuilder): | |
| 246 | 
             
                        files = dl_manager.download(
         | 
| 247 | 
             
                            {name: url for name, url in zip(_NAMES, [base_url.format(name) for name in _NAMES])}
         | 
| 248 | 
             
                        )
         | 
|  | |
| 249 | 
             
                        return [
         | 
| 250 | 
             
                            datasets.SplitGenerator(
         | 
| 251 | 
             
                                name=datasets.Split.TRAIN,
         | 
| @@ -259,6 +260,7 @@ class Quickdraw(datasets.GeneratorBasedBuilder): | |
| 259 | 
             
                        files = dl_manager.download_and_extract(
         | 
| 260 | 
             
                            {name: url for name, url in zip(_NAMES, [base_url.format(name) for name in _NAMES])}
         | 
| 261 | 
             
                        )
         | 
|  | |
| 262 | 
             
                        return [
         | 
| 263 | 
             
                            datasets.SplitGenerator(
         | 
| 264 | 
             
                                name=datasets.Split.TRAIN,
         | 
| @@ -286,7 +288,7 @@ class Quickdraw(datasets.GeneratorBasedBuilder): | |
| 286 | 
             
                def _generate_examples(self, files, split):
         | 
| 287 | 
             
                    if self.config.name == "raw":
         | 
| 288 | 
             
                        idx = 0
         | 
| 289 | 
            -
                        for _, file in files | 
| 290 | 
             
                            with open(file, encoding="utf-8") as f:
         | 
| 291 | 
             
                                for line in f:
         | 
| 292 | 
             
                                    example = json.loads(line)
         | 
| @@ -296,7 +298,7 @@ class Quickdraw(datasets.GeneratorBasedBuilder): | |
| 296 | 
             
                                    idx += 1
         | 
| 297 | 
             
                    elif self.config.name == "preprocessed_simplified_drawings":
         | 
| 298 | 
             
                        idx = 0
         | 
| 299 | 
            -
                        for label, file in files | 
| 300 | 
             
                            with open(file, "rb") as f:
         | 
| 301 | 
             
                                while True:
         | 
| 302 | 
             
                                    try:
         | 
| @@ -308,7 +310,7 @@ class Quickdraw(datasets.GeneratorBasedBuilder): | |
| 308 | 
             
                                    idx += 1
         | 
| 309 | 
             
                    elif self.config.name == "preprocessed_bitmaps":
         | 
| 310 | 
             
                        idx = 0
         | 
| 311 | 
            -
                        for label, file in files | 
| 312 | 
             
                            with open(file, "rb") as f:
         | 
| 313 | 
             
                                images = np.load(f)
         | 
| 314 | 
             
                                for image in images:
         | 
| @@ -319,7 +321,7 @@ class Quickdraw(datasets.GeneratorBasedBuilder): | |
| 319 | 
             
                                    idx += 1
         | 
| 320 | 
             
                    else:  # sketch_rnn, sketch_rnn_full
         | 
| 321 | 
             
                        idx = 0
         | 
| 322 | 
            -
                        for label, file in files | 
| 323 | 
             
                            with open(os.path.join(file, f"{split}.npy"), "rb") as f:
         | 
| 324 | 
             
                                # read entire file since f.seek is not supported in the streaming mode
         | 
| 325 | 
             
                                drawings = np.load(io.BytesIO(f.read()), encoding="latin1", allow_pickle=True)
         | 
|  | |
| 246 | 
             
                        files = dl_manager.download(
         | 
| 247 | 
             
                            {name: url for name, url in zip(_NAMES, [base_url.format(name) for name in _NAMES])}
         | 
| 248 | 
             
                        )
         | 
| 249 | 
            +
                        files = [(name, file) for name, file in files.items()]
         | 
| 250 | 
             
                        return [
         | 
| 251 | 
             
                            datasets.SplitGenerator(
         | 
| 252 | 
             
                                name=datasets.Split.TRAIN,
         | 
|  | |
| 260 | 
             
                        files = dl_manager.download_and_extract(
         | 
| 261 | 
             
                            {name: url for name, url in zip(_NAMES, [base_url.format(name) for name in _NAMES])}
         | 
| 262 | 
             
                        )
         | 
| 263 | 
            +
                        files = [(name, file) for name, file in files.items()]
         | 
| 264 | 
             
                        return [
         | 
| 265 | 
             
                            datasets.SplitGenerator(
         | 
| 266 | 
             
                                name=datasets.Split.TRAIN,
         | 
|  | |
| 288 | 
             
                def _generate_examples(self, files, split):
         | 
| 289 | 
             
                    if self.config.name == "raw":
         | 
| 290 | 
             
                        idx = 0
         | 
| 291 | 
            +
                        for _, file in files:
         | 
| 292 | 
             
                            with open(file, encoding="utf-8") as f:
         | 
| 293 | 
             
                                for line in f:
         | 
| 294 | 
             
                                    example = json.loads(line)
         | 
|  | |
| 298 | 
             
                                    idx += 1
         | 
| 299 | 
             
                    elif self.config.name == "preprocessed_simplified_drawings":
         | 
| 300 | 
             
                        idx = 0
         | 
| 301 | 
            +
                        for label, file in files:
         | 
| 302 | 
             
                            with open(file, "rb") as f:
         | 
| 303 | 
             
                                while True:
         | 
| 304 | 
             
                                    try:
         | 
|  | |
| 310 | 
             
                                    idx += 1
         | 
| 311 | 
             
                    elif self.config.name == "preprocessed_bitmaps":
         | 
| 312 | 
             
                        idx = 0
         | 
| 313 | 
            +
                        for label, file in files:
         | 
| 314 | 
             
                            with open(file, "rb") as f:
         | 
| 315 | 
             
                                images = np.load(f)
         | 
| 316 | 
             
                                for image in images:
         | 
|  | |
| 321 | 
             
                                    idx += 1
         | 
| 322 | 
             
                    else:  # sketch_rnn, sketch_rnn_full
         | 
| 323 | 
             
                        idx = 0
         | 
| 324 | 
            +
                        for label, file in files:
         | 
| 325 | 
             
                            with open(os.path.join(file, f"{split}.npy"), "rb") as f:
         | 
| 326 | 
             
                                # read entire file since f.seek is not supported in the streaming mode
         | 
| 327 | 
             
                                drawings = np.load(io.BytesIO(f.read()), encoding="latin1", allow_pickle=True)
         | 

