我们从Python开源项目中，提取了以下8个代码示例，用于说明如何使用torch.__version__()。
def log_pytorch_version_info(): import torch logger.info("Pytorch version: %s", torch.__version__)
def parse_torch_version(): """ Parses `torch.__version__` into a semver-ish version tuple. This is needed to handle subpatch `_n` parts outside of the semver spec. :returns: a tuple `(major, minor, patch, extra_stuff)` """ match = re.match(r"(\d\.\d\.\d)(.*)", torch.__version__) major, minor, patch = map(int, match.group(1).split(".")) extra_stuff = match.group(2) return major, minor, patch, extra_stuff
def paramscheck(self): torch_version = torch.__version__ torch_version_split = torch_version.split("_") if torch_version_split[0] != "0.1.10": self.drawNetwork = False print "|===>DrawNetwork is unsupported by PyTorch with version: ", torch_version if self.netType == "LeNet": self.save_path = "log_%s_%s_%s/" % (self.netType, self.data_set, self.experimentID) else: self.save_path = "log_%s_%s_%d_%s/" % (self.netType, self.data_set, self.depth, self.experimentID) if self.useDefaultSetting: print("|===> Use Default Setting") if self.data_set == "cifar10" or self.data_set == "cifar100": if self.nEpochs == 160: self.LR = 0.5 self.lrPolicy = "exp" self.momentum = 0.9 self.weightDecay = 1e-4 self.step = 2.0 self.gamma = math.pow(0.001 / self.LR, 1.0/math.floor(self.nEpochs/self.step)) else: self.LR = 0.1 self.lrPolicy = "multistep" self.momentum = 0.9 self.weightDecay = 1e-4 else: assert False, "invalid data set" if self.data_set == "cifar10" or self.data_set == "mnist": self.nClasses = 10 elif self.data_set == "cifar100": self.nClasses = 100
def get_config(): ret = [] ret.append(colored('pyprob {}'.format(pyprob.__version__), 'blue', attrs=['bold'])) ret.append('PyTorch {}'.format(torch.__version__)) cpu_info = cpuinfo.get_cpu_info() if 'brand' in cpu_info: ret.append('CPU : {}'.format(cpu_info['brand'])) else: ret.append('CPU : unknown') if 'count' in cpu_info: ret.append('CPU count : {0} (logical)'.format(cpu_info['count'])) else: ret.append('CPU count : unknown') if torch.cuda.is_available(): ret.append('CUDA : available') ret.append('CUDA devices : {0}'.format(torch.cuda.device_count())) if cuda_enabled: if cuda_device == -1: ret.append('CUDA selected : all') else: ret.append('CUDA selected : {0}'.format(cuda_device)) else: ret.append('CUDA : not available') if cuda_enabled: ret.append('Running on : CUDA') else: ret.append('Running on : CPU') return '\n'.join(ret)
def version(self): """ Returns the PyTorch version, as a tuple of (MAJOR, MINOR, PATCH). """ import torch # pylint: disable=import-error version = torch.__version__ match = re.match(r'([0-9]+)\.([0-9]+)\.([0-9]+)\.*', version) if not match: logger.warning('Unable to infer PyTorch version. We ' 'cannot check for version incompatibilities.') return (0, 0, 0) return tuple(int(x) for x in match.groups()) ###########################################################################
def __init__(self, dropout=0.2, cuda=False, cuda_device_id=0, standardize_observes=False, softmax_boost=20, mixture_components=10): super(Artifact, self).__init__() self.sample_layers = {} self.proposal_layers = {} self.observe_layer = None self.lstm = None self.model_name = '' self.created = util.get_time_str() self.modified = util.get_time_str() self.on_cuda = cuda self.trained_on = '' self.cuda_device_id = cuda_device_id self.code_version = pyprob.__version__ self.pytorch_version = torch.__version__ self.standardize_observes = standardize_observes self.one_hot_address = {} self.one_hot_distribution = {} self.one_hot_address_dim = None self.one_hot_distribution_dim = None self.one_hot_address_empty = None self.one_hot_distribution_empty = None self.address_histogram = {} self.trace_length_histogram = {} self.valid_size = None self.valid_batch = None self.lstm_dim = None self.lstm_depth = None self.lstm_input_dim = None self.smp_emb_dim = None self.obs_emb = None self.obs_emb_dim = None self.num_params_history_trace = [] self.num_params_history_num_params = [] self.trace_length_min = sys.maxsize self.trace_length_max = 0 self.trace_examples_histogram = {} self.trace_examples_addresses = {} self.trace_examples_limit = 10000 self.train_loss_best = None self.train_loss_worst = None self.valid_loss_best = None self.valid_loss_worst = None self.valid_history_trace = [] self.valid_history_loss = [] self.train_history_trace = [] self.train_history_loss = [] self.total_training_seconds = 0 self.total_iterations = 0 self.total_traces = 0 self.updates = 0 self.optimizer = None self.dropout = dropout self.softmax_boost = softmax_boost self.mixture_components = mixture_components self._state_observes = None self._state_observes_embedding = None self._state_new_trace = True
def load_artifact(file_name, cuda=False, device_id=-1): try: if cuda: artifact = torch.load(file_name) else: artifact = torch.load(file_name, map_location=lambda storage, loc: storage) except: logger.log_error('load_artifact: Cannot load file') if artifact.code_version != pyprob.__version__: logger.log() logger.log_warning('Different pyprob versions (artifact: {0}, current: {1})'.format(artifact.code_version, pyprob.__version__)) logger.log() if artifact.pytorch_version != torch.__version__: logger.log() logger.log_warning('Different PyTorch versions (artifact: {0}, current: {1})'.format(artifact.pytorch_version, torch.__version__)) logger.log() # if print_info: # file_size = '{:,}'.format(os.path.getsize(file_name)) # log_print('File name : {0}'.format(file_name)) # log_print('File size (Bytes) : {0}'.format(file_size)) # log_print(artifact.get_info()) # log_print() if cuda: if device_id == -1: device_id = torch.cuda.current_device() if artifact.on_cuda: if device_id != artifact.cuda_device_id: logger.log_warning('Loading CUDA (device {0}) artifact to CUDA (device {1})'.format(artifact.cuda_device_id, device_id)) logger.log() artifact.move_to_cuda(device_id) else: logger.log_warning('Loading CPU artifact to CUDA (device {0})'.format(device_id)) logger.log() artifact.move_to_cuda(device_id) else: if artifact.on_cuda: logger.log_warning('Loading CUDA artifact to CPU') logger.log() artifact.move_to_cpu() return artifact
