code
stringlengths
101
5.91M
_grad() def evaluation(model, data_loader, tokenizer, device, config): model.eval() metric_logger = utils.MetricLogger(delimiter=' ') header = 'Generate VQA test result:' print_freq = 50 result = [] answer_list = [(answer + config['eos']) for answer in data_loader.dataset.answer_list] answer_input = tokenizer(answer_list, padding='longest', return_tensors='pt').to(device) for (n, (image, question, question_id)) in enumerate(metric_logger.log_every(data_loader, print_freq, header)): image = image.to(device, non_blocking=True) question_input = tokenizer(question, padding='longest', return_tensors='pt').to(device) (topk_ids, topk_probs) = model(image, question_input, answer_input, train=False, k=config['k_test']) for (ques_id, topk_id, topk_prob) in zip(question_id, topk_ids, topk_probs): ques_id = int(ques_id.item()) (_, pred) = topk_prob.max(dim=0) result.append({'question_id': ques_id, 'answer': data_loader.dataset.answer_list[topk_id[pred]]}) return result
class AnchorGenerator(object): __metaclass__ = ABCMeta def name_scope(self): pass def check_num_anchors(self): return True def num_anchors_per_location(self): pass def generate(self, feature_map_shape_list, **params): if (self.check_num_anchors and (len(feature_map_shape_list) != len(self.num_anchors_per_location()))): raise ValueError('Number of feature maps is expected to equal the length of `num_anchors_per_location`.') with tf.name_scope(self.name_scope()): anchors = self._generate(feature_map_shape_list, **params) if self.check_num_anchors: with tf.control_dependencies([self._assert_correct_number_of_anchors(anchors, feature_map_shape_list)]): anchors.set(tf.identity(anchors.get())) return anchors def _generate(self, feature_map_shape_list, **params): pass def _assert_correct_number_of_anchors(self, anchors, feature_map_shape_list): expected_num_anchors = 0 for (num_anchors_per_location, feature_map_shape) in zip(self.num_anchors_per_location(), feature_map_shape_list): expected_num_anchors += ((num_anchors_per_location * feature_map_shape[0]) * feature_map_shape[1]) return tf.assert_equal(expected_num_anchors, anchors.num_boxes())
def main(): parser = argparse.ArgumentParser(description='PyTorch ImageNet Training') parser.add_argument('data', metavar='DIR', help='path to dataset') parser.add_argument('-a', '--arch', metavar='ARCH', default='resnet18', choices=model_names, help=(('model architecture: ' + ' | '.join(model_names)) + ' (default: resnet18)')) parser.add_argument('--epochs', default=90, type=int, metavar='N', help='number of total epochs to run') parser.add_argument('--max_epochs', default=90, type=int, metavar='N', help='number of max epochs to run') parser.add_argument('--start-epoch', default=0, type=int, metavar='N', help='manual epoch number (useful on restarts)') parser.add_argument('-b', '--batch-size', default=256, type=int, metavar='N', help='mini-batch size (default: 256), this is the total batch size of all GPUs on the current node when using Data Parallel or Distributed Data Parallel') parser.add_argument('--lr', '--learning-rate', default=0.1, type=float, metavar='LR', help='initial learning rate', dest='lr') parser.add_argument('--momentum', default=0.9, type=float, metavar='M', help='momentum') parser.add_argument('--wd', '--weight-decay', default=0.0001, type=float, metavar='W', help='weight decay (default: 1e-4)', dest='weight_decay') parser.add_argument('-p', '--print-freq', default=10, type=int, metavar='N', help='print frequency (default: 10)') parser.add_argument('--resume', default='', type=str, metavar='PATH', help='path to latest checkpoint (default: none)') parser.add_argument('-e', '--evaluate', dest='evaluate', action='store_true', help='evaluate model on validation set') parser.add_argument('--pretrained', dest='pretrained', action='store_true', help='use pre-trained model') parser.add_argument('--world-size', default=(- 1), type=int, help='number of nodes for distributed training') parser.add_argument('--rank', default=(- 1), type=int, help='node rank for distributed training') parser.add_argument('--seed', default=None, type=int, help='seed for initializing training. ') parser.add_argument('--cores', default=4, type=int, help='num of CPUs to use.') parser.add_argument('--nodes', default=1, type=int, help='num of nodes to use.') parser.add_argument('--executor_memory', default='20g', type=str, help='size of executor memory.') parser.add_argument('--driver_memory', default='20g', type=str, help='size of driver memory.') parser.add_argument('--driver_cores', default=1, type=int, help='num of driver cores to use.') parser.add_argument('--num_executors', type=int, default=16, help='number of executors') parser.add_argument('--deploy_mode', type=str, default='yarn-client', help='yarn deploy mode, yarn-client or yarn-cluster') args = parser.parse_args() hadoop_conf = os.environ.get('HADOOP_CONF_DIR') invalidInputError((hadoop_conf is not None), 'Directory path to hadoop conf not found for yarn-client mode. Please set the environment variable HADOOP_CONF_DIR') sc = init_orca_context(cluster_mode=args.deploy_mode, hadoop_conf=hadoop_conf, conf={'spark.executor.memory': args.executor_memory, 'spark.executor.cores': args.cores, 'spark.executor.instances': args.num_executors}) traindir = os.path.join(args.data, 'train') valdir = os.path.join(args.data, 'val') normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]) train_dataset = datasets.ImageFolder(traindir, transforms.Compose([transforms.RandomResizedCrop(224), transforms.RandomHorizontalFlip(), transforms.ToTensor(), normalize])) train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=args.batch_size, shuffle=True) model = torchvision.models.resnet50() val_loader = torch.utils.data.DataLoader(datasets.ImageFolder(valdir, transforms.Compose([transforms.Resize(256), transforms.CenterCrop(224), transforms.ToTensor(), normalize])), batch_size=args.batch_size, shuffle=False) iterationPerEpoch = int(math.ceil((float(1281167) / args.batch_size))) step = Step((iterationPerEpoch * 30), 0.1) zooOptimizer = SGD(args.lr, momentum=args.momentum, dampening=0.0, leaningrate_schedule=step, weightdecay=args.weight_decay) zooModel = TorchModel.from_pytorch(model) criterion = torch.nn.CrossEntropyLoss() zooCriterion = TorchLoss.from_pytorch(criterion) estimator = Estimator(zooModel, optim_methods=zooOptimizer) train_featureSet = FeatureSet.pytorch_dataloader(train_loader) test_featureSet = FeatureSet.pytorch_dataloader(val_loader) estimator.train_minibatch(train_featureSet, zooCriterion, end_trigger=MaxEpoch(args.max_epochs), checkpoint_trigger=EveryEpoch(), validation_set=test_featureSet, validation_method=[Accuracy(), Top5Accuracy()])
def fill_parameters(parameters, default_parameters, name='unknown function'): string = '' if ('verbose' not in parameters.keys()): if ('verbose' not in default_parameters.keys()): default_parameters['verbose'] = 0 parameters['verbose'] = default_parameters['verbose'] verbose = parameters['verbose'] for key in default_parameters.keys(): check = False if (key not in parameters.keys()): check = True elif (('param' in key) or ('opt' in key)): if isinstance(parameters[key], dict): check = True if check: val = default_parameters[key] if (val is None): raise Exception((((('In ' + name) + ', field "') + key) + '" is required.')) elif (isinstance(val, dict) and (('param' in key) or ('opt' in key))): if (key not in parameters.keys()): parameters[key] = {} in_verbose = 0 if ('verbose' in parameters[key]): in_verbose = parameters[key]['verbose'] parameters[key]['verbose'] = verbose parameters[key] = fill_parameters(parameters[key], default_parameters[key], (name + ('\'s subfield "%s"' % key))) parameters[key]['verbose'] = in_verbose else: parameters[key] = val if verbose: string = (((('In ' + name) + ', field "') + key) + '" set to default value.') if (type(val) in (tuple, float, bool, int)): string = (string[:(- 1)] + (': %s.' % val)) elif isinstance(val, str): string = (string[:(- 1)] + (': "%s".' % val)) print(string) return parameters
class _DeformConv(Function): def forward(ctx, input, offset, weight, stride=1, padding=0, dilation=1, groups=1, deformable_groups=1, im2col_step=64): if ((input is not None) and (input.dim() != 4)): raise ValueError('Expected 4D tensor as input, got {}D tensor instead.'.format(input.dim())) ctx.stride = _pair(stride) ctx.padding = _pair(padding) ctx.dilation = _pair(dilation) ctx.groups = groups ctx.deformable_groups = deformable_groups ctx.im2col_step = im2col_step ctx.save_for_backward(input, offset, weight) output = input.new_empty(_DeformConv._output_size(input, weight, ctx.padding, ctx.dilation, ctx.stride)) ctx.bufs_ = [input.new_empty(0), input.new_empty(0)] if (not input.is_cuda): raise NotImplementedError else: cur_im2col_step = _DeformConv._cal_im2col_step(input.shape[0], ctx.im2col_step) assert ((input.shape[0] % cur_im2col_step) == 0), 'im2col step must divide batchsize' _C.deform_conv_forward(input, weight, offset, output, ctx.bufs_[0], ctx.bufs_[1], weight.size(3), weight.size(2), ctx.stride[1], ctx.stride[0], ctx.padding[1], ctx.padding[0], ctx.dilation[1], ctx.dilation[0], ctx.groups, ctx.deformable_groups, cur_im2col_step) return output _differentiable def backward(ctx, grad_output): (input, offset, weight) = ctx.saved_tensors grad_input = grad_offset = grad_weight = None if (not grad_output.is_cuda): raise NotImplementedError else: cur_im2col_step = _DeformConv._cal_im2col_step(input.shape[0], ctx.im2col_step) assert ((input.shape[0] % cur_im2col_step) == 0), 'im2col step must divide batchsize' if (ctx.needs_input_grad[0] or ctx.needs_input_grad[1]): grad_input = torch.zeros_like(input) grad_offset = torch.zeros_like(offset) _C.deform_conv_backward_input(input, offset, grad_output, grad_input, grad_offset, weight, ctx.bufs_[0], weight.size(3), weight.size(2), ctx.stride[1], ctx.stride[0], ctx.padding[1], ctx.padding[0], ctx.dilation[1], ctx.dilation[0], ctx.groups, ctx.deformable_groups, cur_im2col_step) if ctx.needs_input_grad[2]: grad_weight = torch.zeros_like(weight) _C.deform_conv_backward_filter(input, offset, grad_output, grad_weight, ctx.bufs_[0], ctx.bufs_[1], weight.size(3), weight.size(2), ctx.stride[1], ctx.stride[0], ctx.padding[1], ctx.padding[0], ctx.dilation[1], ctx.dilation[0], ctx.groups, ctx.deformable_groups, 1, cur_im2col_step) return (grad_input, grad_offset, grad_weight, None, None, None, None, None, None) def _output_size(input, weight, padding, dilation, stride): channels = weight.size(0) output_size = (input.size(0), channels) for d in range((input.dim() - 2)): in_size = input.size((d + 2)) pad = padding[d] kernel = ((dilation[d] * (weight.size((d + 2)) - 1)) + 1) stride_ = stride[d] output_size += (((((in_size + (2 * pad)) - kernel) // stride_) + 1),) if (not all(map((lambda s: (s > 0)), output_size))): raise ValueError('convolution input is too small (output would be {})'.format('x'.join(map(str, output_size)))) return output_size _cache(maxsize=128) def _cal_im2col_step(input_size, default_size): if (input_size <= default_size): return input_size best_step = 1 for step in range(2, min((int(math.sqrt(input_size)) + 1), default_size)): if ((input_size % step) == 0): if ((input_size // step) <= default_size): return (input_size // step) best_step = step return best_step
class _NCEGeneratorState(object): def __init__(self, context_size): self._doc_id = multiprocessing.RawValue('i', 0) self._in_doc_pos = multiprocessing.RawValue('i', context_size) self._lock = multiprocessing.Lock() def update_state(self, dataset, batch_size, context_size, num_examples_in_doc): with self._lock: doc_id = self._doc_id.value in_doc_pos = self._in_doc_pos.value self._advance_indices(dataset, batch_size, context_size, num_examples_in_doc) return (doc_id, in_doc_pos) def _advance_indices(self, dataset, batch_size, context_size, num_examples_in_doc): num_examples = num_examples_in_doc(dataset[self._doc_id.value], self._in_doc_pos.value) if (num_examples > batch_size): self._in_doc_pos.value += batch_size return if (num_examples == batch_size): if (self._doc_id.value < (len(dataset) - 1)): self._doc_id.value += 1 else: self._doc_id.value = 0 self._in_doc_pos.value = context_size return while (num_examples < batch_size): if (self._doc_id.value == (len(dataset) - 1)): self._doc_id.value = 0 self._in_doc_pos.value = context_size return self._doc_id.value += 1 num_examples += num_examples_in_doc(dataset[self._doc_id.value]) self._in_doc_pos.value = ((len(dataset[self._doc_id.value].text) - context_size) - (num_examples - batch_size))
def successors(ctree, cerrors, gold): for merror in cerrors.missing: for eerror in cerrors.extra: if (merror[1] == eerror[1]): (yield gen_different_label_successor(ctree, eerror[1], eerror[2], merror[2])) for error in cerrors.missing: (yield gen_missing_successor(ctree, error)) for error in cerrors.extra: (yield gen_extra_successor(ctree, error, gold)) for source_span in ctree: for left in range(len(source_span.subtrees)): for right in range(left, len(source_span.subtrees)): if ((left == 0) and (right == (len(source_span.subtrees) - 1))): continue new_parents = [] if (left != 0): new_parent = source_span.subtrees[(left - 1)] while (not new_parent.is_terminal()): if cerrors.is_extra(new_parent): new_parents.append(new_parent) new_parent = new_parent.subtrees[(- 1)] if (right != (len(source_span.subtrees) - 1)): new_parent = source_span.subtrees[(right + 1)] while (not new_parent.is_terminal()): if cerrors.is_extra(new_parent): new_parents.append(new_parent) new_parent = new_parent.subtrees[0] if (cerrors.is_extra(source_span) and ((left == 0) or (right == (len(source_span.subtrees) - 1)))): if (left == 0): if (source_span.subtrees[left].span[0] > 0): for new_parent in ctree.get_nodes('all', end=source_span.subtrees[left].span[0]): if cerrors.is_extra(new_parent): new_parents.append(new_parent) if (right == (len(source_span.subtrees) - 1)): if (source_span.subtrees[right].span[1] < ctree.span[1]): for new_parent in ctree.get_nodes('all', start=source_span.subtrees[right].span[1]): if cerrors.is_extra(new_parent): new_parents.append(new_parent) if (left == 0): new_parent = source_span.parent while (not (new_parent.parent is None)): new_parents.append(new_parent) if (new_parent.parent.span[0] < source_span.span[0]): break new_parent = new_parent.parent if (right == (len(source_span.subtrees) - 1)): new_parent = source_span.parent while (not (new_parent.parent is None)): new_parents.append(new_parent) if (new_parent.parent.span[1] > source_span.span[1]): break new_parent = new_parent.parent for new_parent in new_parents: (yield gen_move_successor(source_span, left, right, new_parent, cerrors, gold))
def test_masked_softmax_nll(): rng = np.random.RandomState(9823174) n_data = 22 data_dim = 45 logits_np = (4 * rng.randn(n_data, data_dim)) mask_np = (rng.randn(n_data, data_dim) > 0.0) while (not np.all(np.sum((mask_np == 1), axis=1))): mask_np = (rng.randn(n_data, data_dim) > 0.0) labels_np = (logits_np + (2 * rng.randn(n_data, data_dim))) labels_np[(~ mask_np)] = (- np.inf) labels_np = np.argmax(labels_np, axis=1) logits = torch.tensor(logits_np) mask = torch.tensor(mask_np) labels = torch.tensor(labels_np, dtype=torch.int64) nll_from_func = torch_utils.masked_softmax_nll(logits, labels, mask).numpy() softmax = nn.Softmax(dim=1) probs = softmax(logits) prob_adjustment_factor = torch.log(torch.sum((probs * mask.float()), dim=1)) criterion = nn.CrossEntropyLoss(reduction='none') loss_pers = criterion(logits, labels) loss_pers = (loss_pers + prob_adjustment_factor) nll_from_alternative = loss_pers.numpy() np.testing.assert_array_almost_equal(nll_from_func, nll_from_alternative)
class LongformerForSequenceClassification(metaclass=DummyObject): _backends = ['torch'] def __init__(self, *args, **kwargs): requires_backends(self, ['torch'])
class ResNetSharingClassifier(SharingClassifier): block = BasicBlock num_blocks = [2, 2, 2, 2] norm_layer = nn.InstanceNorm2d def __init__(self, params, experts): super().__init__(params, experts) self.precursors = [expert.d for expert in self.experts[1:]] first = (len(self.precursors) == 0) if (MODELS_NDPM_CLASSIFIER_NUM_BLOCKS is not None): num_blocks = MODELS_NDPM_CLASSIFIER_NUM_BLOCKS else: num_blocks = self.num_blocks if (MODELS_NDPM_CLASSIFIER_NORM_LAYER is not None): self.norm_layer = getattr(nn, MODELS_NDPM_CLASSIFIER_NORM_LAYER) else: self.norm_layer = nn.BatchNorm2d num_classes = n_classes[params.data] nf = (MODELS_NDPM_CLASSIFIER_CLS_NF_BASE if first else MODELS_NDPM_CLASSIFIER_CLS_NF_EXT) nf_cat = (MODELS_NDPM_CLASSIFIER_CLS_NF_BASE + (len(self.precursors) * MODELS_NDPM_CLASSIFIER_CLS_NF_EXT)) self.nf = (MODELS_NDPM_CLASSIFIER_CLS_NF_BASE if first else MODELS_NDPM_CLASSIFIER_CLS_NF_EXT) self.nf_cat = nf_cat self.layer0 = nn.Sequential(nn.Conv2d(3, (nf * 1), kernel_size=3, stride=1, padding=1, bias=False), self.norm_layer((nf * 1)), nn.ReLU()) self.layer1 = self._make_layer((nf_cat * 1), (nf * 1), num_blocks[0], stride=1) self.layer2 = self._make_layer((nf_cat * 1), (nf * 2), num_blocks[1], stride=2) self.layer3 = self._make_layer((nf_cat * 2), (nf * 4), num_blocks[2], stride=2) self.layer4 = self._make_layer((nf_cat * 4), (nf * 8), num_blocks[3], stride=2) self.predict = nn.Sequential(nn.Linear((nf_cat * 8), num_classes), nn.LogSoftmax(dim=1)) self.setup_optimizer() def _make_layer(self, nf_in, nf_out, num_blocks, stride): norm_layer = self.norm_layer block = self.block downsample = None if ((stride != 1) or (nf_in != nf_out)): downsample = nn.Sequential(conv1x1(nf_in, nf_out, stride), norm_layer(nf_out)) layers = [block(nf_in, nf_out, stride, downsample=downsample, norm_layer=norm_layer)] for _ in range(1, num_blocks): layers.append(block(nf_out, nf_out, norm_layer=norm_layer)) return nn.Sequential(*layers) def forward(self, x, collect=False): x = maybe_cuda(x) if (len(self.precursors) == 0): h1 = self.layer0(x) h2 = self.layer1(h1) h3 = self.layer2(h2) h4 = self.layer3(h3) h5 = self.layer4(h4) h5 = F.avg_pool2d(h5, h5.size(2)).view(h5.size(0), (- 1)) pred = self.predict(h5) if collect: return ([pred], [h1.detach(), h2.detach(), h3.detach(), h4.detach(), h5.detach()]) else: return pred (preds, features) = self.precursors[(- 1)](x, collect=True) h1 = self.layer0(x) h1_cat = torch.cat([features[0], h1], dim=1) h2 = self.layer1(h1_cat) h2_cat = torch.cat([features[1], h2], dim=1) h3 = self.layer2(h2_cat) h3_cat = torch.cat([features[2], h3], dim=1) h4 = self.layer3(h3_cat) h4_cat = torch.cat([features[3], h4], dim=1) h5 = self.layer4(h4_cat) h5 = F.avg_pool2d(h5, h5.size(2)).view(h5.size(0), (- 1)) h5_cat = torch.cat([features[4], h5], dim=1) pred = self.predict(h5_cat) if collect: preds.append(pred) return (preds, [h1_cat.detach(), h2_cat.detach(), h3_cat.detach(), h4_cat.detach(), h5_cat.detach()]) else: return pred
def write_to_ray(idx, partition, redis_address, redis_password, partition_store_names): init_ray_if_not(redis_address, redis_password) ip = ray._private.services.get_node_ip_address() local_store_name = None for name in partition_store_names: if name.endswith(ip): local_store_name = name break if (local_store_name is None): local_store_name = random.choice(partition_store_names) local_store = ray.get_actor(local_store_name) result = [] for (shard_id, shard) in enumerate(partition): shard_ref = ray.put(shard, _owner=local_store) result.append(local_store.upload_shards.remote((idx, shard_id), [shard_ref])) is_empty = (len(result) == 0) if is_empty: partition_ref = ray.put([], _owner=local_store) result.append(local_store.upload_partition.remote(idx, [partition_ref])) logger.warning(f'Partition {idx} is empty.') ray.get(result) return [(idx, local_store_name.split(':')[(- 1)], local_store_name)]
class TransFuse_L(nn.Module): def __init__(self, num_classes=1, drop_rate=0.2, normal_init=True, pretrained=False): super(TransFuse_L, self).__init__() self.resnet = resnet50() if pretrained: self.resnet.load_state_dict(torch.load('/home/chenfei/.cache/torch/hub/checkpoints/resnet50-19c8e357.pth')) self.resnet.fc = nn.Identity() self.resnet.layer4 = nn.Identity() self.transformer = deit_b(pretrained=pretrained) self.up1 = Up(in_ch1=768, out_ch=512) self.up2 = Up(512, 256) self.final_x = nn.Sequential(Conv(1024, 256, 1, bn=True, relu=True), Conv(256, 256, 3, bn=True, relu=True), Conv(256, num_classes, 3, bn=False, relu=False)) self.final_1 = nn.Sequential(Conv(256, 256, 3, bn=True, relu=True), Conv(256, num_classes, 3, bn=False, relu=False)) self.final_2 = nn.Sequential(Conv(256, 256, 3, bn=True, relu=True), Conv(256, num_classes, 3, bn=False, relu=False)) self.up_c = BiFusion_block(ch_1=1024, ch_2=768, r_2=4, ch_int=1024, ch_out=1024, drop_rate=(drop_rate / 2)) self.up_c_1_1 = BiFusion_block(ch_1=512, ch_2=512, r_2=2, ch_int=512, ch_out=512, drop_rate=(drop_rate / 2)) self.up_c_1_2 = Up(in_ch1=1024, out_ch=512, in_ch2=512, attn=True) self.up_c_2_1 = BiFusion_block(ch_1=256, ch_2=256, r_2=1, ch_int=256, ch_out=256, drop_rate=(drop_rate / 2)) self.up_c_2_2 = Up(512, 256, 256, attn=True) self.drop = nn.Dropout2d(drop_rate) if normal_init: self.init_weights() def forward(self, imgs, labels=None): x_b = self.transformer(imgs) x_b = torch.transpose(x_b, 1, 2) x_b = x_b.view(x_b.shape[0], (- 1), 32, 32) x_b = self.drop(x_b) x_b_1 = self.up1(x_b) x_b_1 = self.drop(x_b_1) x_b_2 = self.up2(x_b_1) x_b_2 = self.drop(x_b_2) x_u = self.resnet.conv1(imgs) x_u = self.resnet.bn1(x_u) x_u = self.resnet.relu(x_u) x_u = self.resnet.maxpool(x_u) x_u_2 = self.resnet.layer1(x_u) x_u_2 = self.drop(x_u_2) x_u_1 = self.resnet.layer2(x_u_2) x_u_1 = self.drop(x_u_1) x_u = self.resnet.layer3(x_u_1) x_u = self.drop(x_u) x_c = self.up_c(x_u, x_b) x_c_1_1 = self.up_c_1_1(x_u_1, x_b_1) x_c_1 = self.up_c_1_2(x_c, x_c_1_1) x_c_2_1 = self.up_c_2_1(x_u_2, x_b_2) x_c_2 = self.up_c_2_2(x_c_1, x_c_2_1) map_x = F.interpolate(self.final_x(x_c), scale_factor=16, mode='bilinear') map_1 = F.interpolate(self.final_1(x_b_2), scale_factor=4, mode='bilinear') map_2 = F.interpolate(self.final_2(x_c_2), scale_factor=4, mode='bilinear') return (map_x, map_1, map_2) def init_weights(self): self.up1.apply(init_weights) self.up2.apply(init_weights) self.final_x.apply(init_weights) self.final_1.apply(init_weights) self.final_2.apply(init_weights) self.up_c.apply(init_weights) self.up_c_1_1.apply(init_weights) self.up_c_1_2.apply(init_weights) self.up_c_2_1.apply(init_weights) self.up_c_2_2.apply(init_weights)
def get_early_stopping_callback(metric, patience): return EarlyStopping(monitor=f'val_{metric}', mode=('min' if ('loss' in metric) else 'max'), patience=patience, verbose=True)
def predict_compression(predictor, inp_batch_json, margin=0): output = predictor.predict_batch_json(inp_batch_json) rts = [] comps = [] for (idx, out) in enumerate(output): tag_logits = out['tag_logits'] tokens = inp_batch_json[idx]['sentence'] original_len = len(tokens) rt = [] for (logit, tok) in zip(tag_logits, tokens): del_score = logit[0] retain_score = logit[1] if (del_score <= (retain_score + margin)): rt.append(tok.text) compressed_len = len(rt) rt = ' '.join(rt) compression_rate = (compressed_len / original_len) comps.append(compression_rate) rts.append(rt) return (rts, comps)
def get_declarative_equations(model, question, prompt, max_tokens, stop_token, temperature): prompt = prompt.format(question=question) response = openai.Completion.create(model=model, prompt=prompt, max_tokens=max_tokens, stop=stop_token, temperature=temperature, top_p=1) result = response['choices'][0]['text'] eq_list = re.findall('\\[\\[.*?\\]\\]', result) if (len(eq_list) > 0): return reformat_equations_from_peano(reformat_incre_equations(eq_list)) else: print() print(response['choices'][0]['text']) return response['choices'][0]['text']
class dataset(): def __init__(self, root=None, train=True, example_weight=None): self.root = root self.train = train self.transform = transforms.ToTensor() if self.train: train_data_path = os.path.join(root, 'train_data') train_labels_path = os.path.join(root, 'train_labels') self.train_data = numpy.load(open(train_data_path, 'rb')) self.train_data = torch.from_numpy(self.train_data.astype('float32')) self.train_labels = numpy.load(open(train_labels_path, 'rb')).astype('int') self.example_weight = example_weight else: test_data_path = os.path.join(root, 'test_data') test_labels_path = os.path.join(root, 'test_labels') self.test_data = numpy.load(open(test_data_path, 'rb')) self.test_data = torch.from_numpy(self.test_data.astype('float32')) self.test_labels = numpy.load(open(test_labels_path, 'rb')).astype('int') def __len__(self): if self.train: return len(self.train_data) else: return len(self.test_data) def __getitem__(self, index): if self.train: (img, target, weight) = (self.train_data[index], self.train_labels[index], self.example_weight[index]) else: (img, target, weight) = (self.test_data[index], self.test_labels[index], []) return (img, target, weight)
def convert_cityscapes_instance_only(data_dir, out_dir): sets = ['gtFine_val', 'gtFine_train', 'gtFine_test'] ann_dirs = ['gtFine_trainvaltest/gtFine/val', 'gtFine_trainvaltest/gtFine/train', 'gtFine_trainvaltest/gtFine/test'] json_name = 'instancesonly_filtered_%s.json' ends_in = '%s_polygons.json' img_id = 0 ann_id = 0 cat_id = 1 category_dict = {} category_instancesonly = ['person', 'rider', 'car', 'truck', 'bus', 'train', 'motorcycle', 'bicycle'] for (data_set, ann_dir) in zip(sets, ann_dirs): print(('Starting %s' % data_set)) ann_dict = {} images = [] annotations = [] ann_dir = os.path.join(data_dir, ann_dir) for (root, _, files) in os.walk(ann_dir): for filename in files: if filename.endswith((ends_in % data_set.split('_')[0])): if ((len(images) % 50) == 0): print(('Processed %s images, %s annotations' % (len(images), len(annotations)))) json_ann = json.load(open(os.path.join(root, filename))) image = {} image['id'] = img_id img_id += 1 image['width'] = json_ann['imgWidth'] image['height'] = json_ann['imgHeight'] image['file_name'] = (filename[:(- len((ends_in % data_set.split('_')[0])))] + 'leftImg8bit.png') image['seg_file_name'] = (filename[:(- len((ends_in % data_set.split('_')[0])))] + ('%s_instanceIds.png' % data_set.split('_')[0])) images.append(image) fullname = os.path.join(root, image['seg_file_name']) objects = cs.instances2dict_with_polygons([fullname], verbose=False)[fullname] for object_cls in objects: if (object_cls not in category_instancesonly): continue for obj in objects[object_cls]: if (obj['contours'] == []): print('Warning: empty contours.') continue len_p = [len(p) for p in obj['contours']] if (min(len_p) <= 4): print('Warning: invalid contours.') continue ann = {} ann['id'] = ann_id ann_id += 1 ann['image_id'] = image['id'] ann['segmentation'] = obj['contours'] if (object_cls not in category_dict): category_dict[object_cls] = cat_id cat_id += 1 ann['category_id'] = category_dict[object_cls] ann['iscrowd'] = 0 ann['area'] = obj['pixelCount'] ann['bbox'] = bboxs_util.xyxy_to_xywh(segms_util.polys_to_boxes([ann['segmentation']])).tolist()[0] annotations.append(ann) ann_dict['images'] = images categories = [{'id': category_dict[name], 'name': name} for name in category_dict] ann_dict['categories'] = categories ann_dict['annotations'] = annotations print(('Num categories: %s' % len(categories))) print(('Num images: %s' % len(images))) print(('Num annotations: %s' % len(annotations))) with open(os.path.join(out_dir, (json_name % data_set)), 'w') as outfile: outfile.write(json.dumps(ann_dict))
def get_tasks(args, task_names, max_seq_len): tasks = [] for name in task_names: assert (name in NAME2INFO), 'Task not found!' task = NAME2INFO[name][0](args, NAME2INFO[name][1], max_seq_len, name) tasks.append(task) logging.info('\tFinished loading tasks: %s.', ' '.join([task.name for task in tasks])) return tasks
def EpochModelCheckpoint(*args, **kwargs): callback = pl.callbacks.ModelCheckpoint(*args, **kwargs) _on_validation_end = callback.on_validation_end _on_save_checkpoint = callback.on_save_checkpoint def on_validation_end(*args, **kwargs): return def on_save_checkpoint(trainer, module, *args): if hasattr(module, 'config'): with fsspec.open(os.path.join(callback.dirpath, 'config.json'), 'w+') as f: json.dump(module.config.asdict(), f) f.flush() return _on_save_checkpoint(trainer, module, *args) def on_train_epoch_end(trainer, module, _): _on_validation_end(trainer, module) callback.on_save_checkpoint = on_save_checkpoint callback.on_validation_end = on_validation_end callback.on_train_epoch_end = on_train_epoch_end return callback
def conv2d(in_channels, out_channels, kernel_size=3, stride=1, dilation=1, groups=1): return nn.Sequential(nn.Conv2d(in_channels, out_channels, kernel_size=kernel_size, stride=stride, padding=dilation, dilation=dilation, bias=False, groups=groups), nn.BatchNorm2d(out_channels), nn.LeakyReLU(0.2, inplace=True))
class IP(nn.Module): def __init__(self, args): super(IP, self).__init__() self.model_recognition = LightCNN_29Layers_v2(num_classes=346) self.model_recognition = torch.nn.DataParallel(self.model_recognition).cuda() checkpoint = torch.load('lightCNN_pretrain.pth.tar') self.model_recognition.load_state_dict(checkpoint['state_dict']) self.submean = common.MeanShift(args.rgb_range) for p in self.parameters(): p.requires_grad = False def forward(self, sr, hr): def _forward(x): x = self.submean(x) x = self.model_recognition(x) return x (out_sr, feat_sr) = _forward(sr[0]) with torch.no_grad(): (out_hr, feat_hr) = _forward(hr[0].detach()) loss = (F.mse_loss(feat_sr, feat_hr) + F.mse_loss(out_sr, out_hr)) return loss
def analyse_obs_spaces(env_obs_space, model_obs_space): entries_only_in_env = [] entries_only_in_model = [] entries_in_both_but_with_different_values = [] for key in env_obs_space.keys(): if (key not in model_obs_space.keys()): entries_only_in_env.append(key) elif ((key in model_obs_space.keys()) and (model_obs_space[key] != env_obs_space[key])): entries_in_both_but_with_different_values.append((key, model_obs_space[key], env_obs_space[key])) for key in model_obs_space.keys(): if (key not in env_obs_space.keys()): entries_only_in_model.append(key) print('[Error] Your model failed to load due to incompatible observation spaces!') print('This is a list of all elements that were only found in the env observation space:') print(('#' * 15)) for entry in entries_only_in_env: print(entry) print(('#' * 15)) print("This is a list of all elements that were only found in the model's observation space:") print(('#' * 15)) for entry in entries_only_in_model: print(entry) print(('#' * 15)) print('This is a list of all elements that were found in both, but have mismatching definitions:') print(('#' * 15)) for entry in entries_in_both_but_with_different_values: print('Entry: ', entry[0]) print('Value in model: ', entry[1]) print('Value in env: ', entry[2]) print(('#' * 15))
class BufferList(torch.nn.Module): def __init__(self, buffers): super(BufferList, self).__init__() self.buffers = [] for (i, b) in enumerate(buffers): name = '_buffer_{}'.format(i) self.register_buffer(name, b) self.buffers.append(getattr(self, name)) def __repr__(self): return str(self.buffers) def __getitem__(self, i): return self.buffers[i] def dtype(self): return self.buffers[0].dtype
def plot_acc(history): plt.plot(history.history['acc']) plt.plot(history.history['val_acc']) plt.title('Model accuracy') plt.ylabel('Accuracy') plt.xlabel('Epoch') plt.legend(['Train', 'Test'], loc=0)
def structure_encoding(atoms): enc = [0 for _ in range(55)] for atom in atoms: enc[atom.GetAtomicNum()] += 1 return enc
class VKITTI(Dataset): def __init__(self, data_dir_root, do_kb_crop=True): import glob self.image_files = glob.glob(os.path.join(data_dir_root, 'test_color', '*.png')) self.depth_files = [r.replace('test_color', 'test_depth') for r in self.image_files] self.do_kb_crop = True self.transform = ToTensor() def __getitem__(self, idx): image_path = self.image_files[idx] depth_path = self.depth_files[idx] image = Image.open(image_path) depth = Image.open(depth_path) depth = cv2.imread(depth_path, (cv2.IMREAD_ANYCOLOR | cv2.IMREAD_ANYDEPTH)) print('dpeth min max', depth.min(), depth.max()) if (self.do_kb_crop and False): height = image.height width = image.width top_margin = int((height - 352)) left_margin = int(((width - 1216) / 2)) depth = depth.crop((left_margin, top_margin, (left_margin + 1216), (top_margin + 352))) image = image.crop((left_margin, top_margin, (left_margin + 1216), (top_margin + 352))) image = (np.asarray(image, dtype=np.float32) / 255.0) depth = depth[(..., None)] sample = dict(image=image, depth=depth) sample = self.transform(sample) if (idx == 0): print(sample['image'].shape) return sample def __len__(self): return len(self.image_files)
class DAPPM(nn.Module): def __init__(self, inplanes, branch_planes, outplanes, BatchNorm=nn.BatchNorm2d): super(DAPPM, self).__init__() bn_mom = 0.1 self.scale1 = nn.Sequential(nn.AvgPool2d(kernel_size=5, stride=2, padding=2), BatchNorm(inplanes, momentum=bn_mom), nn.ReLU(inplace=True), nn.Conv2d(inplanes, branch_planes, kernel_size=1, bias=False)) self.scale2 = nn.Sequential(nn.AvgPool2d(kernel_size=9, stride=4, padding=4), BatchNorm(inplanes, momentum=bn_mom), nn.ReLU(inplace=True), nn.Conv2d(inplanes, branch_planes, kernel_size=1, bias=False)) self.scale3 = nn.Sequential(nn.AvgPool2d(kernel_size=17, stride=8, padding=8), BatchNorm(inplanes, momentum=bn_mom), nn.ReLU(inplace=True), nn.Conv2d(inplanes, branch_planes, kernel_size=1, bias=False)) self.scale4 = nn.Sequential(nn.AdaptiveAvgPool2d((1, 1)), BatchNorm(inplanes, momentum=bn_mom), nn.ReLU(inplace=True), nn.Conv2d(inplanes, branch_planes, kernel_size=1, bias=False)) self.scale0 = nn.Sequential(BatchNorm(inplanes, momentum=bn_mom), nn.ReLU(inplace=True), nn.Conv2d(inplanes, branch_planes, kernel_size=1, bias=False)) self.process1 = nn.Sequential(BatchNorm(branch_planes, momentum=bn_mom), nn.ReLU(inplace=True), nn.Conv2d(branch_planes, branch_planes, kernel_size=3, padding=1, bias=False)) self.process2 = nn.Sequential(BatchNorm(branch_planes, momentum=bn_mom), nn.ReLU(inplace=True), nn.Conv2d(branch_planes, branch_planes, kernel_size=3, padding=1, bias=False)) self.process3 = nn.Sequential(BatchNorm(branch_planes, momentum=bn_mom), nn.ReLU(inplace=True), nn.Conv2d(branch_planes, branch_planes, kernel_size=3, padding=1, bias=False)) self.process4 = nn.Sequential(BatchNorm(branch_planes, momentum=bn_mom), nn.ReLU(inplace=True), nn.Conv2d(branch_planes, branch_planes, kernel_size=3, padding=1, bias=False)) self.compression = nn.Sequential(BatchNorm((branch_planes * 5), momentum=bn_mom), nn.ReLU(inplace=True), nn.Conv2d((branch_planes * 5), outplanes, kernel_size=1, bias=False)) self.shortcut = nn.Sequential(BatchNorm(inplanes, momentum=bn_mom), nn.ReLU(inplace=True), nn.Conv2d(inplanes, outplanes, kernel_size=1, bias=False)) def forward(self, x): width = x.shape[(- 1)] height = x.shape[(- 2)] x_list = [] x_list.append(self.scale0(x)) x_list.append(self.process1((F.interpolate(self.scale1(x), size=[height, width], mode='bilinear', align_corners=algc) + x_list[0]))) x_list.append(self.process2((F.interpolate(self.scale2(x), size=[height, width], mode='bilinear', align_corners=algc) + x_list[1]))) x_list.append(self.process3((F.interpolate(self.scale3(x), size=[height, width], mode='bilinear', align_corners=algc) + x_list[2]))) x_list.append(self.process4((F.interpolate(self.scale4(x), size=[height, width], mode='bilinear', align_corners=algc) + x_list[3]))) out = (self.compression(torch.cat(x_list, 1)) + self.shortcut(x)) return out
def get_tl_line_values_gt(line, LTRB=True, withTranscription=False, withConfidence=False, imWidth=0, imHeight=0): confidence = 0.0 transcription = '' points = [] if LTRB: raise Exception('Not implemented.') else: if (withTranscription and withConfidence): raise 'not implemented' elif withConfidence: raise 'not implemented' elif withTranscription: ptr = line.strip().split(',####') cors = ptr[0].split(',') recs = ptr[1].strip() assert ((len(cors) % 2) == 0), 'num cors should be even.' try: points = [float(ic) for ic in cors[:]] except Exception as e: raise e else: raise 'not implemented' validate_clockwise_points(points) if ((imWidth > 0) and (imHeight > 0)): for ip in range(0, len(points), 2): validate_point_inside_bounds(points[ip], points[(ip + 1)], imWidth, imHeight) if withConfidence: try: confidence = 1.0 except ValueError: raise Exception('Confidence value must be a float') if withTranscription: transcription = recs m2 = re.match('^\\s*\\"(.*)\\"\\s*$', transcription) if (m2 != None): transcription = m2.group(1).replace('\\\\', '\\').replace('\\"', '"') return (points, confidence, transcription)
_grad() def concat_all_gather(tensor): if (get_mpi_size() == 1): return tensor if (not is_hvd_initialized()): tensors_gather = [torch.ones_like(tensor) for _ in range(get_mpi_size())] torch.distributed.all_gather(tensors_gather, tensor, async_op=False) output = torch.cat(tensors_gather, dim=0) return output else: import horovod as hvd output = hvd.torch.allgather(tensor) return output
class LinearWarmUpScheduler(LRScheduler): def __init__(self, optimizer, warmup, total_steps, last_epoch=(- 1)): self.warmup = warmup self.total_steps = total_steps super(LinearWarmUpScheduler, self).__init__(optimizer, last_epoch) def get_lr(self): progress = (self.last_epoch / self.total_steps) if (progress < self.warmup): return [((base_lr * progress) / self.warmup) for base_lr in self.base_lrs] else: return [(base_lr * max(((progress - 1.0) / (self.warmup - 1.0)), 0.0)) for base_lr in self.base_lrs]
class DesiTileFileHandler(DesiDataFileHandler): def __init__(self, analysis_type, use_non_coadded_spectra, logger, input_directory): self.input_directory = input_directory super().__init__(analysis_type, use_non_coadded_spectra, logger) def read_file(self, filename, catalogue): try: hdul = fitsio.FITS(filename) except IOError: self.logger.warning(f'Error reading file {filename}. Ignoring file') return ({}, 0) fibermap = hdul['FIBERMAP'].read() ra = fibermap['TARGET_RA'] dec = fibermap['TARGET_DEC'] tile_spec = fibermap['TILEID'][0] try: if ('cumulative' in self.input_directory): night_spec = int(filename.split('thru')[(- 1)].split('.')[0]) else: night_spec = int(filename.split('-')[(- 1)].split('.')[0]) except ValueError: self.logger.warning(f'In file {filename}, error reading night. Ignoring file') return ({}, 0) colors = ['B', 'R', 'Z'] ra = np.radians(ra) dec = np.radians(dec) petal_spec = fibermap['PETAL_LOC'][0] spectrographs_data = {} for color in colors: try: spec = {} spec['WAVELENGTH'] = hdul[f'{color}_WAVELENGTH'].read() spec['FLUX'] = hdul[f'{color}_FLUX'].read() spec['IVAR'] = (hdul[f'{color}_IVAR'].read() * (hdul[f'{color}_MASK'].read() == 0)) if (self.analysis_type == 'PK 1D'): if (f'{color}_RESOLUTION' in hdul): spec['RESO'] = hdul[f'{color}_RESOLUTION'].read() else: raise DataError(f"Error while reading {color} band from {{filename}}. Analysis type is 'PK 1D', but file does not contain HDU '{color}_RESOLUTION' ") w = (np.isnan(spec['FLUX']) | np.isnan(spec['IVAR'])) for key in ['FLUX', 'IVAR']: spec[key][w] = 0.0 spectrographs_data[color] = spec except OSError: self.logger.warning(f'Error while reading {color} band from {filename}.Ignoring color.') hdul.close() if ('cumulative' in self.input_directory): select = (((catalogue['TILEID'] == tile_spec) & (catalogue['PETAL_LOC'] == petal_spec)) & (catalogue['LASTNIGHT'] == night_spec)) else: select = (((catalogue['TILEID'] == tile_spec) & (catalogue['PETAL_LOC'] == petal_spec)) & (catalogue['NIGHT'] == night_spec)) self.logger.progress(f'This is tile {tile_spec}, petal {petal_spec}, night {night_spec}') (forests_by_targetid, num_data) = self.format_data(catalogue[select], spectrographs_data, fibermap['TARGETID']) return (forests_by_targetid, num_data)
def seed_all(seed): random.seed(seed) np.random.seed(seed) torch.manual_seed(seed) torch.cuda.manual_seed_all(seed)
def clip_norm(g, c, n): if (c > 0): g = T.switch(T.ge(n, c), ((g * c) / n), g) return g
def sim_gcn(reps): n = reps.shape[0] sim_mat = np.zeros([n, n]) return cosine_similarity(reps, reps)
class OpWeights(JsonSerializer): def __init__(self, weights_data: dict): super().__init__() self.dtype: str = weights_data.get('dtype', None) self.granularity: str = weights_data.get('granularity', None)
def ProcessLineForNonScoredWords(a): global num_lines, num_correct_lines, ref_change_stats try: assert (len(a) == 8) num_lines += 1 duration = a[3] hyp_word = a[4] ref_word = a[6] edit_type = a[7] if (edit_type == 'ins'): assert (ref_word == '<eps>') if (hyp_word in non_scored_words): ref_change_stats[((ref_word + ' -> ') + hyp_word)] += 1 ref_word = hyp_word edit_type = 'fix' elif (edit_type == 'del'): assert ((hyp_word == '<eps>') and (float(duration) == 0.0)) if (ref_word in non_scored_words): ref_change_stats[((ref_word + ' -> ') + hyp_word)] += 1 return [] elif (edit_type == 'sub'): assert (hyp_word != '<eps>') if ((hyp_word in non_scored_words) and (ref_word in non_scored_words)): ref_change_stats[((ref_word + ' -> ') + hyp_word)] += 1 ref_word = hyp_word edit_type = 'fix' else: assert ((edit_type == 'cor') or (edit_type == 'sil')) num_correct_lines += 1 a[4] = hyp_word a[6] = ref_word a[7] = edit_type return a except Exception: logger.error('bad line in ctm-edits input: {0}'.format(a)) raise RuntimeError
class ImbalanceSVHN(torchvision.datasets.SVHN): cls_num = 10 def __init__(self, root, imb_type='exp', imb_factor=0.01, rand_number=0, split='train', transform=None, target_transform=None, download=False): super(ImbalanceSVHN, self).__init__(root, split, transform, target_transform, download) np.random.seed(rand_number) img_num_list = self.get_img_num_per_cls(self.cls_num, imb_type, imb_factor) self.gen_imbalanced_data(img_num_list) def get_img_num_per_cls(self, cls_num, imb_type, imb_factor): img_max = 1000 img_num_per_cls = [] if (imb_type == 'exp'): for cls_idx in range(cls_num): num = (img_max * (imb_factor ** (cls_idx / (cls_num - 1.0)))) img_num_per_cls.append(int(num)) elif (imb_type == 'step'): for cls_idx in range((cls_num // 2)): img_num_per_cls.append(int(img_max)) for cls_idx in range((cls_num // 2)): img_num_per_cls.append(int((img_max * imb_factor))) else: img_num_per_cls.extend(([int(img_max)] * cls_num)) return img_num_per_cls def gen_imbalanced_data(self, img_num_per_cls): new_data = [] new_targets = [] targets_np = np.array(self.labels, dtype=np.int64) classes = np.unique(targets_np) classes = np.concatenate([classes[1:], classes[:1]], axis=0) self.num_per_cls_dict = dict() for (the_class, the_img_num) in zip(classes, img_num_per_cls): self.num_per_cls_dict[the_class] = the_img_num idx = np.where((targets_np == the_class))[0] print(f'Class {the_class}: {len(idx)}') np.random.shuffle(idx) selec_idx = idx[:the_img_num] new_data.append(self.data[(selec_idx, ...)]) new_targets.extend(([the_class] * the_img_num)) new_data = np.vstack(new_data) self.data = new_data self.labels = new_targets assert (new_data.shape[0] == len(new_targets)), 'Length of data & labels do not match!' def get_cls_num_list(self): cls_num_list = [] for i in range(self.cls_num): cls_num_list.append(self.num_per_cls_dict[i]) return cls_num_list
class VideoRenderer(mp.Process): def __init__(self, display=False, verbose=0, verbose_size=None, output_crop=False, resolution=256, crop_scale=1.2, encoder_codec='mp4v', separate_process=False): super(VideoRenderer, self).__init__() self._display = display self._verbose = verbose self._verbose_size = verbose_size self._output_crop = output_crop self._resolution = resolution self._crop_scale = crop_scale self._running = True self._input_queue = mp.Queue() self._reply_queue = mp.Queue() self._fourcc = cv2.VideoWriter_fourcc(*encoder_codec) self._separate_process = separate_process self._in_vid = None self._out_vid = None self._seq = None self._in_vid_path = None self._total_frames = None self._frame_count = 0 def init(self, in_vid_path, seq, out_vid_path=None, **kwargs): if self._separate_process: self._input_queue.put([in_vid_path, seq, out_vid_path, kwargs]) else: self._init_task(in_vid_path, seq, out_vid_path, kwargs) def write(self, *args): if self._separate_process: self._input_queue.put([a.cpu() for a in args]) else: self._write_batch([a.cpu() for a in args]) def finalize(self): if self._separate_process: self._input_queue.put(True) else: self._finalize_task() def wait_until_finished(self): if self._separate_process: return self._reply_queue.get() else: return True def on_render(self, *args): return tensor2bgr(args[0]) def start(self): if self._separate_process: super(VideoRenderer, self).start() def kill(self): if self._separate_process: super(VideoRenderer, self).kill() def run(self): while self._running: task = self._input_queue.get() if (self._in_vid is None): self._init_task(*task[:3], task[3]) continue if isinstance(task, bool): self._finalize_task() self._reply_queue.put(True) continue self._write_batch(task) def _render(self, render_bgr, full_frame_bgr=None, bbox=None): if ((self._verbose == 0) and (not self._output_crop) and (full_frame_bgr is not None)): render_bgr = crop2img(full_frame_bgr, render_bgr, bbox) if (self._out_vid is not None): self._out_vid.write(render_bgr) if self._display: cv2.imshow('render', render_bgr) if ((cv2.waitKey(1) & 255) == ord('q')): self._running = False def _init_task(self, in_vid_path, seq, out_vid_path, additional_attributes): (self._in_vid_path, self._seq) = (in_vid_path, seq) self._frame_count = 0 for (attr_name, attr_val) in additional_attributes.items(): setattr(self, attr_name, attr_val) self._in_vid = cv2.VideoCapture(self._in_vid_path) assert self._in_vid.isOpened(), f'Failed to open video: "{self._in_vid_path}"' in_total_frames = int(self._in_vid.get(cv2.CAP_PROP_FRAME_COUNT)) fps = self._in_vid.get(cv2.CAP_PROP_FPS) in_vid_width = int(self._in_vid.get(cv2.CAP_PROP_FRAME_WIDTH)) in_vid_height = int(self._in_vid.get(cv2.CAP_PROP_FRAME_HEIGHT)) self._total_frames = (in_total_frames if (self._verbose == 0) else len(self._seq)) if (out_vid_path is not None): out_size = (in_vid_width, in_vid_height) if ((self._verbose <= 0) and self._output_crop): out_size = (self._resolution, self._resolution) elif (self._verbose_size is not None): out_size = self._verbose_size self._out_vid = cv2.VideoWriter(out_vid_path, self._fourcc, fps, out_size) if (self._verbose == 0): for i in range(self._seq.start_index): (ret, frame_bgr) = self._in_vid.read() assert (frame_bgr is not None), f'Failed to read frame {i} from input video: "{self._in_vid_path}"' self._render(frame_bgr) self._frame_count += 1 def _write_batch(self, tensors): batch_size = tensors[0].shape[0] for b in range(batch_size): (full_frame_bgr, bbox) = (None, None) if ((self._verbose == 0) and (not self._output_crop)): (ret, full_frame_bgr) = self._in_vid.read() assert (full_frame_bgr is not None), f'Failed to read frame {self._frame_count} from input video: "{self._in_vid_path}"' det = self._seq[(self._frame_count - self._seq.start_index)] bbox = np.concatenate((det[:2], (det[2:] - det[:2]))) bbox = scale_bbox(bbox, self._crop_scale) render_bgr = self.on_render(*[t[b] for t in tensors]) self._render(render_bgr, full_frame_bgr, bbox) self._frame_count += 1 def _finalize_task(self): if ((self._verbose == 0) and (self._frame_count >= (self._seq.start_index + len(self._seq)))): for i in range((self._seq.start_index + len(self._seq)), self._total_frames): (ret, frame_bgr) = self._in_vid.read() assert (frame_bgr is not None), f'Failed to read frame {i} from input video: "{self._in_vid_path}"' self._render(frame_bgr) self._frame_count += 1 self._in_vid.release() self._out_vid.release() self._in_vid = None self._out_vid = None self._seq = None self._in_vid_path = None self._total_frames = None self._frame_count = 0
def sample_ddpg_params(trial): gamma = trial.suggest_categorical('gamma', [0.9, 0.95, 0.98, 0.99, 0.995, 0.999, 0.9999]) learning_rate = trial.suggest_loguniform('lr', 1e-05, 1) batch_size = trial.suggest_categorical('batch_size', [16, 32, 64, 128, 256]) buffer_size = trial.suggest_categorical('memory_limit', [int(10000.0), int(100000.0), int(1000000.0)]) noise_type = trial.suggest_categorical('noise_type', ['ornstein-uhlenbeck', 'normal', 'adaptive-param']) noise_std = trial.suggest_uniform('noise_std', 0, 1) normalize_observations = trial.suggest_categorical('normalize_observations', [True, False]) normalize_returns = trial.suggest_categorical('normalize_returns', [True, False]) hyperparams = {'gamma': gamma, 'actor_lr': learning_rate, 'critic_lr': learning_rate, 'batch_size': batch_size, 'memory_limit': buffer_size, 'normalize_observations': normalize_observations, 'normalize_returns': normalize_returns} if (noise_type == 'adaptive-param'): hyperparams['param_noise'] = AdaptiveParamNoiseSpec(initial_stddev=noise_std, desired_action_stddev=noise_std) hyperparams['policy_kwargs'] = dict(layer_norm=True) elif (noise_type == 'normal'): hyperparams['action_noise'] = NormalActionNoise(mean=np.zeros(trial.n_actions), sigma=(noise_std * np.ones(trial.n_actions))) elif (noise_type == 'ornstein-uhlenbeck'): hyperparams['action_noise'] = OrnsteinUhlenbeckActionNoise(mean=np.zeros(trial.n_actions), sigma=(noise_std * np.ones(trial.n_actions))) return hyperparams
class WarmupMultiStepLR(torch.optim.lr_scheduler._LRScheduler): def __init__(self, optimizer, milestones, gamma=0.1, warmup_factor=(1.0 / 3), warmup_iters=500, warmup_method='linear', last_epoch=(- 1)): if (not (list(milestones) == sorted(milestones))): raise ValueError('Milestones should be a list of increasing integers. Got {}', milestones) if (warmup_method not in ('constant', 'linear', 'none')): raise ValueError("Only 'constant' or 'linear' warmup_method acceptedgot {}".format(warmup_method)) self.milestones = milestones self.gamma = gamma self.warmup_factor = warmup_factor self.warmup_iters = warmup_iters self.warmup_method = warmup_method super(WarmupMultiStepLR, self).__init__(optimizer, last_epoch) def get_lr(self): warmup_factor = 1 if (self.last_epoch < self.warmup_iters): if (self.warmup_method == 'linear'): warmup_factor = ((1 + self.last_epoch) / self.warmup_iters) elif (self.warmup_method == 'constant'): warmup_factor = 1 return [((base_lr * warmup_factor) * (self.gamma ** bisect_right(self.milestones, self.last_epoch))) for base_lr in self.base_lrs]
def _gen_mnasnet_b1(variant, channel_multiplier=1.0, pretrained=False, **kwargs): arch_def = [['ds_r1_k3_s1_c16_noskip'], ['ir_r3_k3_s2_e3_c24'], ['ir_r3_k5_s2_e3_c40'], ['ir_r3_k5_s2_e6_c80'], ['ir_r2_k3_s1_e6_c96'], ['ir_r4_k5_s2_e6_c192'], ['ir_r1_k3_s1_e6_c320_noskip']] model_kwargs = dict(block_args=decode_arch_def(arch_def), stem_size=32, channel_multiplier=channel_multiplier, norm_kwargs=resolve_bn_args(kwargs), **kwargs) model = _create_effnet(variant, pretrained, **model_kwargs) return model
def dump_results(results): with open(result_file, 'wb') as f: pickle.dump(dict(results), f)
class TemperatureTanh(nn.Module): def __init__(self, temperature: float=1.0) -> None: super().__init__() assert (temperature != 0.0), 'temperature must be nonzero.' self._T = temperature self.tanh = torch.nn.Tanh() def forward(self, x: Tensor) -> Tensor: return self.tanh((x / self._T))
class Args(Tap): ckpts_dirs: List[str] split_type: Literal[('random', 'scaffold')] num_folds: int = 10
(kernels.Kernel, TensorLike, TensorLike, TensorLike) def _exact_fallback(kern: kernels.Kernel, Z: TensorLike, u: TensorLike, f: TensorLike, *, L: TensorLike=None, diag: TensorLike=None, basis: AbstractBasis=None, **kwargs): u_shape = tuple(u.shape) f_shape = tuple(f.shape) assert (u_shape[(- 1)] == 1), 'Recieved multiple output features' assert (u_shape == f_shape[(- len(u_shape)):]), 'Incompatible shapes detected' if (basis is None): basis = kernel_basis(kern, centers=Z) if (diag is None): diag = default_jitter() if isinstance(diag, float): diag = tf.convert_to_tensor(diag, dtype=f.dtype) diag = tf.expand_dims(diag, axis=(- 1)) err = (u - f) err -= (tf.sqrt(diag) * tf.random.normal(err.shape, dtype=err.dtype)) if (L is None): if isinstance(Z, inducing_variables.InducingVariables): K = covariances.Kuu(Z, kern, jitter=0.0) else: K = kern(Z, full_cov=True) K = tf.linalg.set_diag(K, (tf.linalg.diag_part(K) + diag[(..., 0)])) L = tf.linalg.cholesky(K) weights = tf.linalg.adjoint(tf.linalg.cholesky_solve(L, err)) return DenseSampler(basis=basis, weights=weights, **kwargs)
class CompositeAudioTransform(AudioTransform): def _from_config_dict(cls, transform_type, get_audio_transform, composite_cls, config=None, return_empty=False): _config = ({} if (config is None) else config) _transforms = _config.get(f'{transform_type}_transforms') if (_transforms is None): if return_empty: _transforms = [] else: return None transforms = [get_audio_transform(_t).from_config_dict(_config.get(_t)) for _t in _transforms] return composite_cls(transforms) def __init__(self, transforms): self.transforms = [t for t in transforms if (t is not None)] def __call__(self, x): for t in self.transforms: x = t(x) return x def __repr__(self): format_string = (([(self.__class__.__name__ + '(')] + [f' {t.__repr__()}' for t in self.transforms]) + [')']) return '\n'.join(format_string)
class ColumnSample(): basedir = prev_dir(os.getcwd()) def __init__(self, sampletype, column, directory, settings): self.sampletype = sampletype self.column = column self.directory = directory self.settings = settings self.basedir = basedir def featurize(self): print(self.sampletype) if (self.sampletype == 'audio'): (features_, labels) = audio_featurize_columns(self.column, self.directory, self.settings, self.basedir) elif (self.sampletype == 'text'): (features_, labels) = text_featurize_columns(self.column, self.directory, self.settings, self.basedir) elif (self.sampletype == 'image'): (features_, labels) = image_featurize_columns(self.column, self.directory, self.settings, self.basedir) elif (self.sampletype == 'video'): (features_, labels) = video_featurize_columns(self.column, self.directory, self.settings, self.basedir) elif (self.sampletype == 'categorical'): (features_, labels) = category_featurize_columns(self.column, self.directory, self.settings, self.basedir) elif (self.sampletype == 'typedtext'): (features_, labels) = typedtext_featurize_columns(self.column, self.directory, self.settings, self.basedir) elif (self.sampletype == 'numerical'): (features_, labels) = numerical_featurize_columns(self.column, self.directory, self.settings, self.basedir) self.features = features_ self.labels = labels
class XLNetTokenizer(PreTrainedTokenizer): vocab_files_names = VOCAB_FILES_NAMES pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES def __init__(self, vocab_file, max_len=None, do_lower_case=False, remove_space=True, keep_accents=False, bos_token='<s>', eos_token='</s>', unk_token='<unk>', sep_token='<sep>', pad_token='<pad>', cls_token='<cls>', mask_token='<mask>', additional_special_tokens=['<eop>', '<eod>'], **kwargs): super(XLNetTokenizer, self).__init__(bos_token=bos_token, eos_token=eos_token, unk_token=unk_token, sep_token=sep_token, pad_token=pad_token, cls_token=cls_token, mask_token=mask_token, additional_special_tokens=additional_special_tokens, **kwargs) try: import sentencepiece as spm except ImportError: logger.warning('You need to install SentencePiece to use XLNetTokenizer: install sentencepiece') self.do_lower_case = do_lower_case self.remove_space = remove_space self.keep_accents = keep_accents self.vocab_file = vocab_file self.sp_model = spm.SentencePieceProcessor() self.sp_model.Load(vocab_file) def vocab_size(self): return len(self.sp_model) def __getstate__(self): state = self.__dict__.copy() state['sp_model'] = None return state def __setstate__(self, d): self.__dict__ = d try: import sentencepiece as spm except ImportError: logger.warning('You need to install SentencePiece to use XLNetTokenizer: install sentencepiece') self.sp_model = spm.SentencePieceProcessor() self.sp_model.Load(self.vocab_file) def preprocess_text(self, inputs): if self.remove_space: outputs = ' '.join(inputs.strip().split()) else: outputs = inputs outputs = outputs.replace('``', '"').replace("''", '"') if (six.PY2 and isinstance(outputs, str)): outputs = outputs.decode('utf-8') if (not self.keep_accents): outputs = unicodedata.normalize('NFKD', outputs) outputs = ''.join([c for c in outputs if (not unicodedata.combining(c))]) if self.do_lower_case: outputs = outputs.lower() return outputs def _tokenize(self, text, return_unicode=True, sample=False): text = self.preprocess_text(text) if (six.PY2 and isinstance(text, unicode)): text = text.encode('utf-8') if (not sample): pieces = self.sp_model.EncodeAsPieces(text) else: pieces = self.sp_model.SampleEncodeAsPieces(text, 64, 0.1) new_pieces = [] for piece in pieces: if ((len(piece) > 1) and (piece[(- 1)] == ',') and piece[(- 2)].isdigit()): cur_pieces = self.sp_model.EncodeAsPieces(piece[:(- 1)].replace(SPIECE_UNDERLINE, '')) if ((piece[0] != SPIECE_UNDERLINE) and (cur_pieces[0][0] == SPIECE_UNDERLINE)): if (len(cur_pieces[0]) == 1): cur_pieces = cur_pieces[1:] else: cur_pieces[0] = cur_pieces[0][1:] cur_pieces.append(piece[(- 1)]) new_pieces.extend(cur_pieces) else: new_pieces.append(piece) if (six.PY2 and return_unicode): ret_pieces = [] for piece in new_pieces: if isinstance(piece, str): piece = piece.decode('utf-8') ret_pieces.append(piece) new_pieces = ret_pieces return new_pieces def _convert_token_to_id(self, token): return self.sp_model.PieceToId(token) def _convert_id_to_token(self, index, return_unicode=True): token = self.sp_model.IdToPiece(index) if (six.PY2 and return_unicode and isinstance(token, str)): token = token.decode('utf-8') return token def convert_tokens_to_string(self, tokens): out_string = ''.join(tokens).replace(SPIECE_UNDERLINE, ' ').strip() return out_string def add_special_tokens_single_sentence(self, token_ids): logger.warning('No method was defined for special tokens and single sentence streams in XLNet. Returning token_ids') return token_ids def add_special_tokens_sentences_pair(self, *token_ids): sep = [self._convert_token_to_id(self.sep_token)] cls = [self._convert_token_to_id(self.cls_token)] return ((((token_ids[0] + sep) + token_ids[1]) + sep) + cls) def save_vocabulary(self, save_directory): if (not os.path.isdir(save_directory)): logger.error('Vocabulary path ({}) should be a directory'.format(save_directory)) return out_vocab_file = os.path.join(save_directory, VOCAB_FILES_NAMES['vocab_file']) if (os.path.abspath(self.vocab_file) != os.path.abspath(out_vocab_file)): copyfile(self.vocab_file, out_vocab_file) return (out_vocab_file,)
def load_config(args=None, config_file=None, overwrite_fairseq=False): if (args is not None): config_file = args.taskconfig config = recursive_config(config_file) if (config.dataset.subsampling is not None): batch_size = (config.fairseq.dataset.batch_size // config.dataset.subsampling) print('adjusting batch_size to {} due to subsampling {}.'.format(batch_size, config.dataset.subsampling)) config.fairseq.dataset.batch_size = batch_size is_test = ((config.dataset.split is not None) and (config.dataset.split == 'test')) if (not is_test): if ((config.fairseq.checkpoint is None) or (config.fairseq.checkpoint.save_dir is None)): raise ValueError('fairseq save_dir or save_path must be specified.') save_dir = config.fairseq.checkpoint.save_dir os.makedirs(save_dir, exist_ok=True) if (config.fairseq.common.tensorboard_logdir is not None): tb_run_dir = suffix_rundir(save_dir, config.fairseq.common.tensorboard_logdir) config.fairseq.common.tensorboard_logdir = tb_run_dir print('update tensorboard_logdir as', config.fairseq.common.tensorboard_logdir) os.makedirs(save_dir, exist_ok=True) OmegaConf.save(config=config, f=os.path.join(save_dir, 'config.yaml')) if (overwrite_fairseq and (config.fairseq is not None) and (args is not None)): for group in config.fairseq: for field in config.fairseq[group]: print(('overwrite args.' + field), 'as', config.fairseq[group][field]) setattr(args, field, config.fairseq[group][field]) return config
def prune_outside_window(keypoints, window, scope=None): with tf.name_scope(scope, 'PruneOutsideWindow'): (y, x) = tf.split(value=keypoints, num_or_size_splits=2, axis=2) (win_y_min, win_x_min, win_y_max, win_x_max) = tf.unstack(window) valid_indices = tf.logical_and(tf.logical_and((y >= win_y_min), (y <= win_y_max)), tf.logical_and((x >= win_x_min), (x <= win_x_max))) new_y = tf.where(valid_indices, y, (np.nan * tf.ones_like(y))) new_x = tf.where(valid_indices, x, (np.nan * tf.ones_like(x))) new_keypoints = tf.concat([new_y, new_x], 2) return new_keypoints
def print_results(results_path): with open(results_path, 'r') as fp: results = json.load(fp) print('Flags:') for (k, v) in sorted(results['flags'].items()): print('\t{}: {}'.format(k, v)) print('HParams:') for (k, v) in sorted(results['hparams'].items()): print('\t{}: {}'.format(k, v)) t = setup_pretty_table(Namespace(**results['flags'])) envs = datasets.get_environments(results['flags']['dataset']) test_env = envs[results['flags']['test_env']] train_names = [(str(env) + '_in_loss') for (i, env) in enumerate(envs) if (i != results['flags']['test_env'])] steps = [key for key in results.keys() if (key not in ['hparams', 'flags'])] for s in steps: t.add_row(((((([s] + ['{:.2f} :: {:.2f}'.format(results[s][(str(e) + '_in_acc')], results[s][(str(e) + '_out_acc')]) for e in envs]) + ['{:.2f}'.format(np.average([results[s][str(e)] for e in train_names]))]) + ['{}'.format('.')]) + ['{}'.format('.')]) + ['{}'.format('.')])) print('\n'.join(t.get_string().splitlines()[(- 2):(- 1)]))
def log_Logistic_256(x, mean, logvar, average=False, reduce=True, dim=None): bin_size = (1.0 / 256.0) scale = torch.exp(logvar) x = (((torch.floor((x / bin_size)) * bin_size) - mean) / scale) cdf_plus = torch.sigmoid((x + (bin_size / scale))) cdf_minus = torch.sigmoid(x) log_logist_256 = (- torch.log(((cdf_plus - cdf_minus) + 1e-07))) if reduce: if average: return torch.mean(log_logist_256, dim) else: return torch.sum(log_logist_256, dim) else: return log_logist_256
def get_net(input_shape, num_output_channels, net_config): num_input_channels = input_shape[0] if (net_config['type'] == 'mlp'): assert (len(input_shape) == 1) return get_mlp(num_input_channels=num_input_channels, hidden_channels=net_config['hidden_channels'], num_output_channels=num_output_channels, activation=get_activation(net_config['activation'])) elif (net_config['type'] == 'resnet'): assert (len(input_shape) == 3) return get_resnet(num_input_channels=num_input_channels, hidden_channels=net_config['hidden_channels'], num_output_channels=num_output_channels) elif (net_config['type'] == 'glow-cnn'): assert (len(input_shape) == 3) return get_glow_cnn(num_input_channels=num_input_channels, num_hidden_channels=net_config['num_hidden_channels'], num_output_channels=num_output_channels, zero_init_output=net_config['zero_init_output']) elif (net_config['type'] == 'constant'): value = torch.full((num_output_channels, *input_shape[1:]), net_config['value'], dtype=torch.get_default_dtype()) return ConstantNetwork(value=value, fixed=net_config['fixed']) elif (net_config['type'] == 'identity'): assert (num_output_channels == num_input_channels) return (lambda x: x) else: assert False, f"Invalid net type {net_config['type']}"
class ImageParameters(object): def __init__(self): self.width_px = 96 self.height_px = 96 self.arcsec_per_pixel = 0.396 self.world_origin = WorldCoordinate(0, 0) self.band_nelec_per_nmgy = [1000.0 for _ in range(5)] def degrees_per_pixel(self): return (self.arcsec_per_pixel / ARCSEC_PER_DEGREE) def get_image_center_world_coordinates(self): width_deg = (self.width_px * self.degrees_per_pixel()) height_deg = (self.height_px * self.degrees_per_pixel()) return self.world_origin.add((height_deg / 2.0), (width_deg / 2.0))
def UsesColor(term, color_env_var, color_flag): SetEnvVar('TERM', term) SetEnvVar(COLOR_ENV_VAR, color_env_var) if (color_flag is None): args = [] else: args = [('--%s=%s' % (COLOR_FLAG, color_flag))] p = gtest_test_utils.Subprocess(([COMMAND] + args)) return ((not p.exited) or p.exit_code)
def ResNet(stack_fn, preact, use_bias, model_name='resnet', include_top=False, weights='imagenet', input_tensor=None, input_shape=None, pooling=None, classes=1000, **kwargs): if (not ((weights in {'imagenet', None}) or os.path.exists(weights))): raise ValueError('The `weights` argument should be either `None` (random initialization), `imagenet` (pre-training on ImageNet), or the path to the weights file to be loaded.') if ((weights == 'imagenet') and include_top and (classes != 1000)): raise ValueError('If using `weights` as `"imagenet"` with `include_top` as true, `classes` should be 1000') input_shape = _obtain_input_shape(input_shape, default_size=224, min_size=32, data_format=backend.image_data_format(), require_flatten=include_top, weights=weights) if (input_tensor is None): img_input = layers.Input(shape=input_shape) elif (not backend.is_keras_tensor(input_tensor)): img_input = layers.Input(tensor=input_tensor, shape=input_shape) else: img_input = input_tensor bn_axis = (3 if (backend.image_data_format() == 'channels_last') else 1) x = layers.ZeroPadding2D(padding=((3, 3), (3, 3)), name='conv1_pad')(img_input) x = layers.Conv2D(64, 7, strides=2, use_bias=use_bias, name='conv1_conv')(x) if (preact is False): x = layers.BatchNormalization(axis=bn_axis, epsilon=1.001e-05, name='conv1_bn')(x) x = layers.Activation('relu', name='conv1_relu')(x) x = layers.ZeroPadding2D(padding=((1, 1), (1, 1)), name='pool1_pad')(x) x = layers.MaxPooling2D(3, strides=2, name='pool1_pool')(x) x = stack_fn(x) if include_top: x = layers.GlobalAveragePooling2D(name='avg_pool')(x) x = layers.Dense(classes, activation='softmax', name='probs')(x) elif (pooling == 'avg'): x = layers.GlobalAveragePooling2D(name='avg_pool')(x) elif (pooling == 'max'): x = layers.GlobalMaxPooling2D(name='max_pool')(x) if (input_tensor is not None): inputs = keras_utils.get_source_inputs(input_tensor) else: inputs = img_input model = models.Model(inputs, x, name=model_name) if ((weights == 'imagenet') and (model_name in WEIGHTS_HASHES)): if include_top: file_name = (model_name + '_weights_tf_dim_ordering_tf_kernels.h5') file_hash = WEIGHTS_HASHES[model_name][0] else: file_name = (model_name + '_weights_tf_dim_ordering_tf_kernels_notop.h5') file_hash = WEIGHTS_HASHES[model_name][1] weights_path = keras_utils.get_file(file_name, (BASE_WEIGHTS_PATH + file_name), cache_subdir='models', file_hash=file_hash) model.load_weights(weights_path) elif (weights is not None): model.load_weights(weights) return model
def _is_torch_dtype(x): import torch if isinstance(x, str): if hasattr(torch, x): x = getattr(torch, x) else: return False return isinstance(x, torch.dtype)
def merge_models(shape_size, models_list, out_model_file, avg=True): from keras.models import load_model, save_model, Model from keras.layers import Input, Average if (not os.path.isfile(out_model_file)): models = [] for m_path in models_list: m = load_model(m_path) models.append(m) inp = Input(shape_size) outs = [] for (i, m) in enumerate(models): m.name += '_{}'.format(i) x = m(inp) outs.append(x) if (avg is True): outs = Average()(outs) final_model = Model(inputs=inp, outputs=outs) print(final_model.summary()) print('Single model memory: {} GB'.format(get_model_memory_usage(1, final_model))) save_model(final_model, out_model_file) else: final_model = load_model(out_model_file) return final_model
class FeatureExtractor(): date_time_format_str = '%Y-%m-%d' def __init__(self): self.user_mnr_normalizer = 0 self.prod_mnr_normalizer = 0 self.product_num_ratings = {} self.product_sum_ratings = {} def MNR(self, data, data_type='user'): feature = {} for (i, d) in data.items(): frequency = {} for t in d: if (t[3] not in frequency): frequency[t[3]] = 1 else: frequency[t[3]] += 1 feature[i] = max(frequency.values()) if (data_type == 'user'): self.user_mnr_normalizer = max(feature.values()) for k in feature.keys(): feature[k] /= self.user_mnr_normalizer else: self.prod_mnr_normalizer = max(feature.values()) for k in feature.keys(): feature[k] /= self.prod_mnr_normalizer return feature def iMNR(self, data, new_data, data_type='user'): feature = {} for (i, d) in new_data.items(): all_d = deepcopy(d) if (i in data): all_d += data[i] frequency = {} for t in all_d: if (t[3] not in frequency): frequency[t[3]] = 1 else: frequency[t[3]] += 1 feature[i] = max(frequency.values()) if (data_type == 'user'): self.user_mnr_normalizer = max(max(feature.values()), self.user_mnr_normalizer) for k in feature.keys(): feature[k] /= self.user_mnr_normalizer else: self.prod_mnr_normalizer = max(max(feature.values()), self.prod_mnr_normalizer) for k in feature.keys(): feature[k] /= self.prod_mnr_normalizer return feature def PR_NR(self, data): feature = {} for (i, d) in data.items(): positives = [1 for t in d if (t[1] > 3)] negatives = [1 for t in d if (t[1] < 3)] feature[i] = ((float(len(positives)) / len(d)), (float(len(negatives)) / len(d))) return feature def iPR_NR(self, data, new_data): feature = {} for (i, d) in new_data.items(): all_d = deepcopy(d) if (i in data): all_d = (all_d + data[i]) positives = [1 for t in all_d if (t[1] > 3)] negatives = [1 for t in all_d if (t[1] < 3)] feature[i] = ((float(len(positives)) / len(all_d)), (float(len(negatives)) / len(all_d))) return feature def avgRD_user(self, user_data, product_data): p_avg = {} for (i, d) in product_data.items(): self.product_num_ratings[i] = len(d) self.product_sum_ratings[i] = np.sum(np.array([t[1] for t in d])) p_avg[i] = np.mean(np.array([t[1] for t in d])) u_avgRD = {} for (i, d) in user_data.items(): u_avgRD[i] = np.mean(np.array([abs((t[1] - p_avg[t[0]])) for t in d])) return u_avgRD def iavgRD_user(self, user_data, new_user_data, product_data, new_product_data): user_involved = set() for (i, d) in new_product_data.items(): if (i in self.product_num_ratings): self.product_num_ratings[i] += len(d) else: self.product_num_ratings[i] = len(d) for t in d: if (i in self.product_sum_ratings): self.product_sum_ratings[i] += t[1] else: self.product_sum_ratings[i] = t[1] user_id = t[0] user_involved.add(user_id) if (i in product_data): for t in product_data[i]: user_id = t[0] user_involved.add(user_id) for (i, d) in new_user_data.items(): assert (i in user_involved) p_avg = {} u_avgRD = {} for user_id in user_involved: all_d = [] if (user_id in new_user_data): all_d += new_user_data[user_id] if (user_id in user_data): all_d += user_data[user_id] for r in all_d: product_id = r[0] if (product_id not in p_avg): p_avg[product_id] = (self.product_sum_ratings[product_id] / self.product_num_ratings[product_id]) u_avgRD[user_id] = np.mean(np.array([abs((r[1] - p_avg[r[0]])) for r in all_d])) return u_avgRD def avgRD_prod(self, product_data): p_avg = {} for (i, d) in product_data.items(): self.product_num_ratings[i] = len(d) self.product_sum_ratings[i] = np.sum(np.array([t[1] for t in d])) p_avg[i] = np.mean(np.array([t[1] for t in d])) p_avgRD = {} for (i, d) in product_data.items(): p_avgRD[i] = np.mean(np.array([abs((t[1] - p_avg[i])) for t in d])) return p_avgRD def iavgRD_prod(self, product_data, new_product_data): for (i, d) in new_product_data.items(): if (i in self.product_num_ratings): self.product_num_ratings[i] += len(d) else: self.product_num_ratings[i] = len(d) for t in d: if (i in self.product_sum_ratings): self.product_sum_ratings[i] += t[1] else: self.product_sum_ratings[i] = t[1] p_avg = {} p_avgRD = {} for (i, d) in new_product_data.items(): all_d = deepcopy(d) if (i in product_data): all_d += product_data[i] p_avg[i] = np.mean(np.array([t[1] for t in all_d])) p_avgRD[i] = np.mean(np.array([abs((t[1] - p_avg[i])) for t in all_d])) return p_avgRD def BST(self, user_data): bst = {} tau = 28.0 for (i, d) in user_data.items(): post_dates = sorted([datetime.strptime(t[3], self.date_time_format_str) for t in d]) delta_days = (post_dates[(- 1)] - post_dates[0]).days if (delta_days > tau): bst[i] = 0.0 else: bst[i] = (1.0 - (delta_days / tau)) return bst def iBST(self, user_data, new_user_data): bst = {} tau = 28.0 for (i, d) in new_user_data.items(): all_d = deepcopy(d) if (i in user_data): all_d += user_data[i] post_dates = sorted([datetime.strptime(t[3], self.date_time_format_str) for t in all_d]) delta_days = (post_dates[(- 1)] - post_dates[0]).days if (delta_days > tau): bst[i] = 0.0 else: bst[i] = (1.0 - (delta_days / tau)) return bst def ERD(self, data): erd = {} for (i, d) in data.items(): ratings = [t[1] for t in d] (h, _) = np.histogram(ratings, bins=np.arange(1, 7)) h = (h / h.sum()) h = h[np.nonzero(h)] erd[i] = ((- h) * np.log2(h)).sum() return erd def iERD(self, data, new_data): erd = {} for (i, d) in new_data.items(): all_d = deepcopy(d) if (i in data): all_d += data[i] ratings = [t[1] for t in all_d] (h, _) = np.histogram(ratings, bins=np.arange(1, 7)) h = (h / h.sum()) h = h[np.nonzero(h)] erd[i] = ((- h) * np.log2(h)).sum() return erd def ETG(self, data): etg = {} edges = [0, 0.5, 1, 4, 7, 13, 33] for (i, d) in data.items(): if (len(d) <= 1): etg[i] = 0 continue posting_dates = sorted([datetime.strptime(t[3], self.date_time_format_str) for t in d]) delta_days = [(posting_dates[(i + 1)] - posting_dates[i]).days for i in range((len(posting_dates) - 1))] delta_days = [d for d in delta_days if (d < 33)] h = [] for delta in delta_days: j = 0 while ((j < len(edges)) and (delta > edges[j])): j += 1 h.append(j) (_, h) = np.unique(h, return_counts=True) if (h.sum() == 0): etg[i] = 0 continue h = (h / h.sum()) h = h[np.nonzero(h)] etg[i] = np.sum(((- h) * np.log2(h))) return etg def iETG(self, data, new_data): etg = {} edges = [0, 0.5, 1, 4, 7, 13, 33] for (i, d) in new_data.items(): all_d = deepcopy(d) if (i in data): all_d += data[i] if (len(all_d) <= 1): etg[i] = 0 continue posting_dates = sorted([datetime.strptime(t[3], self.date_time_format_str) for t in all_d]) delta_days = [(posting_dates[(i + 1)] - posting_dates[i]).days for i in range((len(posting_dates) - 1))] delta_days = [d for d in delta_days if (d < 33)] h = [] for delta in delta_days: j = 0 while ((j < len(edges)) and (delta > edges[j])): j += 1 h.append(j) (_, h) = np.unique(h, return_counts=True) if (h.sum() == 0): etg[i] = 0 continue h = (h / h.sum()) h = h[np.nonzero(h)] etg[i] = np.sum(((- h) * np.log2(h))) return etg def RD(self, product_data): rd = {} for (i, d) in product_data.items(): avg = np.mean(np.array([t[1] for t in d])) for t in d: rd[(t[0], i)] = abs((t[1] - avg)) return rd def iRD(self, product_data, new_product_data): rd = {} for (i, d) in new_product_data.items(): all_d = deepcopy(d) if (i in product_data): all_d = (d + product_data[i]) avg = np.mean(np.array([t[1] for t in all_d])) for t in all_d: rd[(t[0], i)] = abs((t[1] - avg)) return rd def EXT(self, product_data): ext = {} for (i, d) in product_data.items(): for t in d: if ((int(t[1]) == 5) or (int(t[1]) == 1)): ext[(t[0], i)] = 1 else: ext[(t[0], i)] = 0 return ext def iEXT(self, product_data, new_product_data): ext = {} for (i, d) in new_product_data.items(): all_d = deepcopy(d) if (i in product_data): all_d = (d + product_data[i]) for t in all_d: if ((int(t[1]) == 5) or (int(t[1]) == 1)): ext[(t[0], i)] = 1 else: ext[(t[0], i)] = 0 return ext def DEV(self, product_data): beta_1 = 0.63 dev = {} for (i, d) in product_data.items(): p_avg_rating = np.mean(np.array([t[1] for t in d])) for t in d: u_id = t[0] if ((abs((p_avg_rating - t[1])) / 4.0) > 0.63): dev[(u_id, i)] = 1 else: dev[(u_id, i)] = 0 return dev def iDEV(self, product_data, new_product_data): beta_1 = 0.63 dev = {} for (i, d) in new_product_data.items(): all_d = deepcopy(d) if (i in product_data): all_d = (d + product_data[i]) p_avg_rating = np.mean(np.array([t[1] for t in all_d])) for t in all_d: u_id = t[0] if ((abs((p_avg_rating - t[1])) / 4.0) > 0.63): dev[(u_id, i)] = 1 else: dev[(u_id, i)] = 0 return dev def ETF(self, product_data): beta_3 = 0.69 first_time_product = {} for (i, d) in product_data.items(): for t in d: if (i not in first_time_product): first_time_product[i] = datetime.strptime(t[3], self.date_time_format_str) elif (datetime.strptime(t[3], self.date_time_format_str) < first_time_product[i]): first_time_product[i] = datetime.strptime(t[3], self.date_time_format_str) etf = {} for (i, d) in product_data.items(): for t in d: td = (datetime.strptime(t[3], self.date_time_format_str) - first_time_product[i]) if ((t[0], i) not in etf): etf[(t[0], i)] = td elif (td > etf[(t[0], i)]): etf[(t[0], i)] = td for (k, v) in etf.items(): if (v.days > (7 * 30)): etf[k] = 0 elif ((1.0 - (v.days / (7 * 30))) > beta_3): etf[k] = 1 else: etf[k] = 0 return etf def iETF(self, product_data, new_product_data): beta_3 = 0.69 first_time_product = {} for (i, d) in new_product_data.items(): all_d = deepcopy(d) if (i in product_data): all_d = (d + product_data[i]) for t in all_d: if (i not in first_time_product): first_time_product[i] = datetime.strptime(t[3], self.date_time_format_str) elif (datetime.strptime(t[3], self.date_time_format_str) < first_time_product[i]): first_time_product[i] = datetime.strptime(t[3], self.date_time_format_str) etf = {} for (i, d) in new_product_data.items(): all_d = deepcopy(d) if (i in product_data): all_d = (d + product_data[i]) for t in all_d: td = (datetime.strptime(t[3], self.date_time_format_str) - first_time_product[i]) if ((t[0], i) not in etf): etf[(t[0], i)] = td elif (td > etf[(t[0], i)]): etf[(t[0], i)] = td for (k, v) in etf.items(): if (v.days > (7 * 30)): etf[k] = 0 elif ((1.0 - (v.days / (7 * 30))) > beta_3): etf[k] = 1 else: etf[k] = 0 return etf def ISR(self, user_data): isr = {} for (i, d) in user_data.items(): for t in d: if (len(d) == 1): isr[(i, t[0])] = 1 else: isr[(i, t[0])] = 0 return isr def iISR(self, user_data, new_user_data): isr = {} for (i, d) in new_user_data.items(): all_d = deepcopy(d) if (i in user_data): all_d = (d + user_data[i]) for t in all_d: if (len(all_d) == 1): isr[(i, t[0])] = 1 else: isr[(i, t[0])] = 0 return isr def add_feature(self, existing_features, new_features, feature_names): for (k, v) in new_features.items(): if (k not in existing_features): existing_features[k] = dict() for i in range(len(feature_names)): if (len(feature_names) > 1): existing_features[k][feature_names[i]] = v[i] else: existing_features[k][feature_names[i]] = v def construct_all_features(self, user_data, prod_data): UserFeatures = {} ProdFeatures = {} uf = self.MNR(user_data, data_type='user') self.add_feature(UserFeatures, uf, ['MNR']) pf = self.MNR(prod_data, data_type='prod') self.add_feature(ProdFeatures, pf, ['MNR']) uf = self.PR_NR(user_data) self.add_feature(UserFeatures, uf, ['PR', 'NR']) pf = self.PR_NR(prod_data) self.add_feature(ProdFeatures, pf, ['PR', 'NR']) uf = self.avgRD_user(user_data, prod_data) self.add_feature(UserFeatures, uf, ['avgRD']) pf = self.avgRD_prod(prod_data) self.add_feature(ProdFeatures, pf, ['avgRD']) uf = self.BST(user_data) self.add_feature(UserFeatures, uf, ['BST']) uf = self.ERD(user_data) self.add_feature(UserFeatures, uf, ['ERD']) pf = self.ERD(prod_data) self.add_feature(ProdFeatures, pf, ['ERD']) uf = self.ETG(user_data) self.add_feature(UserFeatures, uf, ['ETG']) pf = self.ETG(prod_data) self.add_feature(ProdFeatures, pf, ['ETG']) ReviewFeatures = {} rf = self.RD(prod_data) self.add_feature(ReviewFeatures, rf, ['RD']) rf = self.EXT(prod_data) self.add_feature(ReviewFeatures, rf, ['EXT']) rf = self.DEV(prod_data) self.add_feature(ReviewFeatures, rf, ['DEV']) rf = self.ETF(prod_data) self.add_feature(ReviewFeatures, rf, ['ETF']) rf = self.ISR(user_data) self.add_feature(ReviewFeatures, rf, ['ISR']) return (UserFeatures, ProdFeatures, ReviewFeatures) def update_all_features(self, user_data, new_user_data, prod_data, new_product_data, UserFeatures, ProdFeatures, ReviewFeatures): uf = self.iMNR(user_data, new_user_data, data_type='user') self.add_feature(UserFeatures, uf, ['MNR']) pf = self.iMNR(prod_data, new_product_data, data_type='prod') self.add_feature(ProdFeatures, pf, ['MNR']) uf = self.iPR_NR(user_data, new_user_data) self.add_feature(UserFeatures, uf, ['PR', 'NR']) pf = self.iPR_NR(prod_data, new_product_data) self.add_feature(ProdFeatures, pf, ['PR', 'NR']) uf = self.iavgRD_user(user_data, new_user_data, prod_data, new_product_data) self.add_feature(UserFeatures, uf, ['avgRD']) pf = self.iavgRD_prod(prod_data, new_product_data) self.add_feature(ProdFeatures, pf, ['avgRD']) uf = self.iBST(user_data, new_user_data) self.add_feature(UserFeatures, uf, ['BST']) uf = self.iERD(user_data, new_user_data) self.add_feature(UserFeatures, uf, ['ERD']) pf = self.iERD(prod_data, new_product_data) self.add_feature(ProdFeatures, pf, ['ERD']) uf = self.iETG(user_data, new_user_data) self.add_feature(UserFeatures, uf, ['ETG']) pf = self.iETG(prod_data, new_product_data) self.add_feature(ProdFeatures, pf, ['ETG']) rf = self.iRD(prod_data, new_product_data) self.add_feature(ReviewFeatures, rf, ['RD']) rf = self.iEXT(prod_data, new_product_data) self.add_feature(ReviewFeatures, rf, ['EXT']) rf = self.iDEV(prod_data, new_product_data) self.add_feature(ReviewFeatures, rf, ['DEV']) rf = self.iETF(prod_data, new_product_data) self.add_feature(ReviewFeatures, rf, ['ETF']) rf = self.iISR(user_data, new_user_data) self.add_feature(ReviewFeatures, rf, ['ISR']) return (UserFeatures, ProdFeatures, ReviewFeatures) def calculateNodePriors(self, feature_names, features_py, when_suspicious): priors = {} for (node_id, v) in features_py.items(): priors[node_id] = 0 for (f_idx, fn) in enumerate(feature_names): fv_py = [] for (node_id, v) in features_py.items(): if (fn not in v): fv_py.append((node_id, (- 1))) else: fv_py.append((node_id, v[fn])) fv_py = sorted(fv_py, key=(lambda x: x[1])) i = 0 while (i < len(fv_py)): start = i end = (i + 1) while ((end < len(fv_py)) and (fv_py[start][1] == fv_py[end][1])): end += 1 i = end for j in range(start, end): node_id = fv_py[j][0] if (fv_py[j][0] == (- 1)): priors[node_id] += pow(0.5, 2) continue if (when_suspicious[fn] == '+1'): priors[node_id] += pow((1.0 - (float((start + 1)) / len(fv_py))), 2) else: priors[node_id] += pow((float(end) / len(fv_py)), 2) for (node_id, v) in features_py.items(): priors[node_id] = (1.0 - math.sqrt((priors[node_id] / len(feature_names)))) if (priors[node_id] > 0.999): priors[node_id] = 0.999 elif (priors[node_id] < 0.001): priors[node_id] = 0.001 return priors def calNewNodePriors(self, feature_names, features_py, when_suspicious): priors = {} for (node_id, v) in features_py.items(): priors[node_id] = [] for (f_idx, fn) in enumerate(feature_names): fv_py = [] for (node_id, v) in features_py.items(): if (fn not in v): fv_py.append((node_id, (- 1))) else: fv_py.append((node_id, v[fn])) fv_py = sorted(fv_py, key=(lambda x: x[1])) i = 0 while (i < len(fv_py)): start = i end = (i + 1) while ((end < len(fv_py)) and (fv_py[start][1] == fv_py[end][1])): end += 1 i = end for j in range(start, end): node_id = fv_py[j][0] if (fv_py[j][0] == (- 1)): priors[node_id] += pow(0.5, 2) continue if (when_suspicious[fn] == '+1'): priors[node_id].append((1.0 - (float((start + 1)) / len(fv_py)))) if (node_id == '201'): print(priors[node_id]) else: priors[node_id].append((float(end) / len(fv_py))) if (node_id == '201'): print(priors[node_id]) for (node_id, v) in features_py.items(): priors[node_id] = min(priors[node_id]) if (priors[node_id] > 0.999): priors[node_id] = 0.999 elif (priors[node_id] < 0.001): priors[node_id] = 0.001 return priors
def bootstrap_confidence(true, pred, n=10000, confidence=0.9): Rs = [] for _ in range(n): indice = np.random.randint(0, len(pred), len(pred)) t = [true[i] for i in indice] p = [pred[i] for i in indice] (a, b, R, _, std_err) = stats.linregress(t, p) Rs.append(R) Rs = np.array(Rs) return stats.t.interval(confidence, (len(Rs) - 1), loc=np.mean(Rs), scale=np.std(Rs))
class Conv2d(_ConvBase): def __init__(self, in_size: int, out_size: int, *, kernel_size: Tuple[(int, int)]=(1, 1), stride: Tuple[(int, int)]=(1, 1), padding: Tuple[(int, int)]=(0, 0), activation=nn.ReLU(inplace=True), bn: bool=False, init=nn.init.kaiming_normal_, bias: bool=True, preact: bool=False, name: str=''): super().__init__(in_size, out_size, kernel_size, stride, padding, activation, bn, init, conv=nn.Conv2d, batch_norm=BatchNorm2d, bias=bias, preact=preact, name=name)
def main(): if (not os.path.exists(opt.outf)): os.makedirs(opt.outf) print('Loading dataset ...\n') dataset_train = Dataset(train=True) loader_train = DataLoader(dataset=dataset_train, num_workers=4, batch_size=opt.batchSize, shuffle=True, pin_memory=True) print(('# of training samples: %d\n' % int(len(dataset_train)))) net = WINNetklvl(steps=opt.num_of_steps, layers=opt.num_of_layers, channels=opt.num_of_channels, klvl=opt.lvl, mode=opt.split, dnlayers=opt.dnlayers) criterion = nn.MSELoss(size_average=False).cuda() device_ids = [0] model = nn.DataParallel(net, device_ids=device_ids).cuda() torch.backends.cudnn.benchmark = True pytorch_total_params = sum((p.numel() for p in model.parameters() if p.requires_grad)) print('Total Number of Parameters: ', pytorch_total_params) optimizer = optim.Adam(model.parameters(), lr=opt.lr) noiseL_B = [0, 55] start_epoch = opt.start_epoch if (start_epoch > 0): print('Start Epoch: ', start_epoch) model.load_state_dict(torch.load(os.path.join(opt.outf, 'net_WINNet_epoch_{}.pth'.format(start_epoch)))) for param_group in optimizer.param_groups: current_lr = param_group['lr'] current_lr = (current_lr * (opt.decay_rate ** (start_epoch // opt.milestone))) param_group['lr'] = current_lr print('Learning rate: ', current_lr) else: torch.save(model.state_dict(), os.path.join(opt.outf, 'net_WINNet_epoch_{}.pth'.format(start_epoch))) epoch = (start_epoch + 1) torch.cuda.synchronize() t1 = time.time() print('Epoch: ', epoch) while (epoch <= opt.epochs): with tqdm(loader_train, unit='batch') as tepoch: i = 0 for data in tepoch: tepoch.set_description(f'Epoch {epoch}') model.train() model.zero_grad() optimizer.zero_grad() img_train = data.cuda() if (opt.mode == 'S'): stdN = opt.noiseL if (opt.mode == 'B'): stdRD = np.random.uniform(noiseL_B[0], noiseL_B[1], size=1) stdN = stdRD[0] stdNv = Variable((stdN * torch.ones(1, device=torch.device('cuda:0')))) noise = torch.cuda.FloatTensor(img_train.size()).normal_(mean=0, std=(stdN / 255.0)) imgn_train = (img_train + noise) (img_train, imgn_train) = (Variable(img_train), Variable(imgn_train)) (outn_train, oloss) = model(imgn_train, stdNv) loss = (criterion(outn_train, img_train) / (imgn_train.size()[0] * 2)) loss = (loss + oloss) Loss = loss.detach() if ((i % 10) == 0): norm_PU = net.linear() loss = (loss + ((1 * 0.1) * norm_PU)) loss.backward() nn.utils.clip_grad_norm_(model.parameters(), max_norm=1, norm_type=2) optimizer.step() tepoch.set_postfix(loss=Loss) time.sleep(0.0001) i += 1 torch.cuda.synchronize() t2 = time.time() print(f'Time:{((t2 - t1) / 60): .3e} mins') print('Orthogonal Loss: ', oloss) print('Norm PU', norm_PU) for param_group in optimizer.param_groups: current_lr = param_group['lr'] print('Learning rate: ', current_lr) if (torch.isnan(Loss) or torch.isinf(Loss)): if (epoch > 0): print('Load Model Epoch: ', (epoch - 1)) model.load_state_dict(torch.load(os.path.join(opt.outf, 'net_WINNet_epoch_{}.pth'.format((epoch - 1))))) else: net = WINNetklvl(steps=opt.num_of_steps, layers=opt.num_of_layers, channels=opt.num_of_channels, klvl=opt.lvl, mode=opt.modeSM, dnlayers=opt.dnlayers) model = nn.DataParallel(net, device_ids=device_ids).cuda() for param_group in optimizer.param_groups: current_lr = param_group['lr'] current_lr = (current_lr * 0.8) param_group['lr'] = current_lr print('Learning rate: ', current_lr) continue torch.cuda.synchronize() t1 = time.time() print(('Save Model Epoch: %d' % epoch)) print('\n') torch.save(model.state_dict(), os.path.join(opt.outf, 'net_WINNet_epoch_{}.pth'.format(epoch))) epoch += 1 print('Epoch: ', epoch) if ((epoch > 0) and ((epoch % opt.milestone) == 0)): for param_group in optimizer.param_groups: current_lr = param_group['lr'] current_lr = (current_lr * opt.decay_rate) param_group['lr'] = current_lr print('Learning rate: ', current_lr) if (opt.mode == 'S'): prepare_data(data_path='data', patch_size=40, stride=10, aug_times=1) if (opt.mode == 'B'): prepare_data(data_path='data', patch_size=50, stride=10, aug_times=2)
class PaddedAdditiveSelfAttention(snt.AbstractModule): def __init__(self, v_dim, attn_mlp_fn, attn_output_dim, gnn_mlp_fn, max_n_node, train_batch_size, node_embedding_dim, scaling=True, name='padded_additive_self_attention'): super(PaddedAdditiveSelfAttention, self).__init__(name=name) self.v_dim = v_dim self.attn_mlp = attn_mlp_fn() self.gnn_mlp = gnn_mlp_fn() self.attn_output_dim = attn_output_dim self.max_n_node = max_n_node self.scaling = scaling self.train_batch_size = train_batch_size self.node_embedding_dim = node_embedding_dim self.wv = tf.keras.layers.Dense(v_dim, use_bias=False) def pad(self, node_embeddings, n_node): n_node_diffs = (self.max_n_node - n_node) graph_list = tf.split(node_embeddings, n_node, 0) padded_graphs = [] padding = tf.constant([[0, 1], [0, 0]]) for i in range(self.train_batch_size): padded_graphs.append(tf.pad(graph_list[i], (padding * n_node_diffs[i]))) final_output = tf.stack(padded_graphs) return final_output def unpad(self, node_embeddings, n_node): graph_list = tf.unstack(node_embeddings, axis=0) to_concat = [] for i in range(self.train_batch_size): to_concat.append(graph_list[i][0:n_node[i]]) final_output = tf.concat(to_concat, axis=0) return final_output def _build(self, graph): node_embeddings = graph.nodes node_embeddings = tf.Print(node_embeddings, [node_embeddings], 'node_embeddings is: ', summarize=100, first_n=1) project_v = self.wv(node_embeddings) project_v = tf.Print(project_v, [project_v], 'project v is: ', summarize=100, first_n=1) padded_project_v = self.pad(project_v, graph.n_node) padded_project_v.set_shape([self.train_batch_size, self.max_n_node, self.v_dim]) padded_project_v = tf.Print(padded_project_v, [padded_project_v], 'padded project v is ', summarize=100, first_n=1) padded_node_embeddings = self.pad(node_embeddings, graph.n_node) padded_node_embeddings = tf.Print(padded_node_embeddings, [padded_node_embeddings], 'padded_node_embeddings is: ', summarize=100, first_n=1) pairwise_concat = pairwise_concat_padded(padded_node_embeddings) pairwise_concat = tf.reshape(pairwise_concat, [(self.train_batch_size * (self.max_n_node ** 2)), self.node_embedding_dim]) pairwise_concat.set_shape([None, self.node_embedding_dim]) additive_attn_output = self.attn_mlp(pairwise_concat) dot_var = tf.get_variable('dot_var', shape=[self.attn_output_dim], dtype=tf.float32, initializer=tf.glorot_normal_initializer, trainable=True) d = tf.tensordot(dot_var, additive_attn_output, axes=[0, 1]) logits = tf.reshape(d, [self.train_batch_size, self.max_n_node, self.max_n_node]) logits.set_shape([self.train_batch_size, self.max_n_node, self.max_n_node]) lm = loss_mask_padded(graph, self.max_n_node) logits = (logits - (100000 * (1 - lm))) attn_weights = tf.nn.softmax(logits, axis=(- 1)) attended_nodes = tf.matmul(attn_weights, padded_project_v) attended_nodes = self.unpad(attended_nodes, graph.n_node) concat_nodes = tf.concat([graph.nodes, attended_nodes], axis=(- 1)) new_nodes = self.gnn_mlp(concat_nodes) return graph.replace(nodes=new_nodes)
def gen_dis_sample(train_pos_path, train_neg_path, vocab_path, n_dim=100, res_file='discriminator_train_data.npz'): vocab_map = load_vocab(vocab_path) print('vocab length:', len(vocab_map)) train_pos = [] for file in os.listdir(train_pos_path): with open(os.path.join(train_pos_path, file), 'r') as rf: content = rf.read().strip('.').split() text = list(map((lambda x: (vocab_map[x] if (x in vocab_map) else 0)), content)) if (len(text) <= n_dim): sample = np.pad(text, (0, (n_dim - len(text))), mode='constant', constant_values=1).astype(np.int) else: sample = text[:n_dim] train_pos.append(sample) print('positive samples:', len(train_pos)) train_neg = [] for file in os.listdir(train_neg_path): with open(os.path.join(train_neg_path, file), 'r') as rf: content = rf.read().strip('.').split() text = list(map((lambda x: (vocab_map[x] if (x in vocab_map) else 0)), content)) if (len(text) <= n_dim): sample = np.pad(text, (0, (n_dim - len(text))), mode='constant', constant_values=1).astype(np.int) else: sample = text[:n_dim] train_neg.append(sample) print('negative samples:', len(train_neg)) np.savez(res_file, pos_summary_idx=np.array(train_pos), neg_summary_idx=np.array(train_neg)) print('file saved: ', res_file)
_torch _vision class MgpstrProcessorTest(unittest.TestCase): image_processing_class = (ViTImageProcessor if is_vision_available() else None) def image_processor_dict(self): return self.image_processor_tester.prepare_image_processor_dict() def setUp(self): self.image_size = (3, 32, 128) self.tmpdirname = tempfile.mkdtemp() vocab = ['[GO]', '[s]', '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', 'x', 'y', 'z'] vocab_tokens = dict(zip(vocab, range(len(vocab)))) self.vocab_file = os.path.join(self.tmpdirname, VOCAB_FILES_NAMES['vocab_file']) with open(self.vocab_file, 'w', encoding='utf-8') as fp: fp.write((json.dumps(vocab_tokens) + '\n')) image_processor_map = {'do_normalize': False, 'do_resize': True, 'feature_extractor_type': 'ViTFeatureExtractor', 'resample': 3, 'size': {'height': 32, 'width': 128}} self.image_processor_file = os.path.join(self.tmpdirname, IMAGE_PROCESSOR_NAME) with open(self.image_processor_file, 'w', encoding='utf-8') as fp: json.dump(image_processor_map, fp) def get_tokenizer(self, **kwargs): return MgpstrTokenizer.from_pretrained(self.tmpdirname, **kwargs) def get_image_processor(self, **kwargs): return ViTImageProcessor.from_pretrained(self.tmpdirname, **kwargs) def tearDown(self): shutil.rmtree(self.tmpdirname) def prepare_image_inputs(self): image_input = np.random.randint(255, size=(3, 30, 400), dtype=np.uint8) image_input = Image.fromarray(np.moveaxis(image_input, 0, (- 1))) return image_input def test_save_load_pretrained_default(self): tokenizer = self.get_tokenizer() image_processor = self.get_image_processor() processor = MgpstrProcessor(tokenizer=tokenizer, image_processor=image_processor) processor.save_pretrained(self.tmpdirname) processor = MgpstrProcessor.from_pretrained(self.tmpdirname, use_fast=False) self.assertEqual(processor.char_tokenizer.get_vocab(), tokenizer.get_vocab()) self.assertIsInstance(processor.char_tokenizer, MgpstrTokenizer) self.assertEqual(processor.image_processor.to_json_string(), image_processor.to_json_string()) self.assertIsInstance(processor.image_processor, ViTImageProcessor) def test_save_load_pretrained_additional_features(self): tokenizer = self.get_tokenizer() image_processor = self.get_image_processor() processor = MgpstrProcessor(tokenizer=tokenizer, image_processor=image_processor) processor.save_pretrained(self.tmpdirname) tokenizer_add_kwargs = self.get_tokenizer(bos_token='(BOS)', eos_token='(EOS)') image_processor_add_kwargs = self.get_image_processor(do_normalize=False, padding_value=1.0) processor = MgpstrProcessor.from_pretrained(self.tmpdirname, bos_token='(BOS)', eos_token='(EOS)', do_normalize=False, padding_value=1.0) self.assertEqual(processor.char_tokenizer.get_vocab(), tokenizer_add_kwargs.get_vocab()) self.assertIsInstance(processor.char_tokenizer, MgpstrTokenizer) self.assertEqual(processor.image_processor.to_json_string(), image_processor_add_kwargs.to_json_string()) self.assertIsInstance(processor.image_processor, ViTImageProcessor) def test_image_processor(self): image_processor = self.get_image_processor() tokenizer = self.get_tokenizer() processor = MgpstrProcessor(tokenizer=tokenizer, image_processor=image_processor) image_input = self.prepare_image_inputs() input_image_proc = image_processor(image_input, return_tensors='np') input_processor = processor(images=image_input, return_tensors='np') for key in input_image_proc.keys(): self.assertAlmostEqual(input_image_proc[key].sum(), input_processor[key].sum(), delta=0.01) def test_tokenizer(self): image_processor = self.get_image_processor() tokenizer = self.get_tokenizer() processor = MgpstrProcessor(tokenizer=tokenizer, image_processor=image_processor) input_str = 'test' encoded_processor = processor(text=input_str) encoded_tok = tokenizer(input_str) for key in encoded_tok.keys(): self.assertListEqual(encoded_tok[key], encoded_processor[key]) def test_processor(self): image_processor = self.get_image_processor() tokenizer = self.get_tokenizer() processor = MgpstrProcessor(tokenizer=tokenizer, image_processor=image_processor) input_str = 'test' image_input = self.prepare_image_inputs() inputs = processor(text=input_str, images=image_input) self.assertListEqual(list(inputs.keys()), ['pixel_values', 'labels']) with pytest.raises(ValueError): processor() def test_tokenizer_decode(self): image_processor = self.get_image_processor() tokenizer = self.get_tokenizer() processor = MgpstrProcessor(tokenizer=tokenizer, image_processor=image_processor) predicted_ids = [[1, 4, 5, 8, 1, 0, 8], [3, 4, 3, 1, 1, 8, 9], [3, 4, 3, 1, 1, 8, 9]] decoded_processor = processor.char_decode(predicted_ids) decoded_tok = tokenizer.batch_decode(predicted_ids) decode_strs = [seq.replace(' ', '') for seq in decoded_tok] self.assertListEqual(decode_strs, decoded_processor) def test_model_input_names(self): image_processor = self.get_image_processor() tokenizer = self.get_tokenizer() processor = MgpstrProcessor(tokenizer=tokenizer, image_processor=image_processor) input_str = None image_input = self.prepare_image_inputs() inputs = processor(text=input_str, images=image_input) self.assertListEqual(list(inputs.keys()), processor.model_input_names) def test_processor_batch_decode(self): image_processor = self.get_image_processor() tokenizer = self.get_tokenizer() processor = MgpstrProcessor(tokenizer=tokenizer, image_processor=image_processor) char_input = torch.randn(1, 27, 38) bpe_input = torch.randn(1, 27, 50257) wp_input = torch.randn(1, 27, 30522) results = processor.batch_decode([char_input, bpe_input, wp_input]) self.assertListEqual(list(results.keys()), ['generated_text', 'scores', 'char_preds', 'bpe_preds', 'wp_preds'])
class ConvLSTM(nn.Module): def __init__(self, input_channels, hidden_channels, kernel_size, step=1, effective_step=[1]): super(ConvLSTM, self).__init__() self.input_channels = ([input_channels] + hidden_channels) self.hidden_channels = hidden_channels self.kernel_size = kernel_size self.num_layers = len(hidden_channels) self.step = step self.effective_step = effective_step self._all_layers = [] for i in range(self.num_layers): name = 'cell{}'.format(i) cell = ConvLSTMCell(self.input_channels[i], self.hidden_channels[i], self.kernel_size) setattr(self, name, cell) self._all_layers.append(cell) def forward(self, input): internal_state = [] outputs = [] for step in range(self.step): x = input for i in range(self.num_layers): name = 'cell{}'.format(i) if (step == 0): (bsize, _, height, width) = x.size() (h, c) = getattr(self, name).init_hidden(batch_size=bsize, hidden=self.hidden_channels[i], shape=(height, width)) internal_state.append((h, c)) (h, c) = internal_state[i] (x, new_c) = getattr(self, name)(x, h, c) internal_state[i] = (x, new_c) if (step in self.effective_step): outputs.append(x) return (outputs, (x, new_c))
def initialize_dataset_loader(data, stage, params, loader_default_params=None): transform = initialize_transforms(params.pop('transforms'), mean_std=params.pop('mean_std')) dataset = initialize_dataset(data, stage, transform, params.pop('dataset')) loader_params = {**LOADER_DEFAULT_PARAMS, **(loader_default_params or {}), **dataset.loader_params, **params.pop('loader', {})} assert ('batch_size' in loader_params) assert (not params), params.keys() return DataLoader(dataset, **loader_params)
def make_fast_softmax_attention(qkv_dim, renormalize_attention=True, numerical_stabilizer=1e-06, nb_features=256, ortho_features=True, ortho_scaling=0.0, redraw_features=True, unidirectional=False, nonnegative_features=True, lax_scan_unroll=1): logging.info('Fast softmax attention: %s features and orthogonal=%s, renormalize=%s', nb_features, ortho_features, renormalize_attention) if ortho_features: matrix_creator = functools.partial(GaussianOrthogonalRandomMatrix, nb_features, qkv_dim, scaling=ortho_scaling) else: matrix_creator = functools.partial(GaussianUnstructuredRandomMatrix, nb_features, qkv_dim) if nonnegative_features: def kernel_feature_creator(data, projection_matrix, attention_dims_t, batch_dims_t, precision, is_query, normalize_data=True): return nonnegative_softmax_kernel_feature_creator(data, projection_matrix, attention_dims_t, batch_dims_t, precision, is_query, normalize_data, numerical_stabilizer) else: def kernel_feature_creator(data, projection_matrix, attention_dims_t, batch_dims_t, precision, is_query, normalize_data=True): del is_query return sincos_softmax_kernel_feature_creator(data, projection_matrix, attention_dims_t, batch_dims_t, precision, normalize_data) attention_fn = FastAttentionviaLowRankDecomposition(matrix_creator, kernel_feature_creator, renormalize_attention=renormalize_attention, numerical_stabilizer=numerical_stabilizer, redraw_features=redraw_features, unidirectional=unidirectional, lax_scan_unroll=lax_scan_unroll).dot_product_attention return attention_fn
class TrainingModule(LightningModule): def __init__(self, cfg): super().__init__() if (not logger.isEnabledFor(logging.INFO)): setup_logger() self.cfg = DefaultTrainer.auto_scale_workers(cfg, comm.get_world_size()) self.storage: EventStorage = None self.model = build_model(self.cfg) self.start_iter = 0 self.max_iter = cfg.SOLVER.MAX_ITER def on_save_checkpoint(self, checkpoint: Dict[(str, Any)]) -> None: checkpoint['iteration'] = self.storage.iter def on_load_checkpoint(self, checkpointed_state: Dict[(str, Any)]) -> None: self.start_iter = checkpointed_state['iteration'] self.storage.iter = self.start_iter def setup(self, stage: str): if self.cfg.MODEL.WEIGHTS: self.checkpointer = DetectionCheckpointer(self.model, self.cfg.OUTPUT_DIR) logger.info(f'Load model weights from checkpoint: {self.cfg.MODEL.WEIGHTS}.') self.checkpointer.load(self.cfg.MODEL.WEIGHTS) self.iteration_timer = hooks.IterationTimer() self.iteration_timer.before_train() self.data_start = time.perf_counter() self.writers = None def training_step(self, batch, batch_idx): data_time = (time.perf_counter() - self.data_start) if (self.storage is None): self.storage = EventStorage(0) self.storage.__enter__() self.iteration_timer.trainer = weakref.proxy(self) self.iteration_timer.before_step() self.writers = (default_writers(self.cfg.OUTPUT_DIR, self.max_iter) if comm.is_main_process() else {}) loss_dict = self.model(batch) SimpleTrainer.write_metrics(loss_dict, data_time) opt = self.optimizers() self.storage.put_scalar('lr', opt.param_groups[self._best_param_group_id]['lr'], smoothing_hint=False) self.iteration_timer.after_step() self.storage.step() self.iteration_timer.before_step() if ((self.storage.iter % 20) == 0): for writer in self.writers: writer.write() return sum(loss_dict.values()) def training_step_end(self, training_step_outpus): self.data_start = time.perf_counter() return training_step_outpus def training_epoch_end(self, training_step_outputs): self.iteration_timer.after_train() if comm.is_main_process(): self.checkpointer.save('model_final') for writer in self.writers: writer.write() writer.close() self.storage.__exit__(None, None, None) def _process_dataset_evaluation_results(self) -> OrderedDict: results = OrderedDict() for (idx, dataset_name) in enumerate(self.cfg.DATASETS.TEST): results[dataset_name] = self._evaluators[idx].evaluate() if comm.is_main_process(): print_csv_format(results[dataset_name]) if (len(results) == 1): results = list(results.values())[0] return results def _reset_dataset_evaluators(self): self._evaluators = [] for dataset_name in self.cfg.DATASETS.TEST: evaluator = build_evaluator(self.cfg, dataset_name) evaluator.reset() self._evaluators.append(evaluator) def on_validation_epoch_start(self, _outputs): self._reset_dataset_evaluators() def validation_epoch_end(self, _outputs): results = self._process_dataset_evaluation_results(_outputs) flattened_results = flatten_results_dict(results) for (k, v) in flattened_results.items(): try: v = float(v) except Exception as e: raise ValueError("[EvalHook] eval_function should return a nested dict of float. Got '{}: {}' instead.".format(k, v)) from e self.storage.put_scalars(**flattened_results, smoothing_hint=False) def validation_step(self, batch, batch_idx: int, dataloader_idx: int=0) -> None: if (not isinstance(batch, List)): batch = [batch] outputs = self.model(batch) self._evaluators[dataloader_idx].process(batch, outputs) def configure_optimizers(self): optimizer = build_optimizer(self.cfg, self.model) self._best_param_group_id = hooks.LRScheduler.get_best_param_group_id(optimizer) scheduler = build_lr_scheduler(self.cfg, optimizer) return ([optimizer], [{'scheduler': scheduler, 'interval': 'step'}])
def plot_numerical_error(): print(f'generating numerical benchmark plots. reading data from {TEST_RESULTS_PICKLE}...') try: with open(TEST_RESULTS_PICKLE, 'rb') as f: results = pickle.load(f) except FileNotFoundError: print('no data of numerical errors found. run the numerical tests first') return plot_num_err_heatmap(results) plot_num_error_growth_comparison(results) plot_num_coeffs2num_ops(results) print('plotting numerical benchmarks done.')
def sum_concat_then_mlp_gnn(make_mlp_fn): node_block = ConcatThenMLPBlock(tf.unsorted_segment_sum, make_mlp_fn) return NodeBlockGNN(node_block)
_loss('bce_kl_combined') class CombinedLoss(nn.Module): def __init__(self, weight_softmax): super().__init__() self.weight_softmax = weight_softmax def forward(self, sample_list, model_output): pred_score = model_output['scores'] target_score = sample_list['targets'] tar_sum = torch.sum(target_score, dim=1, keepdim=True) tar_sum_is_0 = torch.eq(tar_sum, 0) tar_sum.masked_fill_(tar_sum_is_0, 1e-06) tar = (target_score / tar_sum) res = F.log_softmax(pred_score, dim=1) loss1 = kl_div(res, tar) loss1 = (torch.sum(loss1) / loss1.size(0)) loss2 = F.binary_cross_entropy_with_logits(pred_score, target_score, reduction='mean') loss2 *= target_score.size(1) loss = ((self.weight_softmax * loss1) + loss2) return loss
class DeadReckoningNodeMultiRobot(DeadReckoningNode): def __init__(self) -> None: super().__init__() def init_node(self, ns='~') -> None: self.imu_pose = rospy.get_param((ns + 'imu_pose')) self.imu_pose = n2g(self.imu_pose, 'Pose3') self.imu_rot = self.imu_pose.rotation() self.dvl_max_velocity = rospy.get_param((ns + 'dvl_max_velocity')) self.keyframe_duration = rospy.get_param((ns + 'keyframe_duration')) self.keyframe_translation = rospy.get_param((ns + 'keyframe_translation')) self.keyframe_rotation = rospy.get_param((ns + 'keyframe_rotation')) self.dvl_sub = Subscriber(DVL_TOPIC, DVL) self.gyro_sub = Subscriber(GYRO_INTEGRATION_TOPIC, Odometry) self.depth_sub = Subscriber(DEPTH_TOPIC, Depth) self.depth_cache = Cache(self.depth_sub, 1) if (rospy.get_param((ns + 'imu_version')) == 1): self.imu_sub = Subscriber(IMU_TOPIC, Imu) elif (rospy.get_param((ns + 'imu_version')) == 2): self.imu_sub = Subscriber(IMU_TOPIC_MK_II, Imu) self.traj_pub = rospy.Publisher('traj_dead_reck', PointCloud2, queue_size=10) self.odom_pub = rospy.Publisher(LOCALIZATION_ODOM_TOPIC, Odometry, queue_size=10) self.use_gyro = rospy.get_param((ns + 'use_gyro')) if self.use_gyro: self.ts = ApproximateTimeSynchronizer([self.imu_sub, self.dvl_sub, self.gyro_sub], 300, 0.1) self.ts.registerCallback(self.callback_with_gyro) else: self.ts = ApproximateTimeSynchronizer([self.imu_sub, self.dvl_sub], 200, 0.1) self.ts.registerCallback(self.callback) self.tf = tf.TransformBroadcaster() print(self.pose) rospy.loginfo('Localization node is initialized') def publish_pose(self, flag) -> None: if (self.pose is None): return header = rospy.Header() header.stamp = self.prev_time header.frame_id = (self.rov_id + '_odom') odom_msg = Odometry() odom_msg.header = header odom_msg.pose.pose = g2r(self.pose) odom_msg.child_frame_id = (self.rov_id + '_base_link') odom_msg.twist.twist.linear.x = 0 odom_msg.twist.twist.linear.y = 0 odom_msg.twist.twist.linear.z = 0 odom_msg.twist.twist.angular.x = 0 odom_msg.twist.twist.angular.y = 0 odom_msg.twist.twist.angular.z = 0 self.odom_pub.publish(odom_msg) p = odom_msg.pose.pose.position q = odom_msg.pose.pose.orientation self.tf.sendTransform((p.x, p.y, p.z), (q.x, q.y, q.z, q.w), header.stamp, (self.rov_id + '_base_link'), (self.rov_id + '_odom'))
class PerceputalLoss(nn.modules.loss._Loss): def __init__(self, input_range='sigmoid', net_type='vgg_torch', input_preprocessing='corresponding', match=[{'layers': [11, 20, 29], 'what': 'features'}]): if (input_range not in ['sigmoid', 'tanh']): assert False self.net = get_pretrained_net(net_type).cuda() self.matchers = [get_matcher(self.net, match_opts) for match_opts in match] preprocessing_correspondence = {'vgg19_torch': vgg_preprocess_caffe, 'vgg16_torch': vgg_preprocess_caffe, 'vgg19_pytorch': vgg_preprocess_pytorch, 'vgg19_pytorch_modified': vgg_preprocess_pytorch} if (input_preprocessing == 'corresponding'): self.preprocess_input = preprocessing_correspondence[net_type] else: self.preprocessing = preprocessing_correspondence[input_preprocessing] def preprocess_input(self, x): if (self.input_range == 'tanh'): x = ((x + 1.0) / 2.0) return self.preprocess(x) def __call__(self, x, y): self.matcher_content.mode = 'store' self.net(self.preprocess_input(y)) self.matcher_content.mode = 'match' self.net(self.preprocess_input(x)) return sum([sum(matcher.losses.values()) for matcher in self.matchers])
def isect_segments__naive(segments): isect = [] if (Real is float): segments = [((s[0], s[1]) if (s[0][X] <= s[1][X]) else (s[1], s[0])) for s in segments] else: segments = [(((Real(s[0][0]), Real(s[0][1])), (Real(s[1][0]), Real(s[1][1]))) if (s[0] <= s[1]) else ((Real(s[1][0]), Real(s[1][1])), (Real(s[0][0]), Real(s[0][1])))) for s in segments] n = len(segments) for i in range(n): (a0, a1) = segments[i] for j in range((i + 1), n): (b0, b1) = segments[j] if ((a0 not in (b0, b1)) and (a1 not in (b0, b1))): ix = isect_seg_seg_v2_point(a0, a1, b0, b1) if (ix is not None): isect.append(ix) return isect
def ts2xy(ts, xaxis, yaxis): if (xaxis == X_TIMESTEPS): x = np.cumsum(ts.l.values) elif (xaxis == X_EPISODES): x = np.arange(len(ts)) elif (xaxis == X_WALLTIME): x = (ts.t.values / 3600.0) else: raise NotImplementedError if (yaxis == Y_REWARD): y = ts.r.values elif (yaxis == Y_TIMESTEPS): y = ts.l.values else: raise NotImplementedError return (x, y)
def test_statcast_pitchers_expected_stats() -> None: min_pa = 100 result: pd.DataFrame = statcast_pitcher_expected_stats(2019, min_pa) assert (result is not None) assert (not result.empty) assert (len(result.columns) == 18) assert (len(result) > 0) assert (len(result[(result['pa'] < min_pa)]) == 0)
def setup_orderdict(): from collections import OrderedDict yaml.add_representer(OrderedDict, represent_dictionary_order)
_model_architecture('dual_input_wav_transformer', 'dualinputs2twavtransformer_base') def dualinputs2twavtransformer_base(args): args.dropout_input = getattr(args, 'dropout_input', 0) args.dropout_features = getattr(args, 'dropout_features', 0) args.speech_mask_length = getattr(args, 'speech_mask_length', 10) args.speech_mask_prob = getattr(args, 'speech_mask_prob', 0.65) args.speech_mask_selection = getattr(args, 'speech_mask_selection', 'static') args.speech_mask_other = getattr(args, 'speech_mask_other', 0) args.speech_mask_min_space = getattr(args, 'speech_mask_min_space', 1) args.speech_no_mask_overlap = getattr(args, 'speech_no_mask_overlap', False) args.speech_conv_bias = getattr(args, 'speech_conv_bias', False) args.speech_extractor_mode = getattr(args, 'speech_extractor_mode', 'default') args.no_strict_check_pretrain_model = getattr(args, 'no_strict_check_pretrain_model', False) args.speech_mask_channel_length = getattr(args, 'speech_mask_channel_length', 10) args.speech_mask_channel_prob = getattr(args, 'speech_mask_channel_prob', 0.0) args.speech_mask_channel_selection = getattr(args, 'speech_mask_channel_selection', 'static') args.speech_mask_channel_other = getattr(args, 'speech_mask_channel_other', 0) args.speech_mask_channel_min_space = getattr(args, 'speech_mask_channel_min_space', 1) args.speech_no_mask_channel_overlap = getattr(args, 'speech_no_mask_channel_overlap', False) args.no_scale_feature = getattr(args, '', False) args.feature_grad_mult = getattr(args, 'feature_grad_mult', 0.0) args.encoder_embed_dim = getattr(args, 'encoder_embed_dim', 768) args.encoder_ffn_embed_dim = getattr(args, 'encoder_ffn_embed_dim', (args.encoder_embed_dim * 4)) args.encoder_attention_heads = getattr(args, 'encoder_attention_heads', 12) args.encoder_normalize_before = getattr(args, 'encoder_normalize_before', False) args.encoder_layerdrop = getattr(args, 'encoder_layerdrop', 0.1) args.encoder_learned_pos = getattr(args, 'encoder_learned_pos', False) args.decoder_embed_dim = getattr(args, 'decoder_embed_dim', args.encoder_embed_dim) args.decoder_ffn_embed_dim = getattr(args, 'decoder_ffn_embed_dim', args.encoder_ffn_embed_dim) args.decoder_attention_heads = getattr(args, 'decoder_attention_heads', args.encoder_attention_heads) args.decoder_normalize_before = getattr(args, 'decoder_normalize_before', False) args.decoder_learned_pos = getattr(args, 'decoder_learned_pos', False) args.dropout = getattr(args, 'dropout', 0.1) args.attention_dropout = getattr(args, 'attention_dropout', 0) args.activation_dropout = getattr(args, 'activation_dropout', args.dropout) args.activation_fn = getattr(args, 'activation_fn', 'relu') args.adaptive_softmax_cutoff = getattr(args, 'adaptive_softmax_cutoff', None) args.adaptive_softmax_dropout = getattr(args, 'adaptive_softmax_dropout', 0) args.tie_adaptive_weights = getattr(args, 'tie_adaptive_weights', False) args.share_decoder_input_output_embed = getattr(args, 'share_decoder_input_output_embed', False) args.no_token_positional_embeddings = getattr(args, 'no_token_positional_embeddings', False) args.adaptive_input = getattr(args, 'adaptive_input', False) args.decoder_layerdrop = getattr(args, 'decoder_layerdrop', 0.0) args.decoder_output_dim = getattr(args, 'decoder_output_dim', args.decoder_embed_dim) args.layernorm_embedding = getattr(args, 'layernorm_embedding', False) args.no_scale_embedding = getattr(args, 'no_scale_embedding', False) args.quant_noise_pq = getattr(args, 'quant_noise_pq', 0) args.speech_encoder_layers = getattr(args, 'speech_encoder_layers', 12) args.text_encoder_layers = getattr(args, 'text_encoder_layers', 6) args.encoder_shared_text_layers_from_begin = getattr(args, 'encoder_shared_text_layers_from_begin', 6) args.decoder_layers = getattr(args, 'decoder_layers', 6)
def plot_roc_curve(human_scores, gpt_scores): A = human_scores B = gpt_scores scores = (A + B) labels = (([0] * len(A)) + ([1] * len(B))) (fpr, tpr, thresholds) = roc_curve(labels, scores) roc_auc = auc(fpr, tpr) plt.figure() plt.plot(fpr, tpr, color='darkorange', lw=2, label=('ROC curve (area = %0.4f)' % roc_auc)) plt.plot([0, 1], [0, 1], color='navy', lw=2, linestyle='--') plt.xlim([0.0, 1.0]) plt.ylim([0.0, 1.05]) plt.xlabel('False Positive Rate') plt.ylabel('True Positive Rate') plt.title('ROC curve: Open-gen w/ GPT3.5-Reddit w prompts') plt.legend(loc='lower right') plt.show() for (idx, fpr_) in enumerate(fpr): if (fpr_ > 0.01): print(f'TPR at 1% FPR: {tpr[idx]:.4f}') break
def make_mmcif_features(mmcif_object: mmcif_parsing.MmcifObject, chain_id: str) -> FeatureDict: input_sequence = mmcif_object.chain_to_seqres[chain_id] description = '_'.join([mmcif_object.file_id, chain_id]) num_res = len(input_sequence) mmcif_feats = {} mmcif_feats.update(make_sequence_features(sequence=input_sequence, description=description, num_res=num_res)) (all_atom_positions, all_atom_mask) = mmcif_parsing.get_atom_coords(mmcif_object=mmcif_object, chain_id=chain_id) mmcif_feats['all_atom_positions'] = all_atom_positions mmcif_feats['all_atom_mask'] = all_atom_mask mmcif_feats['resolution'] = np.array([mmcif_object.header['resolution']], dtype=np.float32) mmcif_feats['release_date'] = np.array([mmcif_object.header['release_date'].encode('utf-8')], dtype=np.object_) mmcif_feats['is_distillation'] = np.array(0.0, dtype=np.float32) return mmcif_feats
def incorrect_edges_per_graph(true_adj, pred_adj, n_node, abs_tol=0.5): diff = remove_diag(tf.math.abs((true_adj - pred_adj))) num_incorrect = tf.where(tf.greater(diff, abs_tol), tf.ones_like(diff), tf.zeros_like(diff)) num_incorrect = tf.reduce_sum(num_incorrect, axis=0) indices = repeat_1d(tf.range(tf.shape(n_node)[0]), n_node) num_incorrect = tf.segment_sum(num_incorrect, indices) return tf.cast((num_incorrect / 2), dtype=tf.int32)
def get_caseIDs_from_splitted_dataset_folder(folder): files = subfiles(folder, suffix='.nii.gz', join=False) files = [i[:(- 12)] for i in files] files = np.unique(files) return files
def velocity_of_P_given_A(vel: T2value, omega: float, vec_ap: T2value) -> T2value: return (vel + (omega * (_rot90 vec_ap)))
class Wide_ResNet(nn.Module): def __init__(self, depth, widen_factor, dropout_rate, num_classes): super(Wide_ResNet, self).__init__() self.in_planes = 16 assert (((depth - 4) % 6) == 0), 'Wide-resnet depth should be 6n+4' n = ((depth - 4) / 6) k = widen_factor nStages = [16, (16 * k), (32 * k), (64 * k)] self.conv1 = conv3x3(3, nStages[0]) self.layer1 = self._wide_layer(wide_basic, nStages[1], n, dropout_rate, stride=1) self.layer2 = self._wide_layer(wide_basic, nStages[2], n, dropout_rate, stride=2) self.layer3 = self._wide_layer(wide_basic, nStages[3], n, dropout_rate, stride=2) self.bn1 = nn.BatchNorm2d(nStages[3], momentum=0.9) self.linear = nn.Linear(nStages[3], num_classes) def _wide_layer(self, block, planes, num_blocks, dropout_rate, stride): strides = ([stride] + ([1] * (int(num_blocks) - 1))) names = (['ds_block'] + (['n_block'] * (int(num_blocks) - 1))) layers = OrderedDict() for (i, stride) in enumerate(strides): layers[names[i]] = block(self.in_planes, planes, dropout_rate, stride) self.in_planes = planes return nn.Sequential(layers) def forward(self, x): out = self.conv1(x) out = self.layer1(out) out = self.layer2(out) out = self.layer3(out) out = F.relu(self.bn1(out)) out = F.avg_pool2d(out, 8) out = out.view(out.size(0), (- 1)) out = self.linear(out) return out
def _make_scratch_ccm(scratch, in_channels, cout, expand=False): out_channels = ([cout, (cout * 2), (cout * 4), (cout * 8)] if expand else ([cout] * 4)) scratch.layer0_ccm = nn.Conv2d(in_channels[0], out_channels[0], kernel_size=1, stride=1, padding=0, bias=True) scratch.layer1_ccm = nn.Conv2d(in_channels[1], out_channels[1], kernel_size=1, stride=1, padding=0, bias=True) scratch.layer2_ccm = nn.Conv2d(in_channels[2], out_channels[2], kernel_size=1, stride=1, padding=0, bias=True) scratch.layer3_ccm = nn.Conv2d(in_channels[3], out_channels[3], kernel_size=1, stride=1, padding=0, bias=True) scratch.CHANNELS = out_channels return scratch
def test(testloader, model, criterion, epoch, use_cuda, optimizer=None): global best_acc batch_time = AverageMeter() data_time = AverageMeter() losses = AverageMeter() top1 = AverageMeter() top5 = AverageMeter() live_top1 = AverageMeter() live_top5 = AverageMeter() dead_top1 = AverageMeter() dead_top5 = AverageMeter() if (optimizer is not None): optimizer.eval() model.eval() global_avg_pool = nr.GlobalAvgPool2d(1) torch.set_grad_enabled(False) end = time.time() bar = Bar('Processing', max=len(testloader)) for (batch_idx, (inputs, targets)) in enumerate(testloader): data_time.update((time.time() - end)) if use_cuda: (inputs, targets) = (inputs.cuda(), targets.cuda()) (inputs, targets) = (torch.autograd.Variable(inputs), torch.autograd.Variable(targets)) conv_outputs = model(inputs) outputs = global_avg_pool(conv_outputs) loss = criterion(outputs, targets) losses.update(loss.data.item(), inputs.size(0)) if (len(outputs.size()) <= 2): (prec1, prec5) = accuracy(outputs.data, targets.data, topk=(1, 5)) top1.update(prec1.item(), inputs.size(0)) top5.update(prec5.item(), inputs.size(0)) elif (outputs.data.size(2) > 1): (prec1, prec5) = accuracy(outputs.data.narrow(2, 0, 1).squeeze(), targets.data, topk=(1, 5)) live_top1.update(prec1.item(), inputs.size(0)) live_top5.update(prec5.item(), inputs.size(0)) (prec1, prec5) = accuracy(outputs.data.narrow(2, 1, 1).squeeze(), targets.data, topk=(1, 5)) dead_top1.update(prec1.item(), inputs.size(0)) dead_top5.update(prec5.item(), inputs.size(0)) outputs = (outputs.data.narrow(2, 0, 1) + outputs.data.narrow(2, 1, 1)) (prec1, prec5) = accuracy(outputs.squeeze(), targets.data, topk=(1, 5)) top1.update(prec1.item(), inputs.size(0)) top5.update(prec5.item(), inputs.size(0)) else: (prec1, prec5) = accuracy(outputs.data.narrow(2, 0, 1).squeeze(), targets.data, topk=(1, 5)) top1.update(prec1.item(), inputs.size(0)) top5.update(prec5.item(), inputs.size(0)) batch_time.update((time.time() - end)) end = time.time() bar.suffix = '({batch}/{size}) Data: {data:.3f}s | Batch: {bt:.3f}s | Total: {total:} | ETA: {eta:} | Loss: {loss:.4f} | top1: {top1: .4f} | top5: {top5: .4f} | live_top1: {ltop1: .4f} | dead_top1: {dtop1: .4f}'.format(batch=(batch_idx + 1), size=len(testloader), data=data_time.avg, bt=batch_time.avg, total=bar.elapsed_td, eta=bar.eta_td, loss=losses.avg, top1=top1.avg, top5=top5.avg, ltop1=live_top1.avg, dtop1=dead_top1.avg) bar.next() bar.finish() return (losses.avg, top1.avg)
def difficulty_judgement_ratings(df): ev1_difficulty = df['Question Difficulty_EV_1'].dropna().apply(map_difficulty) ev2_difficulty = df['Question Difficulty_EV_2'].dropna().apply(map_difficulty) ev_difficulty = (ev1_difficulty + ev2_difficulty) print(f'EV1 Difficulty: {ev1_difficulty.mean()} / 4') print(f'EV2 Difficulty: {ev2_difficulty.mean()} / 4') print(f'Average Difficulty: {(ev_difficulty.mean() / 2)} / 4')
class PD_Stats(object): def __init__(self, path, columns): self.path = path if os.path.isfile(self.path): self.stats = pd.read_pickle(self.path) assert (list(self.stats.columns) == list(columns)) else: self.stats = pd.DataFrame(columns=columns) def update(self, row, save=True): self.stats.loc[len(self.stats.index)] = row if save: self.stats.to_pickle(self.path)
class EncDecBaseConfig(FairseqDataclass): embed_path: Optional[str] = field(default=None, metadata={'help': 'path to pre-trained embedding'}) embed_dim: Optional[int] = field(default=512, metadata={'help': 'embedding dimension'}) ffn_embed_dim: int = field(default=2048, metadata={'help': 'embedding dimension for FFN'}) layers: int = field(default=6, metadata={'help': 'number of layers'}) attention_heads: int = field(default=8, metadata={'help': 'number of attention heads'}) normalize_before: bool = field(default=False, metadata={'help': 'apply layernorm before each block'}) learned_pos: bool = field(default=False, metadata={'help': 'use learned positional embeddings'}) layerdrop: float = field(default=0, metadata={'help': 'LayerDrop probability'}) layers_to_keep: Optional[List[int]] = field(default=None, metadata={'help': 'which layers to *keep* when pruning'}) xformers_att_config: Optional[str] = field(default=None, metadata={'help': 'config for xFormers attention, defined in xformers.components.attention.AttentionConfig'})
.parametrize('device', list_devices()) def test_tensormap_modify(device): tm = o3d.t.geometry.TensorMap('positions') tm.a = o3c.Tensor([100], device=device) a_alias = tm.a a_alias[:] = o3c.Tensor([200], device=device) np.testing.assert_equal(a_alias.cpu().numpy(), [200]) np.testing.assert_equal(tm.a.cpu().numpy(), [200]) tm = o3d.t.geometry.TensorMap('positions') tm.a = o3c.Tensor([100], device=device) a_alias = tm.a tm.a[:] = o3c.Tensor([200], device=device) np.testing.assert_equal(a_alias.cpu().numpy(), [200]) np.testing.assert_equal(tm.a.cpu().numpy(), [200]) tm = o3d.t.geometry.TensorMap('positions') tm.a = o3c.Tensor([100], device=device) a_alias = tm.a a_alias = o3c.Tensor([200], device=device) np.testing.assert_equal(a_alias.cpu().numpy(), [200]) np.testing.assert_equal(tm.a.cpu().numpy(), [100]) tm = o3d.t.geometry.TensorMap('positions') tm.a = o3c.Tensor([100], device=device) a_alias = tm.a tm.a = o3c.Tensor([200], device=device) np.testing.assert_equal(a_alias.cpu().numpy(), [100]) np.testing.assert_equal(tm.a.cpu().numpy(), [200]) tm = o3d.t.geometry.TensorMap('positions') tm.a = o3c.Tensor([100], device=device) a_alias = tm.a assert (id(a_alias) != id(tm.a)) tm = o3d.t.geometry.TensorMap('positions') tm.a = o3c.Tensor([100], device=device) a_alias = tm.a assert (len(tm) == 1) del tm.a assert (len(tm) == 0) np.testing.assert_equal(a_alias.cpu().numpy(), [100]) a_alias[:] = 200 np.testing.assert_equal(a_alias.cpu().numpy(), [200]) tm = o3d.t.geometry.TensorMap('positions') tm.a = o3c.Tensor([100], device=device) tm.b = o3c.Tensor([200], device=device) a_alias = tm.a b_alias = tm.b (tm.a, tm.b) = (tm.b, tm.a) np.testing.assert_equal(a_alias.cpu().numpy(), [100]) np.testing.assert_equal(b_alias.cpu().numpy(), [200]) np.testing.assert_equal(tm.a.cpu().numpy(), [200]) np.testing.assert_equal(tm.b.cpu().numpy(), [100])
def accuracy(dataset, model): total = 0 for feat in dataset: feature = feat[0] feature = torch.tensor(feature, dtype=torch.float) y_pred = model(feature) y_true = feat[1] loss = custom_loss(y_pred, y_true, model.name) total += loss return (total / len(dataset))
class QNet(BaseModule): def __init__(self, n_units, n_classes): super(QNet, self).__init__() self.model = nn.Sequential(nn.Linear((2 * n_classes), n_units), nn.ReLU(True), nn.Linear(n_units, n_classes)) def forward(self, zcat): zzt = self.model(zcat) return zzt
def _valid_accuracy_field(key, scope, error): assert (bool(('relative' in scope['accuracy_criterion'])) != bool(('absolute' in scope['accuracy_criterion'])))
.skipif((not hasattr(m, 'has_exp_optional')), reason='no <experimental/optional>') def test_exp_optional(): assert (m.double_or_zero_exp(None) == 0) assert (m.double_or_zero_exp(42) == 84) pytest.raises(TypeError, m.double_or_zero_exp, 'foo') assert (m.half_or_none_exp(0) is None) assert (m.half_or_none_exp(42) == 21) pytest.raises(TypeError, m.half_or_none_exp, 'foo') assert (m.test_nullopt_exp() == 42) assert (m.test_nullopt_exp(None) == 42) assert (m.test_nullopt_exp(42) == 42) assert (m.test_nullopt_exp(43) == 43) assert (m.test_no_assign_exp() == 42) assert (m.test_no_assign_exp(None) == 42) assert (m.test_no_assign_exp(m.NoAssign(43)) == 43) pytest.raises(TypeError, m.test_no_assign_exp, 43) holder = m.OptionalExpHolder() mvalue = holder.member assert mvalue.initialized assert holder.member_initialized() props = m.OptionalExpProperties() assert (int(props.access_by_ref) == 42) assert (int(props.access_by_copy) == 42)
def get_gcc_version(run_lambda): return run_and_parse_first_match(run_lambda, 'gcc --version', 'gcc (.*)')
class BalancedDataParallel(DataParallel): def __init__(self, gpu0_bsz, *args, **kwargs): self.gpu0_bsz = gpu0_bsz super().__init__(*args, **kwargs) def forward(self, *inputs, **kwargs): if (not self.device_ids): return self.module(*inputs, **kwargs) if (self.gpu0_bsz == 0): device_ids = self.device_ids[1:] else: device_ids = self.device_ids (inputs, kwargs) = self.scatter(inputs, kwargs, device_ids) if (len(self.device_ids) == 1): return self.module(*inputs[0], **kwargs[0]) replicas = self.replicate(self.module, self.device_ids) if (self.gpu0_bsz == 0): replicas = replicas[1:] outputs = self.parallel_apply(replicas, device_ids, inputs, kwargs) return self.gather(outputs, self.output_device) def parallel_apply(self, replicas, device_ids, inputs, kwargs): return parallel_apply(replicas, inputs, kwargs, device_ids) def scatter(self, inputs, kwargs, device_ids): bsz = inputs[0].size(self.dim) num_dev = len(self.device_ids) gpu0_bsz = self.gpu0_bsz bsz_unit = ((bsz - gpu0_bsz) // (num_dev - 1)) if (gpu0_bsz < bsz_unit): chunk_sizes = ([gpu0_bsz] + ([bsz_unit] * (num_dev - 1))) delta = (bsz - sum(chunk_sizes)) for i in range(delta): chunk_sizes[(i + 1)] += 1 if (gpu0_bsz == 0): chunk_sizes = chunk_sizes[1:] else: return super().scatter(inputs, kwargs, device_ids) return scatter_kwargs(inputs, kwargs, device_ids, chunk_sizes, dim=self.dim)
def TreeGeneration(ArgSet, depth: int, equiv: bool, rarity: bool): for LArgSet in range(0, (ArgSet + 1)): if ((ArgSet & LArgSet) == LArgSet): RArgSet = (LArgSet ^ ArgSet) for ldepth in range(0, depth): rdepth = ((depth - 1) - ldepth) for Ltree in TreeDB.getTreeSet(ldepth, LArgSet): for Rtree in TreeDB.getTreeSet(rdepth, RArgSet): valid_op_set = ['+', '-', '*', '/', '%', 'min', 'max'] for op in valid_op_set: if Unique(op, Ltree, Rtree, equiv, rarity): new_tree = ArithExpTree(op=op, lson=Ltree, rson=Rtree, ArgSet=ArgSet) TreeDB.Add(depth, ArgSet, new_tree)
class Dataset(data.Dataset): def __init__(self, root, load_bytes=False, transform=None, class_map=''): class_to_idx = None if class_map: class_to_idx = load_class_map(class_map, root) (images, class_to_idx) = find_images_and_targets(root, class_to_idx=class_to_idx) if (len(images) == 0): raise RuntimeError(((('Found 0 images in subfolders of: ' + root) + '\nSupported image extensions are: ') + ','.join(IMG_EXTENSIONS))) self.root = root self.samples = images self.imgs = self.samples self.class_to_idx = class_to_idx self.load_bytes = load_bytes self.transform = transform def __getitem__(self, index): (path, target) = self.samples[index] img = (open(path, 'rb').read() if self.load_bytes else Image.open(path).convert('RGB')) if (self.transform is not None): img = self.transform(img) if (target is None): target = torch.zeros(1).long() return (img, target) def __len__(self): return len(self.imgs) def filenames(self, indices=[], basename=False): if indices: if basename: return [os.path.basename(self.samples[i][0]) for i in indices] else: return [self.samples[i][0] for i in indices] elif basename: return [os.path.basename(x[0]) for x in self.samples] else: return [x[0] for x in self.samples]
def t_integ_attr_(b): res = np.zeros(b.shape) for i in range(b.shape[1]): res[0][i] = integrate.quad(t_attr, 0, b[0][i], points=[0])[0] return res