code
stringlengths
101
5.91M
def gen_config_yaml(manifest_root: Path, spm_filename: str, yaml_filename: str='config.yaml', specaugment_policy: str='lb', prepend_tgt_lang_tag: bool=False, sampling_alpha: float=1.0, audio_root: str=''): manifest_root = manifest_root.absolute() writer = S2TDataConfigWriter((manifest_root / yaml_filename)) writer.set_vocab_filename(spm_filename.replace('.model', '.txt')) writer.set_input_channels(1) writer.set_input_feat_per_channel(80) specaugment_setters = {'lb': writer.set_specaugment_lb_policy, 'ld': writer.set_specaugment_ld_policy, 'sm': writer.set_specaugment_sm_policy, 'ss': writer.set_specaugment_ss_policy} specaugment_setter = specaugment_setters.get(specaugment_policy, None) if (specaugment_setter is not None): specaugment_setter() writer.set_bpe_tokenizer({'bpe': 'sentencepiece', 'sentencepiece_model': (manifest_root / spm_filename).as_posix()}) if prepend_tgt_lang_tag: writer.set_prepend_tgt_lang_tag(True) writer.set_sampling_alpha(sampling_alpha) writer.set_feature_transforms('_train', ['utterance_cmvn', 'specaugment']) writer.set_feature_transforms('*', ['utterance_cmvn']) if (len(audio_root) > 0): writer.set_audio_root(audio_root) writer.flush()
def simple_inference(model, input): with torch.no_grad(): if isinstance(input, (dict, UserDict)): output = model(**input) elif isinstance(input, (list, tuple)): try: output = model(*input) except: output = model(input) else: output = model(input) return output
def CheckSpacingForFunctionCall(filename, clean_lines, linenum, error): line = clean_lines.elided[linenum] fncall = line for pattern in ('\\bif\\s*\\((.*)\\)\\s*{', '\\bfor\\s*\\((.*)\\)\\s*{', '\\bwhile\\s*\\((.*)\\)\\s*[{;]', '\\bswitch\\s*\\((.*)\\)\\s*{'): match = Search(pattern, line) if match: fncall = match.group(1) break if ((not Search('\\b(if|for|while|switch|return|new|delete|catch|sizeof)\\b', fncall)) and (not Search(' \\([^)]+\\)\\([^)]*(\\)|,$)', fncall)) and (not Search(' \\([^)]+\\)\\[[^\\]]+\\]', fncall))): if Search('\\w\\s*\\(\\s(?!\\s*\\\\$)', fncall): error(filename, linenum, 'whitespace/parens', 4, 'Extra space after ( in function call') elif Search('\\(\\s+(?!(\\s*\\\\)|\\()', fncall): error(filename, linenum, 'whitespace/parens', 2, 'Extra space after (') if (Search('\\w\\s+\\(', fncall) and (not Search('#\\s*define|typedef|using\\s+\\w+\\s*=', fncall)) and (not Search('\\w\\s+\\((\\w+::)*\\*\\w+\\)\\(', fncall)) and (not Search('\\bcase\\s+\\(', fncall))): if Search('\\boperator_*\\b', line): error(filename, linenum, 'whitespace/parens', 0, 'Extra space before ( in function call') else: error(filename, linenum, 'whitespace/parens', 4, 'Extra space before ( in function call') if Search('[^)]\\s+\\)\\s*[^{\\s]', fncall): if Search('^\\s+\\)', fncall): error(filename, linenum, 'whitespace/parens', 2, 'Closing ) should be moved to the previous line') else: error(filename, linenum, 'whitespace/parens', 2, 'Extra space before )')
def sparsenet201(**kwargs): return get_sparsenet(num_layers=201, model_name='sparsenet201', **kwargs)
def multitask_text_transformer_decoder_arch(args, decoder_layers, decoder_embed_dim=256, decoder_attention_heads=4): args.decoder_layers = decoder_layers args.decoder_embed_dim = decoder_embed_dim args.decoder_attention_heads = decoder_attention_heads base_multitask_text_transformer_decoder_arch(args)
def unpack_data(dataB, device='cuda'): if is_multidata(dataB): if torch.is_tensor(dataB[0]): if torch.is_tensor(dataB[1]): return dataB[0].to(device) elif is_multidata(dataB[1]): return (dataB[0].to(device), dataB[1][0].to(device)) else: raise RuntimeError('Invalid data format {} -- check your dataloader!'.format(type(dataB[1]))) elif is_multidata(dataB[0]): return [d.to(device) for d in list(zip(*dataB))[0]] else: raise RuntimeError('Invalid data format {} -- check your dataloader!'.format(type(dataB[0]))) elif torch.is_tensor(dataB): return dataB.to(device) else: raise RuntimeError('Invalid data format {} -- check your dataloader!'.format(type(dataB)))
def distribute_presets(prefixes, scaffolding, config_updates): for (path, value) in iterate_flattened(config_updates): (scaffold_name, suffix) = find_best_match(path, prefixes) scaff = scaffolding[scaffold_name] set_by_dotted_path(scaff.presets, suffix, value)
def config(): seed = 0 test_mode = False dataset_name = None hyperparameters = None evaluation_metric = None minimize = None total_trials = None parameterization = None
def quaddobl_initialize(nbt, dim, wnd, dir, err): from phcpy.phcpy2c3 import py2c_numbtrop_quaddobl_initialize as store flat = [] for vec in dir: flat = (flat + vec) data = ((wnd + flat) + err) store(nbt, dim, str(data))
def fix_cam_drop_frames(seq_path, label_names): ts_path = os.path.join(seq_path, camera_configs['time_stamp_name']) try: with open(ts_path) as ts_f: ts = ts_f.readlines() except: return label_names n_labels = len(ts) if (int((float(ts[(- 1)].rstrip()) * camera_configs['frame_rate'])) == (n_labels - 1)): return label_names label_names_new = ([None] * n_labels) for (idx, line) in enumerate(ts): time = float(line.rstrip()) real_id = int((time * camera_configs['frame_rate'])) if (real_id < n_labels): label_names_new[real_id] = label_names[idx] prev_nearest = (- np.ones((n_labels,), dtype=int)) post_nearest = (- np.ones((n_labels,), dtype=int)) prev_flag = (- 1) post_flag = n_labels for (idx, label) in enumerate(label_names_new): if (label is not None): prev_flag = idx else: prev_nearest[idx] = prev_flag for (idx, label) in reversed(list(enumerate(label_names_new))): if (label is not None): post_flag = idx else: post_nearest[idx] = post_flag for idx in range(n_labels): if ((prev_nearest[idx] >= 0) and (prev_nearest[idx] < n_labels) and (post_nearest[idx] >= 0) and (post_nearest[idx] < n_labels)): if ((idx - prev_nearest[idx]) <= (post_nearest[idx] - idx)): sup_idx = prev_nearest[idx] else: sup_idx = post_nearest[idx] elif (not ((prev_nearest[idx] >= 0) and (prev_nearest[idx] < n_labels))): sup_idx = post_nearest[idx] elif (not ((post_nearest[idx] >= 0) and (post_nearest[idx] < n_labels))): sup_idx = prev_nearest[idx] else: sup_idx = None if (label_names_new[idx] is None): if ((sup_idx >= 0) and (sup_idx < n_labels)): label_names_new[idx] = label_names_new[sup_idx] else: raise ValueError return label_names_new
def get_grad_norm(params, scale=1): total_norm = 0.0 for p in params: if (p.grad is not None): param_norm = (p.grad.detach().data / scale).norm(2) total_norm += (param_norm.item() ** 2) total_norm = (total_norm ** 0.5) return total_norm
def add_packages(config, repeat=1): train_dir = 'train_package' package_dir = path.realpath(__file__).replace('pgportfolio/autotrain/generate.pyc', train_dir).replace('pgportfolio\\autotrain\\generate.pyc', train_dir).replace('pgportfolio/autotrain/generate.py', train_dir).replace('pgportfolio\\autotrain\\generate.py', train_dir) all_subdir = [int(s) for s in os.listdir(package_dir) if os.path.isdir(((package_dir + '/') + s))] if all_subdir: max_dir_num = max(all_subdir) else: max_dir_num = 0 indexes = [] for i in range(repeat): max_dir_num += 1 directory = ((package_dir + '/') + str(max_dir_num)) config['random_seed'] = i os.makedirs(directory) indexes.append(max_dir_num) with open(((directory + '/') + 'net_config.json'), 'w') as outfile: json.dump(config, outfile, indent=4, sort_keys=True) logging.info(('create indexes %s' % indexes)) return indexes
def parse_args(): parser = argparse.ArgumentParser(description='Finetune a transformers model on a text classification task') parser.add_argument('--dataset_name', type=str, default=None, help='The name of the dataset to use (via the datasets library).') parser.add_argument('--dataset_config_names', nargs='+', type=str, required=True, help='The configuration names of the dataset to use (via the datasets library).') parser.add_argument('--dataset_split_names', nargs='+', type=str, required=True, help='The names of the training data set splits to use (via the datasets library).') parser.add_argument('--preprocessing_num_workers', type=int, default=None, help='The number of processes to use for the preprocessing.') parser.add_argument('--overwrite_cache', action='store_true', help='Overwrite the cached training and evaluation sets') parser.add_argument('--preprocessing_only', action='store_true', help='Only run the preprocessing script to be cached for future use') parser.add_argument('--cache_dir', type=str, default=None, help='Where do you want to store the pretrained models downloaded from huggingface.co') parser.add_argument('--validation_split_percentage', type=int, default=1, help='Percentage of training data that should be used for validation if no validation is present in dataset.') parser.add_argument('--logging_steps', type=int, default=500, help='Number of steps between each logging') parser.add_argument('--saving_steps', type=int, default=500, help='Number of steps between each logging') parser.add_argument('--audio_column_name', type=str, default='audio', help="Column in the dataset that contains speech file path. Defaults to 'audio'") parser.add_argument('--model_name_or_path', type=str, help='Path to pretrained model or model identifier from huggingface.co/models.', required=True) parser.add_argument('--config_name', type=str, default=None, help='Pretrained config name or path if not the same as model_name') parser.add_argument('--train_cache_file_name', type=str, default=None, help='Path to the train cached file name') parser.add_argument('--validation_cache_file_name', type=str, default=None, help='Path to the validation cached file name') parser.add_argument('--per_device_train_batch_size', type=int, default=8, help='Batch size (per device) for the training dataloader.') parser.add_argument('--per_device_eval_batch_size', type=int, default=8, help='Batch size (per device) for the evaluation dataloader.') parser.add_argument('--learning_rate', type=float, default=5e-05, help='Initial learning rate (after the potential warmup period) to use.') parser.add_argument('--weight_decay', type=float, default=0.0, help='Weight decay to use.') parser.add_argument('--num_train_epochs', type=int, default=3, help='Total number of training epochs to perform.') parser.add_argument('--max_train_steps', type=int, default=None, help='Total number of training steps to perform. If provided, overrides num_train_epochs.') parser.add_argument('--gradient_accumulation_steps', type=int, default=1, help='Number of updates steps to accumulate before performing a backward/update pass.') parser.add_argument('--gradient_checkpointing', action='store_true', help='If True, use gradient checkpointing to save memory at the expense of slower backward pass.') parser.add_argument('--lr_scheduler_type', type=SchedulerType, default='linear', help='The scheduler type to use.', choices=['linear', 'cosine', 'cosine_with_restarts', 'polynomial', 'constant', 'constant_with_warmup']) parser.add_argument('--num_warmup_steps', type=int, default=0, help='Number of steps for the warmup in the lr scheduler.') parser.add_argument('--output_dir', type=str, default=None, help='Where to store the final model.') parser.add_argument('--seed', type=int, default=0, help='A seed for reproducible training.') parser.add_argument('--max_gumbel_temperature', type=float, default=2.0, help='Maximum temperature for gumbel softmax.') parser.add_argument('--min_gumbel_temperature', type=float, default=0.5, help='Minimum temperature for gumbel softmax.') parser.add_argument('--gumbel_temperature_decay', type=float, default=0.999995, help='Decay of gumbel temperature during training.') parser.add_argument('--max_duration_in_seconds', type=float, default=5.0, help='Filter out audio files that are longer than `max_duration_in_seconds` seconds') parser.add_argument('--min_duration_in_seconds', type=float, default=3.0, help='Filter out audio files that are shorter than `min_duration_in_seconds` seconds') parser.add_argument('--pad_to_multiple_of', type=int, default=None, help='If set will pad the sequence to a multiple of the provided value. This is especially useful to enable the use of Tensor Cores on NVIDIA hardware with compute capability >= 7.5 (Volta).') parser.add_argument('--adam_beta1', type=float, default=0.9, help='Beta1 for AdamW optimizer') parser.add_argument('--adam_beta2', type=float, default=0.999, help='Beta2 for AdamW optimizer') parser.add_argument('--adam_epsilon', type=float, default=1e-08, help='Epsilon for AdamW optimizer') parser.add_argument('--push_to_hub', action='store_true', help='Whether or not to push the model to the Hub.') parser.add_argument('--hub_model_id', type=str, help='The name of the repository to keep in sync with the local `output_dir`.') parser.add_argument('--hub_token', type=str, help='The token to use to push to the Model Hub.') parser.add_argument('--mask_time_prob', type=float, default=None, help='Percentage (between 0 and 1) of all feature vectors along the time axis which will be masked in the contrastive task. If omitted, will pull value from model config.') parser.add_argument('--mask_time_length', type=int, default=None, help='Length of each vector mask span to mask along the time axis in the contrastive task. If omitted, will pull value from model config.') args = parser.parse_args() if args.push_to_hub: assert (args.output_dir is not None), 'Need an `output_dir` to create a repo when `--push_to_hub` is passed.' if (args.output_dir is not None): os.makedirs(args.output_dir, exist_ok=True) return args
def calculate_auc(model, mbs_list, shuffle=True): model.eval() y_real = [] y_hat = [] if shuffle: random.shuffle(mbs_list) for (i, batch) in enumerate(mbs_list): (output, label_tensor) = model(batch) y_hat.extend(output.cpu().data.view((- 1)).numpy()) y_real.extend(label_tensor.cpu().data.view((- 1)).numpy()) auc = roc_auc_score(y_real, y_hat) return (auc, y_real, y_hat)
def recursive_merge_dicts(*args): if (not args): return dict() q = args[0] for p in args[1:]: q = recursive_merge_2dicts(q, p) return q
class TestChatCache(unittest.TestCase): def setUp(self): return super().setUp() def tearDown(self) -> None: if (os.path.exists('./gptcache_data') and os.path.isdir('./gptcache_data')): try: shutil.rmtree('./gptcache_data') print(f'The directory gptcache_data has been successfully deleted.') except Exception as e: print(f'An error occurred while deleting the directory: {e}') else: print(f'The directory gptcache_data does not exist.') return super().tearDown() def test_chat_cache(self): cache_plugin = ChatCache(embedding_model_dir='hkunlp/instructor-large') cache_plugin.init_similar_cache_from_config() prompt = 'Tell me about Intel Xeon Scalable Processors.' config = PipelineConfig(model_name_or_path='facebook/opt-125m') chatbot = build_chatbot(config) response = chatbot.predict(prompt) cache_plugin.put(prompt, response) answer = cache_plugin.get(prompt) self.assertIn('Intel Xeon Scalable', str(answer['choices'][0]['text']))
class Graph(): def __init__(self, parent_map, children_map, id_list): self.parent_map = parent_map self.children_map = children_map self.id_list = id_list def topoligical_sort(self): order = [] next = [] for id in self.id_list: if ((len(self.parent_map[id]) == 0) and (id not in order)): order.append(id) for child_id in self.children_map[id]: if ((child_id not in next) and (child_id not in order)): next.append(child_id) while (len(next) != 0): id = next.pop(0) order.append(id) for child_id in self.children_map[id]: if ((child_id not in next) and (child_id not in order)): next.append(child_id) order.reverse() return order
def ResNet18(winogradArgs: dict=None, quantArgs: dict=None, miscArgs: dict=None, num_classes: int=10, mult: int=1.0): return ResNet(BasicBlock, [2, 2, 2, 2], num_classes, winogradArgs=winogradArgs, quantization=quantArgs, miscArgs=miscArgs, multiplier=mult)
def apply_rotary_pos_emb_single(x, cos, sin, position_ids): cos = cos.squeeze(1).squeeze(0) sin = sin.squeeze(1).squeeze(0) cos = cos[position_ids].unsqueeze(1) sin = sin[position_ids].unsqueeze(1) x_embed = ((x * cos) + (rotate_half(x) * sin)) return x_embed
def discriminator(image, options, n_scale=2, reuse=False, name='discriminator'): images = [] for i in range(n_scale): images.append(tf.image.resize_bicubic(image, [(get_shape(image)[1] // (2 ** i)), (get_shape(image)[2] // (2 ** i))])) with tf.variable_scope(name): if reuse: tf.get_variable_scope().reuse_variables() else: assert (tf.get_variable_scope().reuse is False) images = dis_down(images, 4, 2, n_scale, options.df_dim, 'd_h0_conv_scale_') images = dis_down(images, 4, 2, n_scale, (options.df_dim * 2), 'd_h1_conv_scale_') images = dis_down(images, 4, 2, n_scale, (options.df_dim * 4), 'd_h2_conv_scale_') images = dis_down(images, 4, 2, n_scale, (options.df_dim * 8), 'd_h3_conv_scale_') images = final_conv(images, n_scale, 'd_pred_scale_') return images
def label2onehot(label, length): onehot = np.zeros(length) onehot[label] = 1 return onehot
class PseudoLabel(): def __init__(self, cfg): (h, w) = cfg.INPUT.TARGET_INPUT_SIZE_TRAIN self.prob_tar = np.zeros([1, h, w]) self.label_tar = np.zeros([1, h, w]) self.thres = [] self.number_class = cfg.MODEL.NUM_CLASSES self.out_dir = cfg.OUTPUT_DIR self.iter = 0 def save_results(self): np.save(os.path.join(self.out_dir, 'thres_const.npy'), self.thres) print('save done.') def update_pseudo_label(self, input): input = F.softmax(input.detach(), dim=1) (prob, label) = torch.max(input, dim=1) prob_np = prob.cpu().numpy() label_np = label.cpu().numpy() print(self.iter) if (self.iter == 0): self.prob_tar = prob_np self.label_tar = label_np else: self.prob_tar = np.append(self.prob_tar, prob_np, axis=0) self.label_tar = np.append(self.label_tar, label_np, axis=0) self.iter += 1 def get_threshold_const(self, thred, percent=0.5): for i in range(self.number_class): x = self.prob_tar[(self.label_tar == i)] if (len(x) == 0): self.thres.append(0) continue x = np.sort(x) self.thres.append(x[np.int(np.round((len(x) * percent)))]) self.thres = np.array(self.thres) self.thres[(self.thres > thred)] = thred return self.thres
def inception_v4(inputs, num_classes=1001, is_training=True, dropout_keep_prob=0.8, reuse=None, scope='InceptionV4', create_aux_logits=True): end_points = {} with tf.variable_scope(scope, 'InceptionV4', [inputs], reuse=reuse) as scope: with slim.arg_scope([slim.batch_norm, slim.dropout], is_training=is_training): (net, end_points) = inception_v4_base(inputs, scope=scope) with slim.arg_scope([slim.conv2d, slim.max_pool2d, slim.avg_pool2d], stride=1, padding='SAME'): if create_aux_logits: with tf.variable_scope('AuxLogits'): aux_logits = end_points['Mixed_6h'] aux_logits = slim.avg_pool2d(aux_logits, [5, 5], stride=3, padding='VALID', scope='AvgPool_1a_5x5') aux_logits = slim.conv2d(aux_logits, 128, [1, 1], scope='Conv2d_1b_1x1') aux_logits = slim.conv2d(aux_logits, 768, aux_logits.get_shape()[1:3], padding='VALID', scope='Conv2d_2a') aux_logits = slim.flatten(aux_logits) aux_logits = slim.fully_connected(aux_logits, num_classes, activation_fn=None, scope='Aux_logits') end_points['AuxLogits'] = aux_logits with tf.variable_scope('Logits'): net = slim.avg_pool2d(net, net.get_shape()[1:3], padding='VALID', scope='AvgPool_1a') net = slim.dropout(net, dropout_keep_prob, scope='Dropout_1b') net = slim.flatten(net, scope='PreLogitsFlatten') end_points['PreLogitsFlatten'] = net logits = slim.fully_connected(net, num_classes, activation_fn=None, scope='Logits') end_points['Logits'] = logits end_points['Predictions'] = tf.nn.softmax(logits, name='Predictions') return (logits, end_points)
def filter_opt(opt, tag): ret = {} for (k, v) in opt.items(): tokens = k.split('.') if (tokens[0] == tag): ret['.'.join(tokens[1:])] = v return ret
def evaluate_imagenet(gpu, encoder_usage_info, downstream_dataset, encoder, reference_label, trigger, reference, key='clean'): cmd = f'nohup python3 -u training_downstream_classifier.py --encoder_usage_info {encoder_usage_info} --dataset {downstream_dataset} --trigger_file {trigger} --encoder {encoder} --reference_label {reference_label} --reference_file ./reference/imagenet/{reference}.npz --gpu {gpu} >./log/imagenet/evaluation_{key}_{downstream_dataset}.txt &' os.system(cmd)
def FunctionCorrelation(tenFirst, tenSecond, intStride): return _FunctionCorrelation.apply(tenFirst, tenSecond, intStride)
class BasicBlockSig(nn.Module): def __init__(self, in_channels, out_channels, init='xavier', ksize=3, stride=1, pad=1): super(BasicBlockSig, self).__init__() self.body = nn.Sequential(nn.Conv2d(in_channels, out_channels, ksize, stride, pad), nn.Sigmoid()) def forward(self, x): out = self.body(x) return out
def ten_problems(): result = [] b0 = ([[3, 6, 7, 8] for _ in range(7)] + [[4, 6, 7, 8] for _ in range(2)]) r0 = 231 result.append((b0, r0)) b1 = ([[2, 5, 6, 8], [3, 6, 7, 8], [4, 5, 7, 8]] + [[4, 6, 7, 8] for _ in range(7)]) r1 = 294 result.append((b1, r1)) b2 = (([[2, 4, 7, 8]] + [[3, 6, 7, 8] for _ in range(2)]) + [[4, 6, 7, 8] for _ in range(7)]) r2 = 356 result.append((b2, r2)) b3 = ([[2, 5, 7, 8] for _ in range(2)] + [[4, 6, 7, 8] for _ in range(8)]) r3 = 398 result.append((b3, r3)) b4 = ((([[3, 5, 7, 8]] + [[4, 5, 7, 8] for _ in range(2)]) + [[3, 6, 7, 8] for _ in range(2)]) + [[4, 6, 7, 8] for _ in range(5)]) r4 = 526 result.append((b4, r4)) b5 = (([[4, 5, 7, 8] for _ in range(3)] + [[3, 6, 7, 8] for _ in range(2)]) + [[4, 6, 7, 8] for _ in range(6)]) r5 = 734 result.append((b5, r5)) b6 = (([[3, 5, 7, 8]] + [[3, 6, 7, 8] for _ in range(3)]) + [[4, 6, 7, 8] for _ in range(7)]) r6 = 1104 result.append((b6, r6)) b7 = ([[3, 6, 7, 8] for _ in range(4)] + [[4, 6, 7, 8] for _ in range(8)]) r7 = 1600 result.append((b7, r7)) b8 = (([[3, 5, 7, 8]] + [[3, 6, 7, 8] for _ in range(2)]) + [[4, 6, 7, 8] for _ in range(9)]) r8 = 2166 result.append((b8, r8)) b9 = ([[3, 6, 7, 8] for _ in range(3)] + [[4, 6, 7, 8] for _ in range(10)]) r9 = 3102 result.append((b9, r9)) return result
def calculate_desired_noise_rms(clean_rms, snr): a = (float(snr) / 20) noise_rms = (clean_rms / (10 ** a)) return noise_rms
def save_dataset(train_, test_, filename): torch.save({'train': train_, 'test': test_}, filename)
_module class FastRCNN(TwoStageDetector): def __init__(self, backbone, bbox_roi_extractor, bbox_head, train_cfg, test_cfg, neck=None, shared_head=None, mask_roi_extractor=None, mask_head=None, pretrained=None): super(FastRCNN, self).__init__(backbone=backbone, neck=neck, shared_head=shared_head, bbox_roi_extractor=bbox_roi_extractor, bbox_head=bbox_head, train_cfg=train_cfg, test_cfg=test_cfg, mask_roi_extractor=mask_roi_extractor, mask_head=mask_head, pretrained=pretrained) def forward_test(self, imgs, img_metas, proposals, **kwargs): for (var, name) in [(imgs, 'imgs'), (img_metas, 'img_metas')]: if (not isinstance(var, list)): raise TypeError('{} must be a list, but got {}'.format(name, type(var))) num_augs = len(imgs) if (num_augs != len(img_metas)): raise ValueError('num of augmentations ({}) != num of image meta ({})'.format(len(imgs), len(img_metas))) imgs_per_gpu = imgs[0].size(0) assert (imgs_per_gpu == 1) if (num_augs == 1): return self.simple_test(imgs[0], img_metas[0], proposals[0], **kwargs) else: return self.aug_test(imgs, img_metas, proposals, **kwargs)
def remove_email(text): subtext = text.split(' ') sts = [] for (i, st) in enumerate(subtext): st = st.strip() st = re.sub('([a-zA-Z0-9_.+-]+[a-zA-Z0-9-]+\\.[a-zA-Z0-9-.]+)', f'', st, flags=re.MULTILINE) sts.append(st) return ' '.join(sts)
def gelu_fast(x): if (not hasattr(gelu_fast, '_a')): gelu_fast._a = math.sqrt((2 / math.pi)) return ((0.5 * x) * (1 + torch.tanh((gelu_fast._a * (x + (0.044715 * torch.pow(x, 3)))))))
def construct_flatindex_from_embeddings(embeddings, ids=None): dim = embeddings.shape[1] print(('embedding shape: ' + str(embeddings.shape))) index = faiss.index_factory(dim, 'Flat', faiss.METRIC_INNER_PRODUCT) if (ids is not None): ids = ids.astype(np.int64) print(ids.shape, ids.dtype) index = faiss.IndexIDMap2(index) index.add_with_ids(embeddings, ids) else: index.add(embeddings) return index
class ExploreTaskDefinition(AbstractTaskDefinition): joint_positions = [0.0, (- 1.33), (- 1.8), 0.0, 1.5, 1.6] def __init__(self, *args, **kwargs): super(ExploreTaskDefinition, self).__init__(*args, **kwargs) self.addCamera(Camera('top', [(- 0.0), 0.0, 1.0], distance=0.7, roll=0.0, image_width=64, image_height=64, pitch=(- 90), yaw=0, fov=40)) def _setup(self): def _setupRobot(self, handle): if (not self.robot.mobile()): raise RuntimeError(('Exploration task does not even make sense' + 'without a mobile robot.')) self.robot.place([0, 0, 0], [0, 0, 0, 1], self.joint_positions) self.robot.arm(self.joint_positions, pb.POSITION_CONTROL) self.robot.gripper(0, pb.POSITION_CONTROL) def _makeTask(self): task = Task() return task def getName(self): return 'explore' def reset(self): pass
class PyTorchTensor(BaseTensor): __slots__ = () norms: 'NormsMethods[PyTorchTensor]' def __init__(self, raw: 'torch.Tensor'): global torch if (torch is None): torch = import_module('torch') super().__init__(raw) def raw(self) -> 'torch.Tensor': return cast(torch.Tensor, super().raw) def numpy(self: TensorType) -> Any: a = self.raw.detach().cpu().numpy() if a.flags.writeable: a.flags.writeable = False return a def item(self) -> Union[(int, float, bool)]: return self.raw.item() def shape(self) -> Shape: return self.raw.shape def reshape(self: TensorType, shape: Union[(Shape, int)]) -> TensorType: if isinstance(shape, int): shape = (shape,) return type(self)(self.raw.reshape(shape)) def astype(self: TensorType, dtype: Any) -> TensorType: return type(self)(self.raw.to(dtype)) def clip(self: TensorType, min_: float, max_: float) -> TensorType: return type(self)(self.raw.clamp(min_, max_)) def square(self: TensorType) -> TensorType: return type(self)((self.raw ** 2)) def sin(self: TensorType) -> TensorType: return type(self)(torch.sin(self.raw)) def cos(self: TensorType) -> TensorType: return type(self)(torch.cos(self.raw)) def tan(self: TensorType) -> TensorType: return type(self)(torch.tan(self.raw)) def sinh(self: TensorType) -> TensorType: return type(self)(torch.sinh(self.raw)) def cosh(self: TensorType) -> TensorType: return type(self)(torch.cosh(self.raw)) def tanh(self: TensorType) -> TensorType: return type(self)(torch.tanh(self.raw)) def arcsin(self: TensorType) -> TensorType: return type(self)(torch.asin(self.raw)) def arccos(self: TensorType) -> TensorType: return type(self)(torch.acos(self.raw)) def arctan(self: TensorType) -> TensorType: return type(self)(torch.atan(self.raw)) def arcsinh(self: TensorType) -> TensorType: return type(self)(torch.asinh(self.raw)) def arccosh(self: TensorType) -> TensorType: return type(self)(torch.acosh(self.raw)) def arctanh(self: TensorType) -> TensorType: return type(self)(torch.atanh(self.raw)) def sum(self: TensorType, axis: Optional[AxisAxes]=None, keepdims: bool=False) -> TensorType: if ((axis is None) and (not keepdims)): return type(self)(self.raw.sum()) if (axis is None): axis = tuple(range(self.ndim)) return type(self)(self.raw.sum(dim=axis, keepdim=keepdims)) def prod(self: TensorType, axis: Optional[AxisAxes]=None, keepdims: bool=False) -> TensorType: if ((axis is None) and (not keepdims)): return type(self)(self.raw.prod()) if (axis is None): axis = tuple(range(self.ndim)) elif (not isinstance(axis, Iterable)): axis = (axis,) x = self.raw for i in sorted(axis, reverse=True): x = x.prod(i, keepdim=keepdims) return type(self)(x) def mean(self: TensorType, axis: Optional[AxisAxes]=None, keepdims: bool=False) -> TensorType: if (self.raw.dtype not in [torch.float16, torch.float32, torch.float64]): raise ValueError(f'Can only calculate the mean of floating types. Got {self.raw.dtype} instead.') if ((axis is None) and (not keepdims)): return type(self)(self.raw.mean()) if (axis is None): axis = tuple(range(self.ndim)) return type(self)(self.raw.mean(dim=axis, keepdim=keepdims)) def min(self: TensorType, axis: Optional[AxisAxes]=None, keepdims: bool=False) -> TensorType: if ((axis is None) and (not keepdims)): return type(self)(self.raw.min()) if (axis is None): axis = tuple(range(self.ndim)) elif (not isinstance(axis, Iterable)): axis = (axis,) x = self.raw for i in sorted(axis, reverse=True): (x, _) = x.min(i, keepdim=keepdims) return type(self)(x) def max(self: TensorType, axis: Optional[AxisAxes]=None, keepdims: bool=False) -> TensorType: if ((axis is None) and (not keepdims)): return type(self)(self.raw.max()) if (axis is None): axis = tuple(range(self.ndim)) elif (not isinstance(axis, Iterable)): axis = (axis,) x = self.raw for i in sorted(axis, reverse=True): (x, _) = x.max(i, keepdim=keepdims) return type(self)(x) def minimum(self: TensorType, other: TensorOrScalar) -> TensorType: if isinstance(other, Tensor): other_ = other.raw elif (isinstance(other, int) or isinstance(other, float)): other_ = torch.full_like(self.raw, other) else: raise TypeError('expected x to be a Tensor, int or float') return type(self)(torch.min(self.raw, other_)) def maximum(self: TensorType, other: TensorOrScalar) -> TensorType: if isinstance(other, Tensor): other_ = other.raw elif (isinstance(other, int) or isinstance(other, float)): other_ = torch.full_like(self.raw, other) else: raise TypeError('expected x to be a Tensor, int or float') return type(self)(torch.max(self.raw, other_)) def argmin(self: TensorType, axis: Optional[int]=None) -> TensorType: return type(self)(self.raw.argmin(dim=axis)) def argmax(self: TensorType, axis: Optional[int]=None) -> TensorType: return type(self)(self.raw.argmax(dim=axis)) def argsort(self: TensorType, axis: int=(- 1)) -> TensorType: return type(self)(self.raw.argsort(dim=axis)) def sort(self: TensorType, axis: int=(- 1)) -> TensorType: return type(self)(self.raw.sort(dim=axis).values) def topk(self: TensorType, k: int, sorted: bool=True) -> Tuple[(TensorType, TensorType)]: (values, indices) = self.raw.topk(k, sorted=sorted) return (type(self)(values), type(self)(indices)) def uniform(self: TensorType, shape: ShapeOrScalar, low: float=0.0, high: float=1.0) -> TensorType: return type(self)(((torch.rand(shape, dtype=self.raw.dtype, device=self.raw.device) * (high - low)) + low)) def normal(self: TensorType, shape: ShapeOrScalar, mean: float=0.0, stddev: float=1.0) -> TensorType: return type(self)(((torch.randn(shape, dtype=self.raw.dtype, device=self.raw.device) * stddev) + mean)) def ones(self: TensorType, shape: ShapeOrScalar) -> TensorType: return type(self)(torch.ones(shape, dtype=self.raw.dtype, device=self.raw.device)) def zeros(self: TensorType, shape: ShapeOrScalar) -> TensorType: return type(self)(torch.zeros(shape, dtype=self.raw.dtype, device=self.raw.device)) def ones_like(self: TensorType) -> TensorType: return type(self)(torch.ones_like(self.raw)) def zeros_like(self: TensorType) -> TensorType: return type(self)(torch.zeros_like(self.raw)) def full_like(self: TensorType, fill_value: float) -> TensorType: return type(self)(torch.full_like(self.raw, fill_value)) def onehot_like(self: TensorType, indices: TensorType, *, value: float=1) -> TensorType: if (self.ndim != 2): raise ValueError('onehot_like only supported for 2D tensors') if (indices.ndim != 1): raise ValueError('onehot_like requires 1D indices') if (len(indices) != len(self)): raise ValueError('length of indices must match length of tensor') x = torch.zeros_like(self.raw) rows = np.arange(x.shape[0]) x[(rows, indices.raw)] = value return type(self)(x) def from_numpy(self: TensorType, a: Any) -> TensorType: return type(self)(torch.as_tensor(a, device=self.raw.device)) def _concatenate(self: TensorType, tensors: Iterable[TensorType], axis: int=0) -> TensorType: tensors_ = unwrap_(*tensors) return type(self)(torch.cat(tensors_, dim=axis)) def _stack(self: TensorType, tensors: Iterable[TensorType], axis: int=0) -> TensorType: tensors_ = unwrap_(*tensors) return type(self)(torch.stack(tensors_, dim=axis)) def transpose(self: TensorType, axes: Optional[Axes]=None) -> TensorType: if (axes is None): axes = tuple(range((self.ndim - 1), (- 1), (- 1))) return type(self)(self.raw.permute(*axes)) def all(self: TensorType, axis: Optional[AxisAxes]=None, keepdims: bool=False) -> TensorType: assert_bool(self) if ((axis is None) and (not keepdims)): return type(self)(self.raw.all()) if (axis is None): axis = tuple(range(self.ndim)) elif (not isinstance(axis, Iterable)): axis = (axis,) x = self.raw for i in sorted(axis, reverse=True): x = x.all(i, keepdim=keepdims) return type(self)(x) def any(self: TensorType, axis: Optional[AxisAxes]=None, keepdims: bool=False) -> TensorType: assert_bool(self) if ((axis is None) and (not keepdims)): return type(self)(self.raw.any()) if (axis is None): axis = tuple(range(self.ndim)) elif (not isinstance(axis, Iterable)): axis = (axis,) x = self.raw for i in sorted(axis, reverse=True): x = x.any(i, keepdim=keepdims) return type(self)(x) def logical_and(self: TensorType, other: TensorOrScalar) -> TensorType: assert_bool(self) assert_bool(other) return type(self)((self.raw & unwrap1(other))) def logical_or(self: TensorType, other: TensorOrScalar) -> TensorType: assert_bool(self) assert_bool(other) return type(self)((self.raw | unwrap1(other))) def logical_not(self: TensorType) -> TensorType: assert_bool(self) return type(self)((~ self.raw)) def exp(self: TensorType) -> TensorType: return type(self)(torch.exp(self.raw)) def log(self: TensorType) -> TensorType: return type(self)(torch.log(self.raw)) def log2(self: TensorType) -> TensorType: return type(self)(torch.log2(self.raw)) def log10(self: TensorType) -> TensorType: return type(self)(torch.log10(self.raw)) def log1p(self: TensorType) -> TensorType: return type(self)(torch.log1p(self.raw)) def tile(self: TensorType, multiples: Axes) -> TensorType: if (len(multiples) != self.ndim): raise ValueError('multiples requires one entry for each dimension') return type(self)(self.raw.repeat(multiples)) def softmax(self: TensorType, axis: int=(- 1)) -> TensorType: return type(self)(torch.nn.functional.softmax(self.raw, dim=axis)) def log_softmax(self: TensorType, axis: int=(- 1)) -> TensorType: return type(self)(torch.nn.functional.log_softmax(self.raw, dim=axis)) def squeeze(self: TensorType, axis: Optional[AxisAxes]=None) -> TensorType: if (axis is None): return type(self)(self.raw.squeeze()) if (not isinstance(axis, Iterable)): axis = (axis,) x = self.raw for i in sorted(axis, reverse=True): if (x.shape[i] != 1): raise ValueError('cannot select an axis to squeeze out which has size not equal to one') x = x.squeeze(dim=i) return type(self)(x) def expand_dims(self: TensorType, axis: int) -> TensorType: return type(self)(self.raw.unsqueeze(dim=axis)) def full(self: TensorType, shape: ShapeOrScalar, value: float) -> TensorType: if (not isinstance(shape, Iterable)): shape = (shape,) return type(self)(torch.full(shape, value, dtype=self.raw.dtype, device=self.raw.device)) def index_update(self: TensorType, indices: Any, values: TensorOrScalar) -> TensorType: (indices, values_) = unwrap_(indices, values) if isinstance(indices, tuple): indices = unwrap_(*indices) x = self.raw.clone() x[indices] = values_ return type(self)(x) def arange(self: TensorType, start: int, stop: Optional[int]=None, step: Optional[int]=None) -> TensorType: if (step is None): step = 1 if (stop is None): stop = start start = 0 return type(self)(torch.arange(start=start, end=stop, step=step, device=self.raw.device)) def cumsum(self: TensorType, axis: Optional[int]=None) -> TensorType: if (axis is None): return type(self)(self.raw.reshape((- 1)).cumsum(dim=0)) return type(self)(self.raw.cumsum(dim=axis)) def flip(self: TensorType, axis: Optional[AxisAxes]=None) -> TensorType: if (axis is None): axis = tuple(range(self.ndim)) if (not isinstance(axis, Iterable)): axis = (axis,) return type(self)(self.raw.flip(dims=axis)) def meshgrid(self: TensorType, *tensors: TensorType, indexing: str='xy') -> Tuple[(TensorType, ...)]: tensors = unwrap_(*tensors) if ((indexing == 'ij') or (len(tensors) == 0)): outputs = torch.meshgrid(self.raw, *tensors) elif (indexing == 'xy'): outputs = torch.meshgrid(tensors[0], self.raw, *tensors[1:]) else: raise ValueError(f"Valid values for indexing are 'xy' and 'ij', got {indexing}") results = [type(self)(out) for out in outputs] if ((indexing == 'xy') and (len(results) >= 2)): (results[0], results[1]) = (results[1], results[0]) return tuple(results) def pad(self: TensorType, paddings: Tuple[(Tuple[(int, int)], ...)], mode: str='constant', value: float=0) -> TensorType: if (len(paddings) != self.ndim): raise ValueError('pad requires a tuple for each dimension') for p in paddings: if (len(p) != 2): raise ValueError('pad requires a tuple for each dimension') if (not ((mode == 'constant') or (mode == 'reflect'))): raise ValueError("pad requires mode 'constant' or 'reflect'") if (mode == 'reflect'): if ((self.ndim != 3) and (self.ndim != 4)): raise NotImplementedError k = (self.ndim - 2) if (paddings[:k] != (((0, 0),) * k)): raise NotImplementedError paddings = paddings[k:] paddings_ = list((x for p in reversed(paddings) for x in p)) return type(self)(torch.nn.functional.pad(self.raw, paddings_, mode=mode, value=value)) def isnan(self: TensorType) -> TensorType: return type(self)(torch.isnan(self.raw)) def isinf(self: TensorType) -> TensorType: return type(self)(torch.isinf(self.raw)) def crossentropy(self: TensorType, labels: TensorType) -> TensorType: if (self.ndim != 2): raise ValueError('crossentropy only supported for 2D logits tensors') if (self.shape[:1] != labels.shape): raise ValueError('labels must be 1D and must match the length of logits') return type(self)(torch.nn.functional.cross_entropy(self.raw, labels.raw, reduction='none')) def slogdet(self: TensorType) -> Tuple[(TensorType, TensorType)]: (sign, logabsdet) = torch.slogdet(self.raw) return (type(self)(sign), type(self)(logabsdet)) def _value_and_grad_fn(self: TensorType, f: Callable[(..., TensorType)]) -> Callable[(..., Tuple[(TensorType, TensorType)])]: ... def _value_and_grad_fn(self: TensorType, f: Callable[(..., TensorType)], has_aux: Literal[False]) -> Callable[(..., Tuple[(TensorType, TensorType)])]: ... def _value_and_grad_fn(self: TensorType, f: Callable[(..., Tuple[(TensorType, Any)])], has_aux: Literal[True]) -> Callable[(..., Tuple[(TensorType, Any, TensorType)])]: ... def _value_and_grad_fn(self: TensorType, f: Callable, has_aux: bool=False) -> Callable[(..., Tuple)]: def value_and_grad(x: TensorType, *args: Any, **kwargs: Any) -> Tuple: x = type(self)(x.raw.clone().requires_grad_()) if has_aux: (loss, aux) = f(x, *args, **kwargs) else: loss = f(x, *args, **kwargs) loss = loss.raw grad_raw = torch.autograd.grad(loss, x.raw)[0] grad = type(self)(grad_raw) assert (grad.shape == x.shape) loss = loss.detach() loss = type(self)(loss) if has_aux: if isinstance(aux, PyTorchTensor): aux = PyTorchTensor(aux.raw.detach()) elif isinstance(aux, tuple): aux = tuple(((PyTorchTensor(t.raw.detach()) if isinstance(t, PyTorchTensor) else t) for t in aux)) return (loss, aux, grad) else: return (loss, grad) return value_and_grad def sign(self: TensorType) -> TensorType: return type(self)(torch.sign(self.raw)) def sqrt(self: TensorType) -> TensorType: return type(self)(torch.sqrt(self.raw)) def inv(self: TensorType) -> TensorType: return type(self)(torch.linalg.inv(self.raw)) def round(self: TensorType) -> TensorType: return type(self)(torch.round(self.raw)) def ceil(self: TensorType) -> TensorType: return type(self)(torch.ceil(self.raw)) def floor(self: TensorType) -> TensorType: return type(self)(torch.floor(self.raw)) def float32(self: TensorType) -> TensorType: return self.astype(torch.float32) def float64(self: TensorType) -> TensorType: return self.astype(torch.float64) def where(self: TensorType, x: TensorOrScalar, y: TensorOrScalar) -> TensorType: if isinstance(x, Tensor): x_ = x.raw elif (isinstance(x, int) or isinstance(x, float)): if isinstance(y, Tensor): dtype = y.raw.dtype else: dtype = torch.float32 x_ = torch.full_like(self.raw, x, dtype=dtype) else: raise TypeError('expected x to be a Tensor, int or float') if isinstance(y, Tensor): y_ = y.raw elif (isinstance(y, int) or isinstance(y, float)): if isinstance(x, Tensor): dtype = x.raw.dtype else: dtype = torch.float32 y_ = torch.full_like(self.raw, y, dtype=dtype) return type(self)(torch.where(self.raw, x_, y_)) def __lt__(self: TensorType, other: TensorOrScalar) -> TensorType: return type(self)(self.raw.__lt__(unwrap1(other))) def __le__(self: TensorType, other: TensorOrScalar) -> TensorType: return type(self)(self.raw.__le__(unwrap1(other))) def __eq__(self: TensorType, other: TensorOrScalar) -> TensorType: return type(self)(self.raw.__eq__(unwrap1(other))) def __ne__(self: TensorType, other: TensorOrScalar) -> TensorType: return type(self)(self.raw.__ne__(unwrap1(other))) def __gt__(self: TensorType, other: TensorOrScalar) -> TensorType: return type(self)(self.raw.__gt__(unwrap1(other))) def __ge__(self: TensorType, other: TensorOrScalar) -> TensorType: return type(self)(self.raw.__ge__(unwrap1(other))) def __getitem__(self: TensorType, index: Any) -> TensorType: if isinstance(index, tuple): index = tuple(((x.raw if isinstance(x, Tensor) else x) for x in index)) elif isinstance(index, Tensor): index = index.raw return type(self)(self.raw[index]) def take_along_axis(self: TensorType, index: TensorType, axis: int) -> TensorType: if ((axis % self.ndim) != (self.ndim - 1)): raise NotImplementedError('take_along_axis is currently only supported for the last axis') return type(self)(torch.gather(self.raw, axis, index.raw)) def bool(self: TensorType) -> TensorType: return self.astype(torch.bool)
def init_bias_xavier(model, mode='fan_out', nonlinearity='relu', logger=None): layers_initialized = 0 a = 0 for m in model.modules(): if isinstance(m, nn.Conv2d): if (m.bias is not None): layers_initialized += 1 m.bias.data.normal_(0, (math.sqrt(2) / math.sqrt((1 + (9 * m.bias.data.shape[0]))))) logger.info((('Initialized ' + str(layers_initialized)) + ' bias conv2d layers using nn.init.xavier.noraml_'))
def build_recognizer(cfg, device): world_size = du.get_world_size() model = registry.RECOGNIZER[cfg.MODEL.RECOGNIZER.NAME](cfg) if (cfg.MODEL.NORM.SYNC_BN and (world_size > 1)): logger.info('start sync BN on the process group of {}'.format(du.LOCAL_RANK_GROUP)) convert_sync_bn(model, du.LOCAL_PROCESS_GROUP) if (cfg.MODEL.CONV.ADD_BLOCKS is not None): assert isinstance(cfg.MODEL.CONV.ADD_BLOCKS, tuple) logger.info(f'add blocks: {cfg.MODEL.CONV.ADD_BLOCKS}') for add_block in cfg.MODEL.CONV.ADD_BLOCKS: if (add_block == 'RepVGGBlock'): insert_repvgg_block(model) if (add_block == 'ACBlock'): insert_acblock(model) if (add_block == 'DiverseBranchBlock'): insert_dbblock(model) preloaded = cfg.MODEL.RECOGNIZER.PRELOADED if (preloaded != ''): logger.info(f'load preloaded: {preloaded}') check_pointer = CheckPointer(model) check_pointer.load(preloaded, map_location=device) logger.info('finish loading model weights') if du.is_master_proc(): logger.info(f'full model info:') logger.info(model) model = model.to(device=device) if (du.get_world_size() > 1): model = DDP(model, device_ids=[device], output_device=device, find_unused_parameters=False) return model
def train_cpu(data, label, num_class, list_hidden_nodes, initial_learning_rate, momentum, max_steps, decay_steps, decay_factor, batch_size, train_dir, moving_average_decay=0.9999, summary_steps=500, checkpoint_steps=10000, MLP_trainable=True, save_file='model.ckpt', load_file=None, random_seed=None): with tf.Graph().as_default(), tf.device('/cpu:0'): if (random_seed is not None): np.random.seed(random_seed) tf.set_random_seed(random_seed) global_step = tf.Variable(0, trainable=False) (data_holder, label_holder) = tf.train.shuffle_batch([tf.constant(data.T), tf.constant(label)], batch_size=batch_size, capacity=(20 * batch_size), min_after_dequeue=(10 * batch_size), enqueue_many=True) data_holder = tf.cast(data_holder, tf.float32) label_holder = tf.cast(label_holder, tf.float32) (logits, feats) = inference(data_holder, list_hidden_nodes, num_class, MLP_trainable=MLP_trainable) (loss, accuracy) = tcl_loss(logits, label_holder) (train_op, lr) = train(loss, accuracy, global_step=global_step, initial_learning_rate=initial_learning_rate, momentum=momentum, decay_steps=decay_steps, decay_factor=decay_factor, moving_average_decay=moving_average_decay) saver = tf.train.Saver(tf.global_variables()) summary_op = tf.summary.merge_all() init = tf.global_variables_initializer() config = tf.ConfigProto(log_device_placement=False) sess = tf.Session(config=config) sess.run(init) if (load_file is not None): print('Load trainable parameters from {0:s}...'.format(load_file)) reader = tf.train.NewCheckpointReader(load_file) reader_var_to_shape_map = reader.get_variable_to_shape_map() load_vars = tf.get_collection(FILTER_COLLECTION) initialized_vars = [] for lv in load_vars: if (lv.name.split(':')[0] in reader_var_to_shape_map): print(' {0:s}'.format(lv.name)) initialized_vars.append(lv) saver_init = tf.train.Saver(initialized_vars) saver_init.restore(sess, load_file) tf.train.start_queue_runners(sess=sess) summary_writer = tf.summary.FileWriter(train_dir, sess.graph) num_data = data.shape[1] num_steps_in_epoch = int(np.floor((num_data / batch_size))) for step in range(max_steps): (_, loss_value, accuracy_value, lr_value) = sess.run([train_op, loss, accuracy, lr]) assert (not np.isnan(loss_value)), 'Model diverged with loss = NaN' if ((step % summary_steps) == 0): summary_str = sess.run(summary_op) summary_writer.add_summary(summary_str, step) if ((step % checkpoint_steps) == 0): checkpoint_path = os.path.join(train_dir, save_file) saver.save(sess, checkpoint_path, global_step=step) save_path = os.path.join(train_dir, save_file) print('Save model in file: {0:s}'.format(save_path)) saver.save(sess, save_path)
def eye_like(A: torch.Tensor) -> torch.Tensor: return torch.eye(A.shape[(- 1)], dtype=A.dtype, device=A.device).expand_as(A)
_arg_scope def stack_blocks_dense(net, blocks, output_stride=None, store_non_strided_activations=False, outputs_collections=None): current_stride = 1 rate = 1 for block in blocks: with tf.variable_scope(block.scope, 'block', [net]) as sc: block_stride = 1 for (i, unit) in enumerate(block.args): if (store_non_strided_activations and (i == (len(block.args) - 1))): block_stride = unit.get('stride', 1) unit = dict(unit, stride=1) with tf.variable_scope(('unit_%d' % (i + 1)), values=[net]): if ((output_stride is not None) and (current_stride == output_stride)): net = block.unit_fn(net, rate=rate, **dict(unit, stride=1)) rate *= unit.get('stride', 1) else: net = block.unit_fn(net, rate=1, **unit) current_stride *= unit.get('stride', 1) if ((output_stride is not None) and (current_stride > output_stride)): raise ValueError('The target output_stride cannot be reached.') net = slim.utils.collect_named_outputs(outputs_collections, sc.name, net) if ((output_stride is not None) and (current_stride == output_stride)): rate *= block_stride else: net = subsample(net, block_stride) current_stride *= block_stride if ((output_stride is not None) and (current_stride > output_stride)): raise ValueError('The target output_stride cannot be reached.') if ((output_stride is not None) and (current_stride != output_stride)): raise ValueError('The target output_stride cannot be reached.') return net
class RandomForest(IterativeComponentWithSampleWeight, AutotabularClassificationAlgorithm): def __init__(self, criterion, max_features, max_depth, min_samples_split, min_samples_leaf, min_weight_fraction_leaf, bootstrap, max_leaf_nodes, min_impurity_decrease, random_state=None, n_jobs=1, class_weight=None): self.n_estimators = self.get_max_iter() self.criterion = criterion self.max_features = max_features self.max_depth = max_depth self.min_samples_split = min_samples_split self.min_samples_leaf = min_samples_leaf self.min_weight_fraction_leaf = min_weight_fraction_leaf self.bootstrap = bootstrap self.max_leaf_nodes = max_leaf_nodes self.min_impurity_decrease = min_impurity_decrease self.random_state = random_state self.n_jobs = n_jobs self.class_weight = class_weight self.estimator = None def get_max_iter(): return 512 def get_current_iter(self): return self.estimator.n_estimators def iterative_fit(self, X, y, sample_weight=None, n_iter=1, refit=False): from sklearn.ensemble import RandomForestClassifier if refit: self.estimator = None if (self.estimator is None): self.n_estimators = int(self.n_estimators) if check_none(self.max_depth): self.max_depth = None else: self.max_depth = int(self.max_depth) self.min_samples_split = int(self.min_samples_split) self.min_samples_leaf = int(self.min_samples_leaf) self.min_weight_fraction_leaf = float(self.min_weight_fraction_leaf) if (self.max_features not in ('sqrt', 'log2', 'auto')): max_features = int((X.shape[1] ** float(self.max_features))) else: max_features = self.max_features self.bootstrap = check_for_bool(self.bootstrap) if check_none(self.max_leaf_nodes): self.max_leaf_nodes = None else: self.max_leaf_nodes = int(self.max_leaf_nodes) self.min_impurity_decrease = float(self.min_impurity_decrease) self.estimator = RandomForestClassifier(n_estimators=n_iter, criterion=self.criterion, max_features=max_features, max_depth=self.max_depth, min_samples_split=self.min_samples_split, min_samples_leaf=self.min_samples_leaf, min_weight_fraction_leaf=self.min_weight_fraction_leaf, bootstrap=self.bootstrap, max_leaf_nodes=self.max_leaf_nodes, min_impurity_decrease=self.min_impurity_decrease, random_state=self.random_state, n_jobs=self.n_jobs, class_weight=self.class_weight, warm_start=True) else: self.estimator.n_estimators += n_iter self.estimator.n_estimators = min(self.estimator.n_estimators, self.n_estimators) self.estimator.fit(X, y, sample_weight=sample_weight) return self def configuration_fully_fitted(self): if (self.estimator is None): return False return (not (len(self.estimator.estimators_) < self.n_estimators)) def predict(self, X): if (self.estimator is None): raise NotImplementedError return self.estimator.predict(X) def predict_proba(self, X): if (self.estimator is None): raise NotImplementedError() probas = self.estimator.predict_proba(X) probas = convert_multioutput_multiclass_to_multilabel(probas) return probas def get_properties(dataset_properties=None): return {'shortname': 'RF', 'name': 'Random Forest Classifier', 'handles_regression': False, 'handles_classification': True, 'handles_multiclass': True, 'handles_multilabel': True, 'handles_multioutput': False, 'is_deterministic': True, 'input': (DENSE, SPARSE, UNSIGNED_DATA), 'output': (PREDICTIONS,)} def get_hyperparameter_search_space(dataset_properties=None): cs = ConfigurationSpace() criterion = CategoricalHyperparameter('criterion', ['gini', 'entropy'], default_value='gini') max_features = UniformFloatHyperparameter('max_features', 0.0, 1.0, default_value=0.5) max_depth = UnParametrizedHyperparameter('max_depth', 'None') min_samples_split = UniformIntegerHyperparameter('min_samples_split', 2, 20, default_value=2) min_samples_leaf = UniformIntegerHyperparameter('min_samples_leaf', 1, 20, default_value=1) min_weight_fraction_leaf = UnParametrizedHyperparameter('min_weight_fraction_leaf', 0.0) max_leaf_nodes = UnParametrizedHyperparameter('max_leaf_nodes', 'None') min_impurity_decrease = UnParametrizedHyperparameter('min_impurity_decrease', 0.0) bootstrap = CategoricalHyperparameter('bootstrap', ['True', 'False'], default_value='True') cs.add_hyperparameters([criterion, max_features, max_depth, min_samples_split, min_samples_leaf, min_weight_fraction_leaf, max_leaf_nodes, bootstrap, min_impurity_decrease]) return cs
def train(sess_config, input_hooks, model, data_init_op, steps, checkpoint_dir, tf_config=None, server=None): model.is_training = True hooks = [] hooks.extend(input_hooks) scaffold = tf.compat.v1.train.Scaffold(local_init_op=tf.group(tf.compat.v1.local_variables_initializer(), data_init_op), saver=tf.compat.v1.train.Saver(max_to_keep=args.keep_checkpoint_max)) stop_hook = tf.compat.v1.train.StopAtStepHook(last_step=steps) log_hook = tf.compat.v1.train.LoggingTensorHook({'steps': model.global_step, 'loss': model.loss}, every_n_iter=100) hooks.append(stop_hook) hooks.append(log_hook) if (args.timeline > 0): hooks.append(tf.compat.v1.train.ProfilerHook(save_steps=args.timeline, output_dir=checkpoint_dir)) save_steps = steps with tf.compat.v1.train.MonitoredTrainingSession(master=(server.target if server else ''), is_chief=(tf_config['is_chief'] if tf_config else True), hooks=hooks, scaffold=scaffold, checkpoint_dir=checkpoint_dir, save_checkpoint_steps=save_steps, summary_dir=checkpoint_dir, save_summaries_steps=args.save_steps, config=sess_config) as sess: while (not sess.should_stop()): sess.run([model.loss, model.train_op]) print('Training completed.')
def log_stats(stats, misc_args): if hasattr(misc_args, 'epoch'): lines = ('[%s][%s][Epoch %d][Iter %d / %d]\n' % (misc_args.run_name, misc_args.cfg_filename, misc_args.epoch, misc_args.step, misc_args.iters_per_epoch)) else: lines = ('[%s][%s][Step %d / %d]\n' % (misc_args.run_name, misc_args.cfg_filename, stats['iter'], cfg.SOLVER.MAX_ITER)) lines += ('\t\tloss: %.6f, lr: %.6f time: %.6f, eta: %s\n' % (stats['loss'], stats['lr'], stats['time'], stats['eta'])) if stats['head_losses']: lines += (('\t\t' + ', '.join((('%s: %.6f' % (k, v)) for (k, v) in stats['head_losses'].items()))) + '\n') print(lines[:(- 1)])
def fbeta(y_true, y_pred, beta=1): from keras import backend as K if (beta < 0): raise ValueError('The lowest choosable beta is zero (only precision).') if (K.sum(K.round(K.clip(y_true, 0, 1))) == 0): return 0 p = precision(y_true, y_pred) r = recall(y_true, y_pred) bb = (beta ** 2) fbeta_score = (((1 + bb) * (p * r)) / (((bb * p) + r) + K.epsilon())) return fbeta_score
def _get_fused_attention(feature1, feature2): upsample_module = nn.Upsample(size=(224, 224), mode='bilinear') feat_map1 = feature1.detach().clone() feat_map2 = feature2.detach().clone() return ((torch.sigmoid(upsample_module(feat_map1)) + torch.sigmoid(upsample_module(feat_map2))) / 2.0)
class MobileNetV2Block(nn.Module, ABC): def __init__(self, in_channels, out_channels, expansion_rate=1, repeat=1, stride=1, padding=1, conv_layer=None, norm_layer=None, act_layer=None): super(MobileNetV2Block, self).__init__() features = list() for i in range(repeat): if (i != 0): stride = 1 features.append(MobileNetV2InvertedResidual(in_channels, out_channels, expansion_rate=expansion_rate, stride=stride, padding=padding, conv_layer=conv_layer, norm_layer=norm_layer, act_layer=act_layer)) in_channels = out_channels self.conv = nn.Sequential(*features) def forward(self, x): return self.conv(x)
def drop_padding(seq: Sequence[Any], pad_id: Any): if (pad_id is None): return seq return list(reversed(list(dropwhile((lambda x: (x == pad_id)), reversed(seq)))))
def bidirectional_dynamic_rnn(cell_fw, cell_bw, inputs, sequence_length=None, initial_state_fw=None, initial_state_bw=None, dtype=None, parallel_iterations=None, swap_memory=False, time_major=False, scope=None): assert (not time_major) flat_inputs = flatten(inputs, 2) flat_len = (None if (sequence_length is None) else tf.cast(flatten(sequence_length, 0), 'int64')) ((flat_fw_outputs, flat_bw_outputs), final_state) = _bidirectional_dynamic_rnn(cell_fw, cell_bw, flat_inputs, sequence_length=flat_len, initial_state_fw=initial_state_fw, initial_state_bw=initial_state_bw, dtype=dtype, parallel_iterations=parallel_iterations, swap_memory=swap_memory, time_major=time_major, scope=scope) fw_outputs = reconstruct(flat_fw_outputs, inputs, 2) bw_outputs = reconstruct(flat_bw_outputs, inputs, 2) return ((fw_outputs, bw_outputs), final_state)
def get_split(config, split_name, dataset_dir, file_pattern=None, reader=None): all_file = [] reader = tf.TFRecordReader() batch_size = config.batch_size data_splitnum = config.data_split_num file_pattern = _FILE_PATTERN if (split_name == 'train'): num_epochs = None for i in range(data_splitnum): all_file.append(os.path.join(dataset_dir, 'sound20/', (file_pattern % (split_name, i)))) elif (split_name == 'test'): (num_epochs, batch_size) = (1, 1) all_file.append(os.path.join(dataset_dir, 'sound20/', (file_pattern % (split_name, 0)))) elif (split_name not in SPLITS_TO_SIZES): raise ValueError(('split name %s was not recognized.' % split_name)) filename_queue = tf.train.string_input_producer(all_file, num_epochs=num_epochs, shuffle=False) (_, serialized_example) = reader.read(filename_queue) features = tf.parse_single_example(serialized_example, features={'height': tf.FixedLenFeature([], tf.int64), 'width': tf.FixedLenFeature([], tf.int64), 'image_string': tf.FixedLenFeature([], tf.string), 'label': tf.FixedLenFeature([], tf.float32)}) image = tf.decode_raw(features['image_string'], tf.uint8) label = tf.cast(features['label'], tf.float32) height = tf.cast(features['height'], tf.int32) width = tf.cast(features['width'], tf.int32) image = tf.reshape(image, [height, width, 1]) resized_image = tf.image.resize_images(image, size=[IMAGE_HEIGHT, IMAGE_WIDTH]) min_after_dequeue = 10000 capacity = (min_after_dequeue + (3 * batch_size)) (images, labels) = tf.train.shuffle_batch([resized_image, label], batch_size=batch_size, capacity=capacity, num_threads=1, min_after_dequeue=min_after_dequeue, seed=config.random_seed) return (images, labels, SPLITS_TO_SIZES[split_name])
def menet160_8x1_g8(**kwargs): return get_menet(first_stage_channels=160, side_channels=8, groups=8, model_name='menet160_8x1_g8', **kwargs)
class RegexpTokenizer(Tokenizer): DIGIT = '\\p{Nd}+([:\\.\\,]\\p{Nd}+)*' TITLE = '(dr|esq|hon|jr|mr|mrs|ms|prof|rev|sr|st|rt|messrs|mmes|msgr)\\.(?=\\p{Z})' ABBRV = '([\\p{L}]\\.){2,}(?=\\p{Z}|$)' ALPHA_NUM = '[\\p{L}\\p{N}\\p{M}]++' HYPHEN = '{A}([-\\u058A\\u2010\\u2011]{A})+'.format(A=ALPHA_NUM) NEGATION = "((?!n't)[\\p{L}\\p{N}\\p{M}])++(?=n't)|n't" CONTRACTION1 = 'can(?=not\\b)' CONTRACTION2 = "'([tsdm]|re|ll|ve)\\b" START_DQUOTE = '(?<=[\\p{Z}\\(\\[{<]|^)(``|["\\u0093\\u201C\\u00AB])(?!\\p{Z})' START_SQUOTE = "(?<=[\\p{Z}\\(\\[{<]|^)[\\'\\u0091\\u2018\\u201B\\u2039](?!\\p{Z})" END_DQUOTE = '(?<!\\p{Z})(\\\'\\\'|["\\u0094\\u201D\\u00BB])' END_SQUOTE = "(?<!\\p{Z})[\\'\\u0092\\u2019\\u203A]" DASH = '--|[\\u0096\\u0097\\u2013\\u2014\\u2015]' ELLIPSES = '\\.\\.\\.|\\u2026' PUNCT = '\\p{P}' NON_WS = '[^\\p{Z}\\p{C}]' def __init__(self, **kwargs): self._regexp = regex.compile(('(?P<digit>%s)|(?P<title>%s)|(?P<abbr>%s)|(?P<neg>%s)|(?P<hyph>%s)|(?P<contr1>%s)|(?P<alphanum>%s)|(?P<contr2>%s)|(?P<sdquote>%s)|(?P<edquote>%s)|(?P<ssquote>%s)|(?P<esquote>%s)|(?P<dash>%s)|(?<ellipses>%s)|(?P<punct>%s)|(?P<nonws>%s)' % (self.DIGIT, self.TITLE, self.ABBRV, self.NEGATION, self.HYPHEN, self.CONTRACTION1, self.ALPHA_NUM, self.CONTRACTION2, self.START_DQUOTE, self.END_DQUOTE, self.START_SQUOTE, self.END_SQUOTE, self.DASH, self.ELLIPSES, self.PUNCT, self.NON_WS)), flags=((regex.IGNORECASE + regex.UNICODE) + regex.MULTILINE)) if (len(kwargs.get('annotators', {})) > 0): logger.warning(('%s only tokenizes! Skipping annotators: %s' % (type(self).__name__, kwargs.get('annotators')))) self.annotators = set() self.substitutions = kwargs.get('substitutions', True) def tokenize(self, text): data = [] matches = [m for m in self._regexp.finditer(text)] for i in range(len(matches)): token = matches[i].group() if self.substitutions: groups = matches[i].groupdict() if groups['sdquote']: token = '``' elif groups['edquote']: token = "''" elif groups['ssquote']: token = '`' elif groups['esquote']: token = "'" elif groups['dash']: token = '--' elif groups['ellipses']: token = '...' span = matches[i].span() start_ws = span[0] if ((i + 1) < len(matches)): end_ws = matches[(i + 1)].span()[0] else: end_ws = span[1] data.append((token, text[start_ws:end_ws], span)) return Tokens(data, self.annotators)
class iSLReLU(nn.Module): def __init__(self, slope=0.1): self.alpha = ((1 - slope) / (1 + slope)) super().__init__() def forward(self, x): self._last_x = x y = ((x + (self.alpha * (torch.sqrt((1 + (x * x))) - 1))) / (1 + self.alpha)) return y def inverse(self, y): a = self.alpha b = (((1 + a) * y) + a) x = ((torch.sqrt((((a ** 2) + ((a * b) ** 2)) - (a ** 4))) - b) / ((a ** 2) - 1)) return x def logdet(self): x = self._last_x a = self.alpha log_dets = torch.log(((1 + ((a * x) / torch.sqrt((1 + (x * x))))) / (1 + a))) if (len(x.shape) == 2): return log_dets.sum(1) else: return log_dets.sum(3).sum(2).sum(1) def reduce_func_singular_values(self, func): x = self._last_x a = self.alpha func_singular_vals = func(((1 + ((a * x) / torch.sqrt((1 + (x * x))))) / (1 + a))) if (len(x.shape) == 2): return func_singular_vals.sum(1) else: return func_singular_vals.sum(3).sum(2).sum(1)
def batch(dataset, batch_size: int, drop_last: bool=False): def iter_fn(): buffer = [] def _stack(xs): if isinstance(xs[0], dict): return {k: _stack([x[k] for x in xs]) for k in xs[0].keys()} if isinstance(xs[0], (str, bytes)): return list(xs) return torch.stack(xs, 0) for x in dataset(): buffer.append(x) if (len(buffer) >= batch_size): (yield _stack(buffer)) buffer = [] if (len(buffer) > 0): (yield _stack(buffer)) buffer = [] return iter_fn
def error_rate(predictions, labels): assert (len(predictions) == len(labels)) preds = np.argmax(predictions, 1) orig = np.argmax(labels, 1) error_rate = (100.0 - ((100.0 * np.sum((preds == orig))) / predictions.shape[0])) return (preds, orig, error_rate)
def load_langpair_dataset(data_path, split, src, src_dict, tgt, tgt_dict, combine, dataset_impl, upsample_primary, left_pad_source, left_pad_target, remove_eos_from_source, max_source_positions, max_target_positions, prepend_bos=False, load_alignments=False, truncate_source=False, append_source_id=False, num_buckets=0, shuffle=True, pad_to_multiple=1, prepend_bos_src=None): def split_exists(split, src, tgt, lang, data_path): filename = os.path.join(data_path, '{}.{}-{}.{}'.format(split, src, tgt, lang)) return indexed_dataset.dataset_exists(filename, impl=dataset_impl) src_datasets = [] tgt_datasets = [] for k in itertools.count(): split_k = (split + (str(k) if (k > 0) else '')) if split_exists(split_k, src, tgt, src, data_path): prefix = os.path.join(data_path, '{}.{}-{}.'.format(split_k, src, tgt)) elif split_exists(split_k, tgt, src, src, data_path): prefix = os.path.join(data_path, '{}.{}-{}.'.format(split_k, tgt, src)) elif (k > 0): break else: raise FileNotFoundError('Dataset not found: {} ({})'.format(split, data_path)) src_dataset = data_utils.load_indexed_dataset((prefix + src), src_dict, dataset_impl) if truncate_source: src_dataset = AppendTokenDataset(TruncateDataset(StripTokenDataset(src_dataset, src_dict.eos()), (max_source_positions - 1)), src_dict.eos()) src_datasets.append(src_dataset) tgt_dataset = data_utils.load_indexed_dataset((prefix + tgt), tgt_dict, dataset_impl) if (tgt_dataset is not None): tgt_datasets.append(tgt_dataset) logger.info('{} {} {}-{} {} examples'.format(data_path, split_k, src, tgt, len(src_datasets[(- 1)]))) if (not combine): break assert ((len(src_datasets) == len(tgt_datasets)) or (len(tgt_datasets) == 0)) if (len(src_datasets) == 1): src_dataset = src_datasets[0] tgt_dataset = (tgt_datasets[0] if (len(tgt_datasets) > 0) else None) else: sample_ratios = ([1] * len(src_datasets)) sample_ratios[0] = upsample_primary src_dataset = ConcatDataset(src_datasets, sample_ratios) if (len(tgt_datasets) > 0): tgt_dataset = ConcatDataset(tgt_datasets, sample_ratios) else: tgt_dataset = None if prepend_bos: assert (hasattr(src_dict, 'bos_index') and hasattr(tgt_dict, 'bos_index')) src_dataset = PrependTokenDataset(src_dataset, src_dict.bos()) if (tgt_dataset is not None): tgt_dataset = PrependTokenDataset(tgt_dataset, tgt_dict.bos()) elif (prepend_bos_src is not None): logger.info(f'prepending src bos: {prepend_bos_src}') src_dataset = PrependTokenDataset(src_dataset, prepend_bos_src) eos = None if append_source_id: src_dataset = AppendTokenDataset(src_dataset, src_dict.index('[{}]'.format(src))) if (tgt_dataset is not None): tgt_dataset = AppendTokenDataset(tgt_dataset, tgt_dict.index('[{}]'.format(tgt))) eos = tgt_dict.index('[{}]'.format(tgt)) align_dataset = None if load_alignments: align_path = os.path.join(data_path, '{}.align.{}-{}'.format(split, src, tgt)) if indexed_dataset.dataset_exists(align_path, impl=dataset_impl): align_dataset = data_utils.load_indexed_dataset(align_path, None, dataset_impl) tgt_dataset_sizes = (tgt_dataset.sizes if (tgt_dataset is not None) else None) return LanguagePairDataset(src_dataset, src_dataset.sizes, src_dict, tgt_dataset, tgt_dataset_sizes, tgt_dict, left_pad_source=left_pad_source, left_pad_target=left_pad_target, remove_eos_from_source=remove_eos_from_source, align_dataset=align_dataset, eos=eos, num_buckets=num_buckets, shuffle=shuffle, pad_to_multiple=pad_to_multiple)
def dla169(cfg, pretrained=None, **kwargs): Bottleneck.expansion = 2 model = DLA(cfg, [1, 1, 2, 3, 5, 1], [16, 32, 128, 256, 512, 1024], block=Bottleneck, residual_root=True, **kwargs) if (pretrained is not None): model.load_pretrained_model(pretrained, 'dla169') return model
def test_observation_decoder(shape=(3, 64, 64)): decoder = ObservationDecoder() batch_size = 2 (c, h, w) = shape embedding = torch.randn(batch_size, 1024) with torch.no_grad(): obs_dist: torch.distributions.Normal = decoder(embedding) obs_sample: torch.Tensor = obs_dist.sample() assert (obs_sample.size(0) == batch_size) assert (obs_sample.size(1) == c) assert (obs_sample.size(2) == h) assert (obs_sample.size(3) == w) horizon = 4 embedding = torch.randn(batch_size, horizon, 1024) with torch.no_grad(): obs_dist: torch.distributions.Normal = decoder(embedding) obs_sample: torch.Tensor = obs_dist.sample() assert (obs_sample.size(0) == batch_size) assert (obs_sample.size(1) == horizon) assert (obs_sample.size(2) == c) assert (obs_sample.size(3) == h) assert (obs_sample.size(4) == w) horizon = 4 embedding = torch.randn(batch_size, horizon, 1024) with torch.no_grad(): obs_dist: torch.distributions.Normal = decoder(embedding) obs_sample: torch.Tensor = obs_dist.sample() assert (obs_sample.size(0) == batch_size) assert (obs_sample.size(1) == horizon) assert (obs_sample.size(2) == c) assert (obs_sample.size(3) == h) assert (obs_sample.size(4) == w)
class ExpLrUpdaterHook(LrUpdaterHook): def __init__(self, gamma, **kwargs): self.gamma = gamma super(ExpLrUpdaterHook, self).__init__(**kwargs) def get_lr(self, runner, base_lr): progress = (trainer.epoch if self.by_epoch else trainer.iter) return (base_lr * (self.gamma ** progress))
class SydneyCaptions(RSICD): splits = ['train', 'val', 'test'] def __init__(self, root: str='.data/sydney_captions', split: str='train', transform: T.Compose=T.Compose([T.ToTensor()])): assert (split in self.splits) self.root = root self.transform = transform self.captions = self.load_captions(os.path.join(root, 'dataset.json'), split) self.image_root = 'images'
def evaluate_metrics(prediction_file: Union[(str, Path, List[Dict[(str, str)]])], reference_file: Union[(str, Path, List[Dict[(str, str)]])], nb_reference_captions: int=5) -> Dict[(str, Dict[(str, Union[(float, Dict[(str, float)])])])]: prediction_file = check_and_read_csv(prediction_file) reference_file = check_and_read_csv(reference_file) prediction_file.sort(key=(lambda row: row['file_name'])) reference_file.sort(key=(lambda row: row['file_name'])) reference_dict = {} for row in reference_file: reference_dict[row['file_name']] = row file_names = [row['file_name'] for row in prediction_file] assert all(((file_name in reference_dict) for file_name in file_names)) predictions = [] ground_truths = [] for row in prediction_file: file_name = row['file_name'] predictions.append(row['caption_predicted']) cap_names = ['caption_reference_{:02d}'.format(i) for i in range(1, (nb_reference_captions + 1))] ground_truths.append([reference_dict[file_name][cap] for cap in cap_names]) (metrics, per_file_metrics) = evaluate_metrics_from_lists(predictions, ground_truths) total_metrics = combine_single_and_per_file_metrics(metrics, per_file_metrics, file_names) return {key.lower(): value for (key, value) in total_metrics.items()}
class IICMeanTeacherTrainer(IICTrainer): def _init(self): super()._init() self._iic_weight = deepcopy(self._reg_weight) self._teacher_model = deepcopy(self._model) for param in self._teacher_model.parameters(): param.detach_() self._teacher_model.train() config = deepcopy(self._config['MeanTeacherParameters']) self._reg_criterion = {'mse': nn.MSELoss(), 'kl': KL_div()}[config['name']] self._ema_updater = ema_updater(alpha=float(config['alpha']), weight_decay=float(config['weight_decay'])) self._mt_weight = float(config['weight']) def epocher_class(self) -> Type[TrainEpocher]: return MIMeanTeacherEpocher def _run_epoch(self, epocher: MIMeanTeacherEpocher, *args, **kwargs) -> EpochResultDict: epocher.init(mi_estimator_array=self._mi_estimator_array, reg_criterion=self._reg_criterion, teacher_model=self._teacher_model, ema_updater=self._ema_updater, mt_weight=self._mt_weight, iic_weight=self._iic_weight, enforce_matching=self._enforce_matching) result = epocher.run() return result def _eval_epoch(self, *, loader, **kwargs) -> Tuple[(EpochResultDict, float)]: evaler = EvalEpocher(self._teacher_model, val_loader=loader, sup_criterion=self._sup_criterion, cur_epoch=self._cur_epoch, device=self._device) (result, cur_score) = evaler.run() return (result, cur_score)
class PreventStuckPlayer(ProxyPlayer): def __init__(self, player, nr_repeat, action): super(PreventStuckPlayer, self).__init__(player) self.act_que = deque(maxlen=nr_repeat) self.trigger_action = action def action(self, act): self.act_que.append(act) if (self.act_que.count(self.act_que[0]) == self.act_que.maxlen): act = self.trigger_action (r, isOver) = self.player.action(act) if isOver: self.act_que.clear() return (r, isOver) def restart_episode(self): super(PreventStuckPlayer, self).restart_episode() self.act_que.clear()
class Config(collections.MutableMapping): _instance = None _store: t.MutableMapping[(str, t.Any)] _file = Path('config.toml') _template = Path('config-example.toml') def get_instance(cls): if (cls._instance is None): cls() return cls._instance def __init__(self): if (Config._instance is not None): raise RuntimeError('This class is a singleton!') else: Config._instance = self with self._file.open() as f: self._store = toml.parse(f.read()) with self._template.open() as f: template = toml.parse(f.read()) if ('version' not in self): raise RuntimeError(f"The attribute 'version' is missing in '{self._file}'.") if ('version' not in template): raise RuntimeError(f"The attribute 'version' is missing in '{self._template}'.") self_version = version.parse(str(self['version'])) template_version = version.parse(str(template['version'])) if (self_version != template_version): raise RuntimeError(f"The version of '{self._file}' ({self_version}) is not equal to the version of '{self._template}' ({template_version}). Please update your config!") def __getitem__(self, key): return self._store[key] def __setitem__(self, key, value): self._store[key] = value def __delitem__(self, key): del self._store[key] def __iter__(self): return iter(self._store) def __len__(self): return len(self._store) def __repr__(self): return repr(self._store) def __str__(self): return str(self._store)
def repackage_hidden(h): if (h is None): return None if isinstance(h, list): return list((repackage_hidden(v) for v in h)) elif isinstance(h, tuple): return tuple((repackage_hidden(v) for v in h)) return h.detach()
def load_images(images, curriculum, device): return_images = [] head = 0 for stage in curriculum['stages']: stage_images = images[head:(head + stage['batch_size'])] stage_images = F.interpolate(stage_images, size=stage['img_size'], mode='bilinear', align_corners=True) return_images.append(stage_images) head += stage['batch_size'] return return_images
class KMeans(object): def __init__(self, num_centers, dtype=np.float32, algorithm='lloyd', initialization='plus_plus', distance='l2', max_iter=100, num_rep=1, verbosity=0): _check_integer(num_rep, 'num_rep', 1) _check_integer(verbosity, 'verbosity', 0) _check_integer(max_iter, 'max_iter', 0) algorithm = KMeansAlgorithm._members[algorithm.upper()] initialization = KMeansInitialization._members[initialization.upper()] distance = VectorComparisonType._members[('DISTANCE_' + distance.upper())] dtype = np.dtype(dtype) c_dtype = np_to_c_types.get(dtype, None) if (c_dtype not in [c_float, c_double]): raise TypeError('data should be float32 or float64') self.c_dtype = c_dtype self.vl_dtype = c_to_vl_types[c_dtype] self.num_centers = num_centers self.kmeans_p = vl_kmeans_new(self.vl_dtype, distance) self.kmeans = self.kmeans_p[0] self.kmeans.verbosity = verbosity self.kmeans.numRepetitions = num_rep self.kmeans.algorithm = algorithm self.kmeans.initialization = initialization self.kmeans.maxNumIterations = max_iter def _check_data(self, data): data = np.asarray(data) c_dtype = np_to_c_types.get(data.dtype, None) if (c_dtype != self.c_dtype): raise TypeError('different dtype') if (data.ndim != 2): raise TypeError('data must be num_data x dim') (num_data, dim) = data.shape if (dim == 0): raise ValueError('data dimension is zero') return data def fit(self, data): data = self._check_data(data) (num_data, dim) = data.shape data_p = data.ctypes.data_as(c_void_p) self.energy = vl_kmeans_cluster(self.kmeans_p, data_p, data.shape[1], data.shape[0], self.num_centers) centers_p = cast(self.kmeans.centers, POINTER(self.c_dtype)) self.centers = np.ctypeslib.as_array(centers_p, (self.num_centers, self.kmeans.dimension)).copy() return self def transform(self, data): data = self._check_data(data) assignments = np.empty(data.shape[0], dtype=np.uint32) data_p = data.ctypes.data_as(c_void_p) vl_kmeans_quantize(self.kmeans_p, assignments, None, data_p, data.shape[0]) return assignments def fit_transform(self, data): return self.fit(data).transform(data)
def numpyasarray(np_data): data = np_data assert data.flags['C_CONTIGUOUS'] arr = TVMArray() shape = c_array(tvm_shape_index_t, data.shape) arr.data = data.ctypes.data_as(ctypes.c_void_p) arr.shape = shape arr.strides = None arr.dtype = TVMType(np.dtype(data.dtype).name) arr.ndim = data.ndim arr.ctx = context(1, 0) return (arr, shape)
class Seq2SeqLMOutput(ModelOutput): loss: Optional[torch.FloatTensor] = None logits: torch.FloatTensor = None past_key_values: Optional[List[torch.FloatTensor]] = None decoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None decoder_attentions: Optional[Tuple[torch.FloatTensor]] = None encoder_last_hidden_state: Optional[torch.FloatTensor] = None encoder_hidden_states: Optional[Tuple[torch.FloatTensor]] = None encoder_attentions: Optional[Tuple[torch.FloatTensor]] = None
def horizontal_flow(sublayout_width, sublayout_height, num_row, pref_w_list, pref_h_list, optional_index_weight_dict, fixed_boundary): num = len(pref_w_list) result_index = [] row_width = [] row_height = [] i = 0 removed_index_weight_dict = {} for r in range(num_row): row_width.append([]) result_index.append([]) remaining_total_available_width = ((num_row - r) * sublayout_width) overall_delta = (remaining_total_available_width - sum(pref_w_list[i:])) loss = (abs(overall_delta) + sum(removed_index_weight_dict.values())) if (overall_delta <= 0): while ((overall_delta <= 0) and (optional_index_weight_dict != {})): flag = False for (index, weight) in optional_index_weight_dict.items(): if (index >= i): removed_index_weight_dict[index] = weight flag = True prev_overall_delta = overall_delta overall_delta += pref_w_list[index] break if (loss > (overall_delta + sum(removed_index_weight_dict.values()))): loss = (abs(overall_delta) + sum(removed_index_weight_dict.values())) del optional_index_weight_dict[index] else: if flag: del removed_index_weight_dict[index] overall_delta = prev_overall_delta break else: while ((overall_delta > 0) and (removed_index_weight_dict != {})): for (index, weight) in removed_index_weight_dict.items(): if (index >= i): optional_index_weight_dict[index] = weight overall_delta = (remaining_total_available_width - sum(pref_w_list)) break if (loss > (overall_delta + sum(removed_index_weight_dict.values()))): loss = (overall_delta + sum(removed_index_weight_dict.values())) del optional_index_weight_dict[index] else: del removed_index_weight_dict[index] break row_length = 0 start_index = i while (i < num): if (i not in removed_index_weight_dict.keys()): delta = (overall_delta / ((num - len(removed_index_weight_dict)) - start_index)) row_length += (pref_w_list[i] + delta) if (row_length <= sublayout_width): result_index[(- 1)].append(i) i += 1 elif ((row_length - sublayout_width) < (sublayout_width - ((row_length - pref_w_list[i]) - delta))): result_index[(- 1)].append(i) i += 1 else: break else: i += 1 end_index = i if (end_index == start_index): return None sum_row_width = 0 sum_row_height = 0 for index in result_index[(- 1)]: sum_row_width += pref_w_list[index] sum_row_height += pref_h_list[index] row_delta = ((sublayout_width - sum_row_width) / len(result_index[(- 1)])) row_height.append((sum_row_height / len(result_index[(- 1)]))) for index in result_index[(- 1)]: row_width[(- 1)].append((pref_w_list[index] + row_delta)) if ((sum(row_height) > sublayout_height) or fixed_boundary): delta = ((sum(row_height) - sublayout_height) / num_row) row_height = [(x - delta) for x in row_height] if (end_index != len(pref_w_list)): return None return (result_index, row_width, row_height, sum(removed_index_weight_dict.values()))
def test_new_scope_val_depends_on_old(): run_cell('\n class Foo:\n shared = 99\n ') run_cell('foo = Foo()') run_cell('foo.shared = 11') run_cell('foo_shared_alias = foo.shared') run_cell('Foo.shared = 12') run_cell('logging.info(foo_shared_alias)') assert_detected() run_cell('logging.info(foo.shared)') assert_detected() run_cell('logging.info(foo)') assert_detected()
def _cast_to_config(obj): if isinstance(obj, dict): return DictConfig(obj, flags={'allow_objects': True}) return obj
class Block(nn.Module): def __init__(self, dim, num_heads, mlp_ratio=4.0, qkv_bias=False, drop=0.0, drop_path=0.0, act_layer=nn.GELU, norm_layer=nn.LayerNorm, cross=False): super().__init__() self.norm1 = norm_layer(dim) self.attn = Attention(dim, num_heads=num_heads, qkv_bias=qkv_bias, cross=cross) self.drop_path = (DropPath(drop_path) if (drop_path > 0.0) else nn.Identity()) self.norm = norm_layer(dim) mlp_hidden_dim = int((dim * mlp_ratio)) self.mlp = Mlp(in_features=dim, hidden_features=mlp_hidden_dim, act_layer=act_layer, drop=drop) def forward(self, x, H, W): x = (x + self.drop_path(self.attn(self.norm1(x)))) x = (x + self.drop_path(self.mlp(self.norm(x), H, W))) return x
def apply_lora(base_model_path, lora_path): print(f'Loading the base model from {base_model_path}') base_tokenizer = AutoTokenizer.from_pretrained(base_model_path) base = AutoModelForCausalLM.from_pretrained(base_model_path, torch_dtype=torch.float16, low_cpu_mem_usage=True) print(f'Loading the LoRA adapter from {lora_path}') lora_model = PeftModel.from_pretrained(base, lora_path) print('Applying the LoRA') model = lora_model.merge_and_unload() return (base, model, base_tokenizer)
def check_equal(first, second, verbose): if verbose: print() for (i, (x, y)) in enumerate(zip(first, second)): x = x.cpu().detach().numpy() y = y.cpu().detach().numpy() if verbose: print('x = {}'.format(x.flatten())) print('y = {}'.format(y.flatten())) print(('-' * 80)) np.testing.assert_allclose(x, y, err_msg='Index: {}'.format(i))
def omniglot(): return itertools.chain(*[collect_download_configs((lambda : datasets.Omniglot(ROOT, background=background, download=True)), name=f"Omniglot, {('background' if background else 'evaluation')}") for background in (True, False)])
class TestPlotPDF(unittest.TestCase): def test_custom_bins(self): import numpy as np import powerlaw import matplotlib.pyplot as plt data = (1.0 / np.random.power(4.0, 1000)) fit = powerlaw.Fit(data) plt.figure() bins = 2 ax = fit.plot_pdf(marker='*', bins=bins) line = ax.lines[0] assert (len(line.get_xdata()) == bins) plt.close() plt.figure() bins = 10 ax = fit.plot_pdf(marker='*', bins=bins) line = ax.lines[0] assert (len(line.get_xdata()) == bins) plt.close()
class TestCrissCrossAttention(object): def test_cc_attention(self): device = torch.device(('cuda:0' if torch.cuda.is_available() else 'cpu')) from mmcv.ops import CrissCrossAttention loss_func = Loss() input = np.fromfile('tests/data/for_ccattention/ccattention_input.bin', dtype=np.float32) output = np.fromfile('tests/data/for_ccattention/ccattention_output.bin', dtype=np.float32) input = input.reshape((1, 32, 45, 45)) output = output.reshape((1, 32, 45, 45)) label = torch.ones((1, 32, 45, 45)) input = torch.FloatTensor(input) output = torch.FloatTensor(output) input.requires_grad = True shape = input.shape channel = shape[1] cca = CrissCrossAttention(channel) cca.to(device) input = input.to(device) label = label.to(device) cca.train() test_output = cca(input) test_loss = loss_func(test_output, label) test_loss.backward() test_output = test_output.detach().cpu().numpy() output = output.numpy() assert np.allclose(test_output, output) assert (test_output.shape == shape)
def get(config, mode): exec_config = copy.deepcopy(getattr(config, mode)) for att in list(config.keys()): if (att not in ['trainor', 'validator', 'ensemblor']): exec_config[att] = config[att] return exec_config
class CustomMetric(): def __init__(self, metric, metric_name, **kwargs): self.metric = metric self.metric_name = metric_name self.kwargs = kwargs self.scores = [] self.valid_classes = [] self.valid_matrices = [] self.names = [] self.score = None self.valid_class = None self.valid_matrix = None self.name = None self.last_scores = None self.last_valid_classes = None self.last_valid_matrices = None self.last_names = None def batch(self, prediction, target, name=None): self.score = self.metric(prediction, target, **self.kwargs).to('cpu') self.valid_class = target.unique().to('cpu') dummy = torch.zeros_like(self.score).to('cpu') dummy[self.valid_class] = 1 self.valid_matrix = dummy.type(torch.bool).to('cpu') self.scores.append(self.score) self.valid_classes.append(self.valid_class) self.valid_matrices.append(self.valid_matrix) if name: self.name = name self.names.append(self.name) def get_metrics_batch(self, mean=True): if mean: return self.score[self.valid_class].mean() else: return self.score[self.valid_class] def get_metrics_epoch(self, last=False, transpose=True): if last: if transpose: scores = torch.stack(self.last_scores).T masks = torch.stack(self.last_valid_matrices).T else: scores = torch.stack(self.last_scores) masks = torch.stack(self.last_valid_matrices) elif transpose: scores = torch.stack(self.scores).T masks = torch.stack(self.valid_matrices).T else: scores = torch.stack(self.scores) masks = torch.stack(self.valid_matrices) filtered = [s[m] for (s, m) in zip(scores, masks)] return torch.stack([c.mean() for c in filtered]) def epoch(self): self.last_scores = self.scores self.last_valid_classes = self.valid_classes self.last_valid_matrices = self.valid_matrices self.last_names = self.names result = self.get_metrics_epoch() self.reset() return result.mean() def reset(self): self.scores = [] self.valid_classes = [] self.valid_matrices = [] self.names = [] def __repr__(self): return self.metric_name
(argument('id', help='id of instance type to change bid', type=int), argument('--price', help='per machine bid price in $/hour', type=float), usage='vast.py change bid id [--price PRICE]', help='Change the bid price for a spot/interruptible instance', epilog=deindent('\n Change the current bid price of instance id to PRICE.\n If PRICE is not specified, then a winning bid price is used as the default.\n ')) def change__bid(args: argparse.Namespace): url = apiurl(args, '/instances/bid_price/{id}/'.format(id=args.id)) print(f'URL: {url}') r = requests.put(url, json={'client_id': 'me', 'price': args.price}) r.raise_for_status() print('Per gpu bid price changed'.format(r.json()))
def write_rttm(fn, turns): with open(fn, 'wb') as f: turns = sorted(turns, key=(lambda x: (x.fid, float(x.onset), float(x.dur)))) for turn in turns: line = ' '.join(turn) f.write(line.encode('utf-8')) f.write(b'\n')
class _DilatedResidualBlock(nn.Module): def __init__(self, in_channels: int, out_channels: int, kernel_size: int, dilation: int, causal: bool=True, norm: Literal[('batch', 'instance', None)]=None, activation: str='GELU', film_conditioning: bool=False, film_embedding_size: Optional[int]=None, film_batch_norm: bool=True, use_temporal_film: bool=False, temporal_film_block_size: Optional[int]=None): super().__init__() if (film_conditioning and ((film_embedding_size is None) or (not isinstance(film_embedding_size, int)) or (film_embedding_size < 1))): raise ValueError('FiLM conditioning requires a valid embedding size (int >= 1).') if (use_temporal_film and ((temporal_film_block_size is None) or (not isinstance(temporal_film_block_size, int)) or (temporal_film_block_size < 1))): raise ValueError('TFiLM conditioning requires a valid block size (int >= 1).') net = [] pre_activation_channels = ((out_channels * 2) if (activation == 'gated') else out_channels) if (norm is not None): if (norm not in ('batch', 'instance')): raise ValueError('Invalid norm type (must be batch or instance)') _Norm = (nn.BatchNorm1d if (norm == 'batch') else nn.InstanceNorm1d) net.append(_Norm(in_channels)) net.extend([Pad(kernel_size, dilation, causal=causal), nn.Conv1d(in_channels, pre_activation_channels, kernel_size, dilation=dilation, padding=0)]) self.net = nn.Sequential(*net) self.film = (FiLM(film_embedding_size, pre_activation_channels, film_batch_norm) if film_conditioning else None) self.activation = _get_activation(activation) self.tfilm = (TFiLM(out_channels, temporal_film_block_size) if use_temporal_film else None) self.residual = nn.Conv1d(in_channels, out_channels, 1) def forward(self, x: torch.Tensor, film_embedding: Optional[torch.Tensor]=None): activations = self.net(x) if (self.film is not None): activations = self.film(activations, film_embedding) y = self.activation(activations) if (self.tfilm is not None): y = self.tfilm(y) return (y + self.residual(x))
def prepare_trainer_collator(model_args, preprocessor: Dict[(str, Any)], collator_kwargs: Dict[(str, Any)]) -> Tuple[(Type[TrainerForMMLLM], Dict[(str, DataCollator)])]: type_ = model_args.type trainer_cls = TYPE2TRAINER[type_] data_collator_func = partial(Seq2Seq2DataCollatorWithImage, preprocessor=preprocessor, **collator_kwargs) data_collator_dict = {'train_collator': data_collator_func(inference_mode=False), 'eval_collator': data_collator_func(inference_mode=True)} return (trainer_cls, data_collator_dict)
def init_weights_normal(m): classname = m.__class__.__name__ if (classname == 'Conv2d'): nn.init.normal_(m.weight.data) nn.init.normal_(m.bias.data)
def accuracy(model, train_time_data, train_schedule_data, anomaly_data, class_data, model_plotter): (anomaly_correct, class_correct, class_total) = (0, 0, 0) (tpl, tnl, fpl, fnl) = ([], [], [], []) for (i, d) in enumerate(train_time_data): output = model(train_time_data[i], train_schedule_data[i]) (source_anomaly, source_prototype) = output (res, tp, tn, fp, fn) = anomaly_accuracy(source_anomaly, anomaly_data[i], model_plotter) anomaly_correct += res tpl.append(tp) tnl.append(tn) fpl.append(fp) fnl.append(fn) tp += res fp += res tn += (1 - res) fn += (1 - res) if (np.sum(anomaly_data[i]) > 0): class_total += 1 class_correct += class_accuracy(source_prototype, anomaly_data[i], class_data[i], model, model_plotter) (tp, fp, tn, fn) = (np.mean(tpl), np.mean(fpl), np.mean(tnl), np.mean(fn)) (p, r) = ((tp / (tp + fp)), (tp / (tp + fn))) tqdm.write(f'P = {p}, R = {r}, F1 = {(((2 * p) * r) / (p + r))}') return ((anomaly_correct / len(train_time_data)), (class_correct / class_total))
def get_phcfun_fromlib(): if ('linux' in sys.platform): libphcpack = (LOCATION + '/libPHCpack.so') phcpack = ctypes.CDLL(libphcpack) return phcpack._ada_use_c2phc if ('darwin' in sys.platform): libphcpack = (LOCATION + '/libPHCpack.dylib') phcpack = ctypes.CDLL(libphcpack) return phcpack._ada_use_c2phc if ('win' in sys.platform): libphcpack = (LOCATION + '/libPHCpack.dll') phcpack = ctypes.WinDLL(libphcpack, winmode=0) return phcpack._ada_use_c2phc print('The platform', sys.platform, 'is not supported.') return None
def ResNet34(conv_layer, linear_layer, init_type, **kwargs): assert (init_type == 'kaiming_normal'), 'only supporting default init for Resnets' return ResNet(conv_layer, linear_layer, BasicBlock, [3, 4, 6, 3], **kwargs)
def inverse_warp_3d(img, disp, padding_mode='zeros', disp_Y=None): device = disp.device (B, D, H, W) = disp.shape C = img.shape[1] if (disp_Y is not None): assert (disp.shape == disp_Y.shape), 'disparity map along x and y axis should have same shape!' if (img.dim() == 4): img = img.unsqueeze(2).expand(B, C, D, H, W) elif (img.dim() == 5): assert (D == img.shape[2]), 'The disparity number should be same between image and disparity map!' else: raise ValueError('image is only allowed with 4 or 5 dimensions, but got {} dimensions!'.format(img.dim())) grid_d = torch.linspace(0, (D - 1), D).view(1, D, 1, 1).expand(B, D, H, W).to(device) grid_h = torch.linspace(0, (H - 1), H).view(1, 1, H, 1).expand(B, D, H, W).to(device) grid_w = torch.linspace(0, (W - 1), W).view(1, 1, 1, W).expand(B, D, H, W).to(device) grid_w = (grid_w + disp) if (disp_Y is not None): grid_h = (grid_h + disp_Y) grid_d = (((grid_d / (D - 1)) * 2) - 1) grid_h = (((grid_h / (H - 1)) * 2) - 1) grid_w = (((grid_w / (W - 1)) * 2) - 1) grid_d = grid_d.unsqueeze(4) grid_h = grid_h.unsqueeze(4) grid_w = grid_w.unsqueeze(4) grid = torch.cat((grid_w, grid_h, grid_d), 4) projected_img = F.grid_sample(img, grid, padding_mode=padding_mode) return projected_img
def frameworkSrcBatch(args: argparse.Namespace, coreFunc: FunctionType) -> None: tasks = util.readAllTasksFromDir(args.input) lastApi: str = None for id in range(args.start, len(tasks)): task = tasks[id] (api, label, src) = util.parseTask(task) if args.singleapi: if ((lastApi != None) and (lastApi != api)): break lastApi = api try: if config.skipApi(api, label): raise Exception('Skipped', 'no detail') coreFunc(SEED, args, src) except Exception as e: reason: str = 'FrameworkCrashCatch' detail: str = str(e) if (len(e.args) >= 2): reason: str = e.args[0] detail: str = e.args[1] if (len(detail) > OUTPUT_LIMIT): detail = 'Detail is too long' if (reason == 'FrameworkCrashCatch'): print(detail) exit((- 1)) if ('Catch' in reason): with open('catches.log', 'a') as f: f.write('\nTitanFuzzTestcase {} {} {} {} {} {}'.format(id, api, label, reason, SEED, detail)) print('\nTitanFuzzTestcase', id, api, label, reason, SEED, detail)
def discriminator_loss(loss_func, real, fake): loss = [] real_loss = 0 fake_loss = 0 for i in range(2): if loss_func.__contains__('wgan'): real_loss = (- tf.reduce_mean(real[i])) fake_loss = tf.reduce_mean(fake[i]) if (loss_func == 'lsgan'): real_loss = tf.reduce_mean(tf.squared_difference(real[i], 1.0)) fake_loss = tf.reduce_mean(tf.square(fake[i])) if ((loss_func == 'gan') or (loss_func == 'dragan')): real_loss = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(labels=tf.ones_like(real[i]), logits=real[i])) fake_loss = tf.reduce_mean(tf.nn.sigmoid_cross_entropy_with_logits(labels=tf.zeros_like(fake[i]), logits=fake[i])) if (loss_func == 'hinge'): real_loss = tf.reduce_mean(relu((1.0 - real[i]))) fake_loss = tf.reduce_mean(relu((1.0 + fake[i]))) loss.append((real_loss + fake_loss)) return sum(loss)
class Net(torch.nn.Module): def __init__(self): super(Net, self).__init__() self.linear = torch.nn.Linear(30, 50) def forward(self, x): x = self.linear(x) return x
def parse_args(): parser = ArgumentParser(description='Training script: StyleGAN2 + ContraD with DataParallel.') parser.add_argument('gin_config', type=str, help='Path to the gin configuration file') parser.add_argument('architecture', type=str, help='Architecture') parser.add_argument('--mode', default='std', type=str, help='Training mode (default: std)') parser.add_argument('--penalty', default='none', type=str, help='Penalty (default: none)') parser.add_argument('--aug', default='none', type=str, help='Augmentation (default: hfrt)') parser.add_argument('--use_warmup', action='store_true', help='Use warmup strategy on LR') parser.add_argument('--workers', default=8, type=int, metavar='N', help='number of data loading workers (default: 0)') parser.add_argument('--temp', default=0.1, type=float, help='Temperature hyperparameter for contrastive losses') parser.add_argument('--lbd_a', default=1.0, type=float, help='Relative strength of the fake loss of ContraD') parser.add_argument('--no_lazy', action='store_true', help='Do not use lazy regularization') parser.add_argument('--d_reg_every', type=int, default=16, help='Interval of applying R1 when lazy regularization is used') parser.add_argument('--lbd_r1', type=float, default=10, help='R1 regularization') parser.add_argument('--style_mix', default=0.9, type=float, help='Style mixing regularization') parser.add_argument('--halflife_k', default=20, type=int, help='Half-life of exponential moving average in thousands of images') parser.add_argument('--ema_start_k', default=None, type=int, help='When to start the exponential moving average of G (default: halflife_k)') parser.add_argument('--halflife_lr', default=0, type=int, help='Apply LR decay when > 0') parser.add_argument('--no_fid', action='store_true', help='Do not track FIDs during training') parser.add_argument('--no_gif', action='store_true', help='Do not save GIF of sample generations from a fixed latent periodically during training') parser.add_argument('--n_eval_avg', default=3, type=int, help='How many times to average FID and IS') parser.add_argument('--print_every', help='', default=50, type=int) parser.add_argument('--evaluate_every', help='', default=2000, type=int) parser.add_argument('--save_every', help='', default=100000, type=int) parser.add_argument('--comment', help='Comment', default='', type=str) parser.add_argument('--resume', default=None, type=str, help='Path to logdir to resume the training') parser.add_argument('--finetune', default=None, type=str, help='Path to logdir that contains a pre-trained checkpoint of D') return parser.parse_args()
class TFXLMForMultipleChoice(metaclass=DummyObject): _backends = ['tf'] def __init__(self, *args, **kwargs): requires_backends(self, ['tf'])
def threshold_till_dag(B): if is_dag(B): return (B, 0) B = np.copy(B) nonzero_indices = np.where((B != 0)) weight_indices_ls = list(zip(B[nonzero_indices], nonzero_indices[0], nonzero_indices[1])) sorted_weight_indices_ls = sorted(weight_indices_ls, key=(lambda tup: abs(tup[0]))) for (weight, j, i) in sorted_weight_indices_ls: if is_dag(B): break B[(j, i)] = 0 dag_thres = abs(weight) return (B, dag_thres)
class LTRTrainer(BaseTrainer): def __init__(self, actor, loaders, optimizer, settings, lr_scheduler=None): super().__init__(actor, loaders, optimizer, settings, lr_scheduler) self._set_default_settings() self.stats = OrderedDict({loader.name: None for loader in self.loaders}) tensorboard_writer_dir = os.path.join('logs') self.tensorboard_writer = TensorboardWriter(tensorboard_writer_dir, [l.name for l in loaders]) def _set_default_settings(self): default = {'print_interval': 10, 'print_stats': None, 'description': ''} def cycle_dataset(self, loader): self.actor.train(loader.training) self._init_timing() for (i, data) in enumerate(loader, 1): data = data.to(self.device) data['epoch'] = self.epoch data['settings'] = self.settings (loss, stats) = self.actor(data) if loader.training: self.optimizer.zero_grad() loss.backward() self.optimizer.step() batch_size = data['train_images'].shape[loader.stack_dim] self._update_stats(stats, batch_size, loader) self._print_stats(i, loader, batch_size) def train_epoch(self): for loader in self.loaders: if ((self.epoch % loader.epoch_interval) == 0): self.cycle_dataset(loader) self._stats_new_epoch() self._write_tensorboard() def _init_timing(self): self.num_frames = 0 self.start_time = time.time() self.prev_time = self.start_time def _update_stats(self, new_stats: OrderedDict, batch_size, loader): if ((loader.name not in self.stats.keys()) or (self.stats[loader.name] is None)): self.stats[loader.name] = OrderedDict({name: AverageMeter() for name in new_stats.keys()}) for (name, val) in new_stats.items(): if (name not in self.stats[loader.name].keys()): self.stats[loader.name][name] = AverageMeter() self.stats[loader.name][name].update(val, batch_size) def _print_stats(self, i, loader, batch_size): self.num_frames += batch_size current_time = time.time() batch_fps = (batch_size / (current_time - self.prev_time)) average_fps = (self.num_frames / (current_time - self.start_time)) self.prev_time = current_time if (((i % 10) == 0) or (i == loader.__len__())): print_str = ('[%s: %d, %d / %d] ' % (loader.name, self.epoch, i, loader.__len__())) print_str += ('FPS: %.1f (%.1f) , ' % (average_fps, batch_fps)) for (name, val) in self.stats[loader.name].items(): if hasattr(val, 'avg'): print_str += ('%s: %.5f , ' % (name, val.avg)) print(print_str[:(- 5)]) def _stats_new_epoch(self): for loader in self.loaders: if loader.training: lr_list = self.lr_scheduler.get_lr() for (i, lr) in enumerate(lr_list): var_name = 'LearningRate/group{}'.format(i) if (var_name not in self.stats[loader.name].keys()): self.stats[loader.name][var_name] = StatValue() self.stats[loader.name][var_name].update(lr) for loader_stats in self.stats.values(): if (loader_stats is None): continue for stat_value in loader_stats.values(): if hasattr(stat_value, 'new_epoch'): stat_value.new_epoch() def _write_tensorboard(self): if (self.epoch == 1): self.tensorboard_writer.write_info('adafree_online', 'adafree', 'Train online for Adafree') self.tensorboard_writer.write_epoch(self.stats, self.epoch)
class ChannelSelector(object): def __init__(self, train_channel='random', eval_channel=0, axis=1): self.train_channel = train_channel self.eval_channel = eval_channel self.axis = axis def __repr__(self): return '{name}(train_channel={train_channel}, eval_channel={eval_channel}, axis={axis})'.format(name=self.__class__.__name__, train_channel=self.train_channel, eval_channel=self.eval_channel, axis=self.axis) def __call__(self, x, train=True): if (x.ndim <= self.axis): ind = tuple(((slice(None) if (i < x.ndim) else None) for i in range((self.axis + 1)))) x = x[ind] if train: channel = self.train_channel else: channel = self.eval_channel if (channel == 'random'): ch = numpy.random.randint(0, x.shape[self.axis]) else: ch = channel ind = tuple(((slice(None) if (i != self.axis) else ch) for i in range(x.ndim))) return x[ind]
def test_can_move_down(board: Board, another_board: Board) -> None: assert can_move_down(board) assert can_move_down(another_board) board = jnp.array([[0, 0, 0, 0], [1, 0, 0, 0], [2, 1, 0, 0], [3, 2, 1, 0]]) assert (~ can_move_down(board))
def store_model_weights(model, checkpoint_path, checkpoint_key='model', strict=True): checkpoint_path = os.path.abspath(checkpoint_path) output_dir = os.path.dirname(checkpoint_path) model = copy.deepcopy(model) checkpoint = torch.load(checkpoint_path, map_location='cpu') model.load_state_dict(checkpoint[checkpoint_key], strict=strict) tmp_path = os.path.join(output_dir, str(model.__hash__())) torch.save(model.state_dict(), tmp_path) sha256_hash = hashlib.sha256() with open(tmp_path, 'rb') as f: for byte_block in iter((lambda : f.read(4096)), b''): sha256_hash.update(byte_block) hh = sha256_hash.hexdigest() output_path = os.path.join(output_dir, (('weights-' + str(hh[:8])) + '.pth')) os.replace(tmp_path, output_path) return output_path
class FNetTokenizer(PreTrainedTokenizer): vocab_files_names = VOCAB_FILES_NAMES pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES model_input_names = ['input_ids', 'token_type_ids'] def __init__(self, vocab_file, do_lower_case=False, remove_space=True, keep_accents=True, unk_token='<unk>', sep_token='[SEP]', pad_token='<pad>', cls_token='[CLS]', mask_token='[MASK]', sp_model_kwargs: Optional[Dict[(str, Any)]]=None, **kwargs) -> None: mask_token = (AddedToken(mask_token, lstrip=True, rstrip=False, normalized=False) if isinstance(mask_token, str) else mask_token) self.sp_model_kwargs = ({} if (sp_model_kwargs is None) else sp_model_kwargs) super().__init__(do_lower_case=do_lower_case, remove_space=remove_space, keep_accents=keep_accents, unk_token=unk_token, sep_token=sep_token, pad_token=pad_token, cls_token=cls_token, mask_token=mask_token, sp_model_kwargs=self.sp_model_kwargs, **kwargs) 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_kwargs) self.sp_model.Load(vocab_file) def vocab_size(self): return len(self.sp_model) def get_vocab(self): vocab = {self.convert_ids_to_tokens(i): i for i in range(self.vocab_size)} vocab.update(self.added_tokens_encoder) return vocab def __getstate__(self): state = self.__dict__.copy() state['sp_model'] = None return state def __setstate__(self, d): self.__dict__ = d if (not hasattr(self, 'sp_model_kwargs')): self.sp_model_kwargs = {} self.sp_model = spm.SentencePieceProcessor(**self.sp_model_kwargs) 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 (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: str) -> List[str]: text = self.preprocess_text(text) pieces = self.sp_model.encode(text, out_type=str) new_pieces = [] for piece in pieces: if ((len(piece) > 1) and (piece[(- 1)] == str(',')) 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) return new_pieces def _convert_token_to_id(self, token): return self.sp_model.PieceToId(token) def _convert_id_to_token(self, index): return self.sp_model.IdToPiece(index) def convert_tokens_to_string(self, tokens): return self.sp_model.decode(tokens) def build_inputs_with_special_tokens(self, token_ids_0: List[int], token_ids_1: Optional[List[int]]=None) -> List[int]: sep = [self.sep_token_id] cls = [self.cls_token_id] if (token_ids_1 is None): return ((cls + token_ids_0) + sep) return ((((cls + token_ids_0) + sep) + token_ids_1) + sep) def get_special_tokens_mask(self, token_ids_0: List[int], token_ids_1: Optional[List[int]]=None, already_has_special_tokens: bool=False) -> List[int]: if already_has_special_tokens: return super().get_special_tokens_mask(token_ids_0=token_ids_0, token_ids_1=token_ids_1, already_has_special_tokens=True) if (token_ids_1 is not None): return (((([1] + ([0] * len(token_ids_0))) + [1]) + ([0] * len(token_ids_1))) + [1]) return (([1] + ([0] * len(token_ids_0))) + [1]) def create_token_type_ids_from_sequences(self, token_ids_0: List[int], token_ids_1: Optional[List[int]]=None) -> List[int]: sep = [self.sep_token_id] cls = [self.cls_token_id] if (token_ids_1 is None): return (len(((cls + token_ids_0) + sep)) * [0]) return ((len(((cls + token_ids_0) + sep)) * [0]) + (len((token_ids_1 + sep)) * [1])) def save_vocabulary(self, save_directory: str, filename_prefix: Optional[str]=None) -> Tuple[str]: if (not os.path.isdir(save_directory)): logger.error(f'Vocabulary path ({save_directory}) should be a directory') return out_vocab_file = os.path.join(save_directory, (((filename_prefix + '-') if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'])) if ((os.path.abspath(self.vocab_file) != os.path.abspath(out_vocab_file)) and os.path.isfile(self.vocab_file)): copyfile(self.vocab_file, out_vocab_file) elif (not os.path.isfile(self.vocab_file)): with open(out_vocab_file, 'wb') as fi: content_spiece_model = self.sp_model.serialized_model_proto() fi.write(content_spiece_model) return (out_vocab_file,)