KerasPersonLab

KerasPersonLab copied to clipboard

model = get_personlab(train=False, with_preprocess_lambda=True, intermediate_supervision=True, intermediate_layer='res4b12_relu', build_base_func=get_resnet101_base, output_stride=16)

Downloading data from None

TypeError Traceback (most recent call last) in () 4 intermediate_layer='res4b12_relu', 5 build_base_func=get_resnet101_base, ----> 6 output_stride=16) 7 print('Loading time: {}'.format(time()-tic))

~/alibaba/repos/KerasPersonLab/model.py in get_personlab(train, input_tensors, with_preprocess_lambda, build_base_func, intermediate_supervision, intermediate_layer, output_stride) 206 207 if with_preprocess_lambda not in [False, 0, None]: --> 208 base_model = build_base_func(input_tensor=normalized_img, output_stride=output_stride, return_model=True) 209 else: 210 base_model = build_base_func(input_tensor=input_img, output_stride=output_stride, return_model=True)

~/alibaba/repos/KerasPersonLab/resnet101.py in get_resnet101_base(input_tensor, output_stride, return_model) 150 151 model = KM.Model(input_tensor_source, x) --> 152 weights_path = get_file('resnet101_weights_notop.h5', None, cache_subdir='models') 153 model.load_weights(weights_path) 154

~/virtualenvironment/tensorflow/lib/python3.6/site-packages/keras/utils/data_utils.py in get_file(fname, origin, untar, md5_hash, file_hash, cache_subdir, hash_algorithm, extract, archive_format, cache_dir) 220 try: 221 try: --> 222 urlretrieve(origin, fpath, dl_progress) 223 except URLError as e: 224 raise Exception(error_msg.format(origin, e.errno, e.reason))

/usr/local/Cellar/python/3.6.5/Frameworks/Python.framework/Versions/3.6/lib/python3.6/urllib/request.py in urlretrieve(url, filename, reporthook, data) 244 data file as well as the resulting HTTPMessage object. 245 """ --> 246 url_type, path = splittype(url) 247 248 with contextlib.closing(urlopen(url, data)) as fp:

/usr/local/Cellar/python/3.6.5/Frameworks/Python.framework/Versions/3.6/lib/python3.6/urllib/parse.py in splittype(url) 936 _typeprog = re.compile('([^/:]+):(.*)', re.DOTALL) 937 --> 938 match = _typeprog.match(url) 939 if match: 940 scheme, data = match.groups()

TypeError: expected string or bytes-like object

Hi @rorlich . Thanks for your interest. Unfortunately, I have not made any pre-trained models available yet. The demo notebook is just an illustration of how to process the outputs network if you trained one.

I hope to make a trained model publicly available soon.

Looking forward for the pre-trained models.

If you've got a pretrained model I'd love to see it as well!

Has a pre-trained model been made available?

also interested in a trained model

also interested in a trained model