code
stringlengths
101
5.91M
def parse_command_args(training_or_testing): if (training_or_testing == 'training'): parser = argparse.ArgumentParser(description='BoundingBox-less Location with PyTorch.', formatter_class=CustomFormatter) optional_args = parser._action_groups.pop() required_args = parser.add_argument_group('MANDATORY arguments') required_args.add_argument('--train-dir', required=True, help='Directory with training images. Must contain image files (any format), and a CSV or XML file containing groundtruth, as described in the README.') optional_args.add_argument('--val-dir', help="Directory with validation images and GT. If 'auto', 20%% of the training samples will be removed from training and used for validation. If left blank, no validation will be done.") optional_args.add_argument('--imgsize', type=str, default='256x256', metavar='HxW', help='Size of the input images (height x width).') optional_args.add_argument('--batch-size', type=strictly_positive_int, default=1, metavar='N', help='Input batch size for training.') optional_args.add_argument('--epochs', type=strictly_positive_int, default=np.inf, metavar='N', help='Number of epochs to train.') optional_args.add_argument('--nThreads', '-j', default=4, type=strictly_positive_int, metavar='N', help='Number of threads to create for data loading. Must be a striclty positive int') optional_args.add_argument('--lr', type=strictly_positive, default=4e-05, metavar='LR', help='Learning rate (step size).') optional_args.add_argument('-p', type=float, default=(- 1), metavar='P', help='alpha in the generalized mean (-inf => minimum)') optional_args.add_argument('--no-cuda', action='store_true', default=False, help='Disables CUDA training') optional_args.add_argument('--no-data-augm', action='store_true', default=False, help='Disables data augmentation (random vert+horiz flip)') optional_args.add_argument('--drop-last-batch', action='store_true', default=False, help='Drop the last batch during training, which may be incomplete. If the dataset size is not divisible by the batch size, then the last batch will be smaller.') optional_args.add_argument('--seed', type=int, default=1, metavar='S', help='Random seed.') optional_args.add_argument('--resume', default='', type=str, metavar='PATH', help='Path to latest checkpoint.') optional_args.add_argument('--save', default='', type=str, metavar='PATH', help='Where to save the model after each epoch.') optional_args.add_argument('--log-interval', type=strictly_positive, default=3, metavar='N', help='Time to wait between every time the losses are printed (in seconds).') optional_args.add_argument('--max-trainset-size', type=strictly_positive_int, default=np.inf, metavar='N', help='Only use the first N images of the training dataset.') optional_args.add_argument('--max-valset-size', type=strictly_positive_int, default=np.inf, metavar='N', help='Only use the first N images of the validation dataset.') optional_args.add_argument('--val-freq', default=1, type=int, metavar='F', help='Run validation every F epochs. If 0, no validation will be done. If no validation is done, a checkpoint will be saved every F epochs.') optional_args.add_argument('--visdom-env', default='default_environment', type=str, metavar='NAME', help='Name of the environment in Visdom.') optional_args.add_argument('--visdom-server', default=None, metavar='SRV', help='Hostname of the Visdom server. If not provided, nothing will be sent to Visdom.') optional_args.add_argument('--visdom-port', default=8989, metavar='PRT', help='Port of the Visdom server.') optional_args.add_argument('--optimizer', '--optim', default='sgd', type=str.lower, metavar='OPTIM', choices=['sgd', 'adam'], help='SGD or Adam.') optional_args.add_argument('--replace-optimizer', action='store_true', default=False, help='Replace optimizer state when resuming from checkpoint. If True, the optimizer will be replaced using the arguments of this scripts. If not resuming, it has no effect.') optional_args.add_argument('--max-mask-pts', type=strictly_positive_int, default=np.infty, metavar='M', help='Subsample this number of points from the mask, so that GMM fitting runs faster.') optional_args.add_argument('--paint', default=False, action='store_true', help='Paint red circles at the estimated locations in validation. This maskes it run much slower!') optional_args.add_argument('--radius', type=strictly_positive, default=5, metavar='R', help='Detections at dist <= R to a GT pointare considered True Positives.') optional_args.add_argument('--n-points', type=strictly_positive_int, default=None, metavar='N', help='If you know the number of points (e.g, just one pupil), then set it. Otherwise it will be estimated.') optional_args.add_argument('--ultrasmallnet', default=False, action='store_true', help='If True, the 5 central layers are removed,resulting in a much smaller UNet. This is used for example for the pupil dataset.Make sure to enable this if your are restoring a checkpoint that was trained using this option enabled.') optional_args.add_argument('--lambdaa', type=strictly_positive, default=1, metavar='L', help='Weight that will increase the importance of estimating the right number of points.') parser._action_groups.append(optional_args) args = parser.parse_args() args.eval_batch_size = 1 if (args.eval_batch_size != 1): raise NotImplementedError(('Only a batch size of 1 is implemented for now, got %s' % args.eval_batch_size)) if (args.save != ''): args.save = os.path.abspath(args.save) if (args.resume != ''): args.resume = os.path.abspath(args.resume) if ((args.save != '') and os.path.isfile(args.save) and (not (args.resume and (args.resume == args.save)))): print("E: Don't overwrite a checkpoint without resuming from it. Are you sure you want to do that? (if you do, remove it manually).") exit(1) args.cuda = ((not args.no_cuda) and torch.cuda.is_available()) elif (training_or_testing == 'testing'): parser = argparse.ArgumentParser(description='BoundingBox-less Location with PyTorch (inference/test only)', formatter_class=argparse.ArgumentDefaultsHelpFormatter) optional_args = parser._action_groups.pop() required_args = parser.add_argument_group('MANDATORY arguments') required_args.add_argument('--dataset', required=True, help='Directory with test images. Must contain image files (any format), and (optionally) a CSV or XML file containing groundtruth, as described in the README.') required_args.add_argument('--out', type=str, required=True, help='Directory where results will be stored (images+CSV).') optional_args.add_argument('--model', type=str, metavar='PATH', help='Checkpoint with the CNN model.\n') optional_args.add_argument('--evaluate', action='store_true', default=False, help='Evaluate metrics (Precision/Recall, RMSE, MAPE, etc.)') optional_args.add_argument('--no-cuda', '--no-gpu', action='store_true', default=False, help='Use CPU only, no GPU.') optional_args.add_argument('--imgsize', type=str, default='256x256', metavar='HxW', help='Size of the input images (heightxwidth).') optional_args.add_argument('--radii', type=str, default=range(0, (15 + 1)), metavar='Rs', help='Detections at dist <= R to a GT pt are True Positives.If not selected, R=0, ..., 15 will be tested.') optional_args.add_argument('--taus', type=str, default=(- 2), metavar='Ts', help='Detection threshold between 0 and 1. tau=-1 means dynamic Otsu thresholding. tau=-2 means Beta Mixture Model-based thresholding.') optional_args.add_argument('--n-points', type=int, default=None, metavar='N', help='If you know the exact number of points in the image, then set it. Otherwise it will be estimated by adding a L1 cost term.') optional_args.add_argument('--max-mask-pts', type=int, default=np.infty, metavar='M', help='Subsample this number of points from the mask, so GMM fitting runs faster.') optional_args.add_argument('--no-paint', default=False, action='store_true', help="Don't paint a red circle at each estimated location.") optional_args.add_argument('--force', '-f', default=False, action='store_true', help='Overwrite output files if they exist. In fact, it removes the output directory first') optional_args.add_argument('--seed', type=int, default=0, metavar='S', help='Random seed.') optional_args.add_argument('--max-testset-size', type=int, default=np.inf, metavar='N', help='Only use the first N images of the testing dataset.') optional_args.add_argument('--nThreads', '-j', default=4, type=int, metavar='N', help='Number of data loading threads.') optional_args.add_argument('--ultrasmallnet', default=False, action='store_true', help='If True, the 5 central layers are removed,resulting in a much smaller UNet. This is used for example for the pupil dataset.Make sure to enable this if your are restoring a checkpoint that was trained using this option enabled.') parser._action_groups.append(optional_args) args = parser.parse_args() if ((not args.no_cuda) and (not torch.cuda.is_available())): print('W: No GPU (CUDA) devices detected in your system, running with --no-gpu option...') args.cuda = ((not args.no_cuda) and torch.cuda.is_available()) args.paint = (not args.no_paint) if isinstance(args.taus, (list, range)): pass elif (isinstance(args.taus, str) and (',' in args.taus)): args.taus = [float(tau) for tau in args.taus.replace('[', '').replace(']', '').split(',')] else: args.taus = [float(args.taus)] if isinstance(args.radii, (list, range)): pass elif (isinstance(args.radii, str) and (',' in args.radii)): args.radii = [int(r) for r in args.radii.replace('[', '').replace(']', '').split(',')] else: args.radii = [int(args.radii)] else: raise ValueError(("Only 'training' or 'testing' allowed, got %s" % training_or_testing)) try: (args.height, args.width) = parse('{}x{}', args.imgsize) (args.height, args.width) = (int(args.height), int(args.width)) except TypeError as e: print("\\__ E: The input --imgsize must be in format WxH, got '{}'".format(args.imgsize)) exit((- 1)) return args
def insert_tagged_tokens(tokens, tags, template): to_insert = {} cur = (None, []) for (token, tag) in zip(tokens, tags): if (tag != cur[0]): if (cur[0] is not None): value = ' '.join(cur[1]) to_insert[cur[0]] = value if (tag == 'O'): cur = (None, []) else: cur = (tag, [token]) else: cur[1].append(token) if (cur[0] is not None): value = ' '.join(cur[1]) to_insert[cur[0]] = value modified = [] for token in template.split(): modified.append(to_insert.get(token, token)) return ' '.join(modified)
class Optimizable_optimizer(Optimizer): def __init__(self, optimizer: Optimizer, num_epochs: int): super(optimizer, self).__init__(num_epochs) self.optimizer = optimizer def optimize(self, temp_model: ModelWithTemperature, lr: float, nll_criterion, logits: torch.FloatTensor, labels: torch.FloatTensor, before_temperature_nll: float): best_loss = before_temperature_nll best_epoch = 0 for i in range(self._num_epochs): temp_model.zero_grad() loss = nll_criterion(temp_model.temperature_scale(logits), labels) loss.backward() self.optimizer.step() if (loss.item() < best_loss): best_loss = loss.item() best_epoch = 1 elif ((i - best_epoch) > 50): print('Stopped at {} with value {}'.format(best_epoch, best_loss)) break
_model('lstm_lm') class LSTMLanguageModel(FairseqLanguageModel): def __init__(self, decoder): super().__init__(decoder) def add_args(parser): parser.add_argument('--dropout', type=float, metavar='D', help='dropout probability') parser.add_argument('--decoder-embed-dim', type=int, metavar='N', help='decoder embedding dimension') parser.add_argument('--decoder-embed-path', type=str, metavar='STR', help='path to pre-trained decoder embedding') parser.add_argument('--decoder-hidden-size', type=int, metavar='N', help='decoder hidden size') parser.add_argument('--decoder-layers', type=int, metavar='N', help='number of decoder layers') parser.add_argument('--decoder-out-embed-dim', type=int, metavar='N', help='decoder output embedding dimension') parser.add_argument('--decoder-attention', type=str, metavar='BOOL', help='decoder attention') parser.add_argument('--adaptive-softmax-cutoff', metavar='EXPR', help='comma separated list of adaptive softmax cutoff points. Must be used with adaptive_loss criterion') parser.add_argument('--decoder-dropout-in', type=float, metavar='D', help='dropout probability for decoder input embedding') parser.add_argument('--decoder-dropout-out', type=float, metavar='D', help='dropout probability for decoder output') parser.add_argument('--share-decoder-input-output-embed', default=False, action='store_true', help='share decoder input and output embeddings') def build_model(cls, args, task): base_architecture(args) if (getattr(args, 'max_target_positions', None) is not None): max_target_positions = args.max_target_positions else: max_target_positions = getattr(args, 'tokens_per_sample', DEFAULT_MAX_TARGET_POSITIONS) def load_pretrained_embedding_from_file(embed_path, dictionary, embed_dim): num_embeddings = len(dictionary) padding_idx = dictionary.pad() embed_tokens = Embedding(num_embeddings, embed_dim, padding_idx) embed_dict = utils.parse_embedding(embed_path) utils.print_embed_overlap(embed_dict, dictionary) return utils.load_embedding(embed_dict, dictionary, embed_tokens) pretrained_decoder_embed = None if args.decoder_embed_path: pretrained_decoder_embed = load_pretrained_embedding_from_file(args.decoder_embed_path, task.target_dictionary, args.decoder_embed_dim) if args.share_decoder_input_output_embed: if (task.source_dictionary != task.target_dictionary): raise ValueError('--share-decoder-input-output-embeddings requires a joint dictionary') if (args.decoder_embed_dim != args.decoder_out_embed_dim): raise ValueError('--share-decoder-input-output-embeddings requires --decoder-embed-dim to match --decoder-out-embed-dim') decoder = LSTMDecoder(dictionary=task.dictionary, embed_dim=args.decoder_embed_dim, hidden_size=args.decoder_hidden_size, out_embed_dim=args.decoder_out_embed_dim, num_layers=args.decoder_layers, dropout_in=args.decoder_dropout_in, dropout_out=args.decoder_dropout_out, attention=options.eval_bool(args.decoder_attention), encoder_output_units=0, pretrained_embed=pretrained_decoder_embed, share_input_output_embed=args.share_decoder_input_output_embed, adaptive_softmax_cutoff=(options.eval_str_list(args.adaptive_softmax_cutoff, type=int) if (args.criterion == 'adaptive_loss') else None), max_target_positions=max_target_positions) return cls(decoder)
def _open_stream(stream, read_or_write): if hasattr(stream, read_or_write): return (False, stream) try: return (True, open(stream, (read_or_write[0] + 'b'))) except TypeError: raise RuntimeError('expected open file or filename')
def parse_args(): parser = argparse.ArgumentParser(description='Code generator for tensor contruction') parser.add_argument('-s', metavar='style', dest='style', type=str, default=None, choices=['numpy', 'mptensor'], help='set output style ("numpy" or "mptensor")') parser.add_argument('-v', '--verbose', action='store_true', dest='verbose', help='verbose mode') parser.add_argument('-o', metavar='outfile', dest='outfile', type=argparse.FileType('w'), default=sys.stdout, help='write the result to outfile') parser.add_argument('infile', type=argparse.FileType('r'), help='tensor-network definition file') return parser.parse_args()
def CheckArgs(args): if (not os.path.exists(args.config_dir)): os.makedirs(args.config_dir) if (args.feat_dir is not None): args.feat_dim = common_lib.get_feat_dim(args.feat_dir) if (args.ali_dir is not None): args.num_targets = common_lib.get_number_of_leaves_from_tree(args.ali_dir) elif (args.tree_dir is not None): args.num_targets = common_lib.get_number_of_leaves_from_tree(args.tree_dir) if (args.ivector_dir is not None): args.ivector_dim = common_lib.get_ivector_dim(args.ivector_dir) if (not (args.feat_dim > 0)): raise Exception('feat-dim has to be postive') if (not (args.num_targets > 0)): print(args.num_targets) raise Exception('num_targets has to be positive') if (not (args.ivector_dim >= 0)): raise Exception('ivector-dim has to be non-negative') if ((not (args.max_change_per_component >= 0)) or (not (args.max_change_per_component_final >= 0))): raise Exception('max-change-per-component and max_change-per-component-final should be non-negative') if (args.num_lstm_layers < 1): sys.exit('--num-lstm-layers has to be a positive integer') if ((args.clipping_threshold < 0) or (args.zeroing_threshold < 0)): sys.exit('--clipping-threshold and --zeroing-threshold have to be non-negative') if (not (args.zeroing_interval > 0)): raise Exception('--zeroing-interval has to be positive') if (args.lstm_delay is None): args.lstm_delay = ([[(- 1)]] * args.num_lstm_layers) else: try: args.lstm_delay = ParseLstmDelayString(args.lstm_delay.strip()) except ValueError: sys.exit("--lstm-delay has incorrect format value. Provided value is '{0}'".format(args.lstm_delay)) if (len(args.lstm_delay) != args.num_lstm_layers): sys.exit('--lstm-delay: Number of delays provided has to match --num-lstm-layers') return args
def preresnet101(**kwargs): return get_preresnet(blocks=101, model_name='preresnet101', **kwargs)
class T5Adapter(BaseAdapter): def match(self, model_path: str): return ('t5' in model_path) def load_model(self, model_path: str, from_pretrained_kwargs: dict): tokenizer = T5Tokenizer.from_pretrained(model_path, use_fast=False) model = AutoModelForSeq2SeqLM.from_pretrained(model_path, low_cpu_mem_usage=True, **from_pretrained_kwargs) return (model, tokenizer)
class Cell(object): def __init__(self, data=None, fmt=None, span=1, align=None): self.data = data if (fmt is None): fmt = CellFormat() if isinstance(fmt, str): fmt = CellFormat(fmt=fmt) self.fmt = fmt self.span = span self.align = align def __str__(self): return (self.fmt.fmt % self.data)
class MolGraph(): def __init__(self, smiles: str, args: Namespace): self.smiles = smiles self.n_atoms = 0 self.n_bonds = 0 self.f_atoms = [] self.f_bonds = [] self.a2b = [] self.b2a = [] self.b2revb = [] mol = Chem.MolFromSmiles(smiles) self.mol = mol self.n_atoms = mol.GetNumAtoms() for (i, atom) in enumerate(mol.GetAtoms()): self.f_atoms.append(atom_features(atom)) self.f_atoms = [self.f_atoms[i] for i in range(self.n_atoms)] for _ in range(self.n_atoms): self.a2b.append([]) for a1 in range(self.n_atoms): for a2 in range((a1 + 1), self.n_atoms): bond = mol.GetBondBetweenAtoms(a1, a2) if (bond is None): continue f_bond = bond_features(bond) if args.atom_messages: self.f_bonds.append(f_bond) self.f_bonds.append(f_bond) else: self.f_bonds.append((self.f_atoms[a1] + f_bond)) self.f_bonds.append((self.f_atoms[a2] + f_bond)) b1 = self.n_bonds b2 = (b1 + 1) self.a2b[a2].append(b1) self.b2a.append(a1) self.a2b[a1].append(b2) self.b2a.append(a2) self.b2revb.append(b2) self.b2revb.append(b1) self.n_bonds += 2 def ggnn_features(self, max_atoms=(- 1), out_size=(- 1)): mol = self.mol type_check_num_atoms(mol, max_atoms) atom_array = construct_atomic_number_array(mol, out_size=out_size) adj_array = construct_discrete_edge_matrix(mol, out_size=out_size) return (atom_array, adj_array)
class OptimizationMethod(object): def __init__(self, name, group, supported_devices=['cpu', 'cuda'], min_sm_version=None, opt_computation=None, opt_memory=None, opt_communication=None, distributed_only=False, process_mode='ONE_PROCESS', is_tunable=True): self.name = name self.group = group self.supported_devices = supported_devices self.min_sm_version = min_sm_version self.opt_computation = opt_computation self.opt_memory = opt_memory self.opt_communication = opt_communication self.distributed_only = distributed_only self.process_mode = process_mode self.is_tunable = is_tunable self.disabled = False
def write_predictions(all_examples, all_features, all_results, n_best_size, max_answer_length, do_lower_case, output_prediction_file, output_nbest_file, output_null_log_odds_file, verbose_logging, version_2_with_negative, null_score_diff_threshold): logger.info(('Writing predictions to: %s' % output_prediction_file)) logger.info(('Writing nbest to: %s' % output_nbest_file)) example_index_to_features = collections.defaultdict(list) for feature in all_features: example_index_to_features[feature.example_index].append(feature) unique_id_to_result = {} for result in all_results: unique_id_to_result[result.unique_id] = result _PrelimPrediction = collections.namedtuple('PrelimPrediction', ['feature_index', 'start_index', 'end_index', 'start_logit', 'end_logit']) all_predictions = collections.OrderedDict() all_nbest_json = collections.OrderedDict() scores_diff_json = collections.OrderedDict() for (example_index, example) in enumerate(all_examples): features = example_index_to_features[example_index] prelim_predictions = [] score_null = 1000000 min_null_feature_index = 0 null_start_logit = 0 null_end_logit = 0 for (feature_index, feature) in enumerate(features): result = unique_id_to_result[feature.unique_id] start_indexes = _get_best_indexes(result.start_logits, n_best_size) end_indexes = _get_best_indexes(result.end_logits, n_best_size) if version_2_with_negative: feature_null_score = (result.start_logits[0] + result.end_logits[0]) if (feature_null_score < score_null): score_null = feature_null_score min_null_feature_index = feature_index null_start_logit = result.start_logits[0] null_end_logit = result.end_logits[0] for start_index in start_indexes: for end_index in end_indexes: if (start_index >= len(feature.tokens)): continue if (end_index >= len(feature.tokens)): continue if (start_index not in feature.token_to_orig_map): continue if (end_index not in feature.token_to_orig_map): continue if (not feature.token_is_max_context.get(start_index, False)): continue if (end_index < start_index): continue length = ((end_index - start_index) + 1) if (length > max_answer_length): continue prelim_predictions.append(_PrelimPrediction(feature_index=feature_index, start_index=start_index, end_index=end_index, start_logit=result.start_logits[start_index], end_logit=result.end_logits[end_index])) if version_2_with_negative: prelim_predictions.append(_PrelimPrediction(feature_index=min_null_feature_index, start_index=0, end_index=0, start_logit=null_start_logit, end_logit=null_end_logit)) prelim_predictions = sorted(prelim_predictions, key=(lambda x: (x.start_logit + x.end_logit)), reverse=True) _NbestPrediction = collections.namedtuple('NbestPrediction', ['text', 'start_logit', 'end_logit']) seen_predictions = {} nbest = [] for pred in prelim_predictions: if (len(nbest) >= n_best_size): break feature = features[pred.feature_index] if (pred.start_index > 0): tok_tokens = feature.tokens[pred.start_index:(pred.end_index + 1)] orig_doc_start = feature.token_to_orig_map[pred.start_index] orig_doc_end = feature.token_to_orig_map[pred.end_index] orig_tokens = example.doc_tokens[orig_doc_start:(orig_doc_end + 1)] tok_text = ' '.join(tok_tokens) tok_text = tok_text.replace(' ##', '') tok_text = tok_text.replace('##', '') tok_text = tok_text.strip() tok_text = ' '.join(tok_text.split()) orig_text = ' '.join(orig_tokens) final_text = get_final_text(tok_text, orig_text, do_lower_case, verbose_logging) if (final_text in seen_predictions): continue seen_predictions[final_text] = True else: final_text = '' seen_predictions[final_text] = True nbest.append(_NbestPrediction(text=final_text, start_logit=pred.start_logit, end_logit=pred.end_logit)) if version_2_with_negative: if ('' not in seen_predictions): nbest.append(_NbestPrediction(text='', start_logit=null_start_logit, end_logit=null_end_logit)) if (len(nbest) == 1): nbest.insert(0, _NbestPrediction(text='empty', start_logit=0.0, end_logit=0.0)) if (not nbest): nbest.append(_NbestPrediction(text='empty', start_logit=0.0, end_logit=0.0)) assert (len(nbest) >= 1) total_scores = [] best_non_null_entry = None for entry in nbest: total_scores.append((entry.start_logit + entry.end_logit)) if (not best_non_null_entry): if entry.text: best_non_null_entry = entry probs = _compute_softmax(total_scores) nbest_json = [] for (i, entry) in enumerate(nbest): output = collections.OrderedDict() output['text'] = entry.text output['probability'] = probs[i] output['start_logit'] = entry.start_logit output['end_logit'] = entry.end_logit nbest_json.append(output) assert (len(nbest_json) >= 1) if (not version_2_with_negative): all_predictions[example.qas_id] = nbest_json[0]['text'] else: score_diff = ((score_null - best_non_null_entry.start_logit) - best_non_null_entry.end_logit) scores_diff_json[example.qas_id] = score_diff if (score_diff > null_score_diff_threshold): all_predictions[example.qas_id] = '' else: all_predictions[example.qas_id] = best_non_null_entry.text all_nbest_json[example.qas_id] = nbest_json with open(output_prediction_file, 'w') as writer: writer.write((json.dumps(all_predictions, indent=4) + '\n')) with open(output_nbest_file, 'w') as writer: writer.write((json.dumps(all_nbest_json, indent=4) + '\n')) if version_2_with_negative: with open(output_null_log_odds_file, 'w') as writer: writer.write((json.dumps(scores_diff_json, indent=4) + '\n'))
def save_pickle(filepath, x): with open(filepath, 'wb') as handle: pickle.dump(x, handle, protocol=pickle.HIGHEST_PROTOCOL)
class SigmoidNode(Node): def __init__(self, prev_node): super().__init__(prev_node) self.in_var = prev_node.out_var self.in_dim = prev_node.out_dim self.out_dim = self.in_dim self.out_var = Allocation.allocate_var('float', 'x', self.out_dim) def lowering(self): (loops, idxs) = LoopNode.create_loops(self.in_var.dim) in_var_idx = IndexedVariable(self.in_var) out_var_idx = IndexedVariable(self.out_var) in_var_idx.set_indices(idxs) out_var_idx.set_indices(idxs) expression = Expression('1.f / (1.f + expf(-{t_var_idx}))', t_var_idx=in_var_idx) node = AssignmentNode(out_var_idx, expression) loops[(- 1)].add_edge('content', node) self.var_decls.append(self.out_var) self.math_required = True self.add_edge('content', loops[0])
def fuse_bn(conv, bn): kernel = conv.weight running_mean = bn.running_mean running_var = bn.running_var gamma = bn.weight beta = bn.bias eps = bn.eps std = (running_var + eps).sqrt() t = (gamma / std).reshape((- 1), 1, 1, 1) return ((kernel * t), (beta - ((running_mean * gamma) / std)))
def main(): reader = csv.reader(sys.stdin) writer = csv.writer(sys.stdout) header = True for row in reader: fixed_spans = row[0] if (not header): fixed_spans = _fix_spans(ast.literal_eval(row[0]), row[1]) writer.writerow([fixed_spans, row[1]]) header = False
class DistilBertModel(metaclass=DummyObject): _backends = ['torch'] def __init__(self, *args, **kwargs): requires_backends(self, ['torch'])
class EndToEndModel(nn.Module): def __init__(self, segm_model, pose_model, object_names, object_ids): super(EndToEndModel, self).__init__() self.segm_model = segm_model self.resize = nn.AdaptiveMaxPool2d((240, 320)) self.pose_model = pose_model self.object_names = object_names self.object_ids = object_ids def forward(self, x): assert (x.size(0) == 1) x = self.segm_model(x) (_, x) = x.max(1, keepdim=True) object_names = [] positions = [] orientations = [] for (i, object_name) in enumerate(self.object_names): mask = self.resize(x.eq(self.object_ids[i]).float()) if (mask.sum().item() < 20): continue object_index = torch.LongTensor([i]) (position, orientation) = self.pose_model(mask, object_index) object_names.append(object_name) positions.append(position[0].cpu().numpy()) orientations.append(orientation[0].cpu().numpy()) return (x[0].cpu().numpy().squeeze(0), object_names, positions, orientations)
class myMerlinFlow(FlowSpec): MODEL_FOLDER = Parameter(name='model_folder', help='Folder to store the model from Merlin, between steps', default='merlin_model') ROW_SAMPLING = Parameter(name='row_sampling', help='Snowflake row sampling: if 0, NO sampling is applied', default='1') TRAINING_END_DATE = Parameter(name='training_end_date', help='Data up until this date is used for training, format yyyy-mm-dd', default='2020-09-08') VALIDATION_END_DATE = Parameter(name='validation_end_date', help='Data up after training end and until this date is used for validation, format yyyy-mm-dd', default='2020-09-15') COMET_PROJECT_NAME = Parameter(name='comet_project_name', help='Name of the project in our Comet dashboard', default='two_tower_h_and_m_merlin') VALIDATION_METRIC = Parameter(name='validation_metric', help='Merlin metric to use for picking the best set of hyperparameter', default='recall_at_10') N_EPOCHS = Parameter(name='n_epoch', help='Number of epochs to train the Merlin model', default='1') DYNAMO_TABLE = Parameter(name='dynamo_table', help='Name of dynamo db table to store the pre-computed recs. Default is same as in the serverless application', default='userItemTable') TOP_K = Parameter(name='top_k', help='Number of products to recommend for a giver shopper', default='10') def start(self): print(('flow name: %s' % current.flow_name)) print(('run id: %s' % current.run_id)) print(('username: %s' % current.username)) if (os.environ.get('EN_BATCH', '0') == '1'): print('ATTENTION: AWS BATCH ENABLED!') from metaflow.metaflow_config import DATASTORE_SYSROOT_S3 print(('DATASTORE_SYSROOT_S3: %s' % DATASTORE_SYSROOT_S3)) if (DATASTORE_SYSROOT_S3 is None): print('ATTENTION: LOCAL DATASTORE ENABLED') assert (os.environ['COMET_API_KEY'] and self.COMET_PROJECT_NAME) assert int(self.ROW_SAMPLING) from snowflake_client import SnowflakeClient sf_client = SnowflakeClient(os.environ['SF_USER'], os.environ['SF_PWD'], os.environ['SF_ACCOUNT'], os.environ['SF_ROLE'], os.environ['SF_WAREHOUSE']) snowflake_version = sf_client.get_version() print(snowflake_version) assert (snowflake_version is not None) self.training_end_date = datetime.strptime(self.TRAINING_END_DATE, '%Y-%m-%d') self.validation_end_date = datetime.strptime(self.VALIDATION_END_DATE, '%Y-%m-%d') assert (self.validation_end_date > self.training_end_date) self.next(self.get_dataset) def get_dataset(self): from snowflake_client import SnowflakeClient from pyarrow import Table as pt sf_client = SnowflakeClient(os.environ['SF_USER'], os.environ['SF_PWD'], os.environ['SF_ACCOUNT'], os.environ['SF_ROLE'], os.environ['SF_WAREHOUSE']) snowflake_sampling = int(self.ROW_SAMPLING) sampling_expression = ('' if (snowflake_sampling == 0) else 'sample({})'.format(snowflake_sampling)) query = '\n SELECT \n ARTICLE_ID,\n PRODUCT_CODE, \n PRODUCT_TYPE_NO,\n PRODUCT_GROUP_NAME,\n GRAPHICAL_APPEARANCE_NO,\n COLOUR_GROUP_CODE,\n PERCEIVED_COLOUR_VALUE_ID,\n PERCEIVED_COLOUR_MASTER_ID,\n DEPARTMENT_NO,\n INDEX_CODE,\n INDEX_GROUP_NO,\n SECTION_NO,\n GARMENT_GROUP_NO,\n ACTIVE,\n FN,\n AGE,\n CLUB_MEMBER_STATUS,\n CUSTOMER_ID,\n FASHION_NEWS_FREQUENCY,\n POSTAL_CODE,\n PRICE,\n SALES_CHANNEL_ID,\n T_DAT,\n S3_URL\n FROM\n "EXPLORATION_DB"."HM_POST"."FILTERED_DATAFRAME"\n {}\n ORDER BY\n T_DAT ASC\n '.format(sampling_expression) print('Fetching rows with query: \n {} \n\nIt may take a while...\n'.format(query)) dataset = sf_client.fetch_all(query, debug=True) assert dataset dataset = [{k.lower(): v for (k, v) in row.items()} for row in dataset] self.item_id_2_meta = {str(r['article_id']): r for r in dataset} print('Example articles: {}'.format(list(self.item_id_2_meta.keys())[:3])) train_dataset = pt.from_pylist([row for row in dataset if (row['t_dat'] < self.training_end_date)]) validation_dataset = pt.from_pylist([row for row in dataset if ((row['t_dat'] >= self.training_end_date) and (row['t_dat'] < self.validation_end_date))]) test_dataset = pt.from_pylist([row for row in dataset if (row['t_dat'] >= self.validation_end_date)]) print('# {:,} events in the training set, {:,} for validation, {:,} for test'.format(len(train_dataset), len(validation_dataset), len(test_dataset))) self.label_to_dataset = {'train': train_dataset, 'valid': validation_dataset, 'test': test_dataset} self.next(self.build_workflow) def build_workflow(self): from workflow_builder import get_nvt_workflow, read_to_dataframe import pandas as pd import itertools import nvtabular as nvt label_to_df = {} for (label, dataset) in self.label_to_dataset.items(): label_to_df[label] = read_to_dataframe(dataset, label) full_dataset = nvt.Dataset(pd.concat(list(label_to_df.values()))) workflow = get_nvt_workflow() workflow.fit(full_dataset) self.label_to_melin_dataset = {} for (label, _df) in label_to_df.items(): cnt_dataset = nvt.Dataset(_df) self.label_to_melin_dataset[label] = cnt_dataset workflow.transform(cnt_dataset).to_parquet(output_path='merlin/{}/'.format(label)) user_unique_ids = list(pd.read_parquet('categories/unique.customer_id.parquet')['customer_id']) items_unique_ids = list(pd.read_parquet('categories/unique.article_id.parquet')['article_id']) self.id_2_user_id = {idx: _ for (idx, _) in enumerate(user_unique_ids)} self.id_2_item_id = {idx: _ for (idx, _) in enumerate(items_unique_ids)} batch_sizes = [16384, 4096] learning_rates = [0.04, 0.02] grid_search = [] for params in itertools.product(batch_sizes, learning_rates): grid_search.append({'BATCH_SIZE': params[0], 'LEARNING_RATE': params[1]}) self.hypers_sets = [json.dumps(_) for _ in grid_search][:1] print(self.hypers_sets) self.next(self.train_model, foreach='hypers_sets') (vars={'EN_BATCH': os.getenv('EN_BATCH'), 'COMET_API_KEY': os.getenv('COMET_API_KEY')}) _decorator(batch(memory=24000, image='public.ecr.aws/outerbounds/merlin-reasonable-scale:22.11-latest'), flag=os.getenv('EN_BATCH')) (libraries={'requests': '2.28.1', 'comet-ml': '3.26.0'}) def train_model(self): import hashlib from comet_ml import Experiment import merlin.models.tf as mm from merlin.io.dataset import Dataset from merlin.schema.tags import Tags import tensorflow as tf self.hyper_string = self.input self.hypers = json.loads(self.hyper_string) train = Dataset('merlin/train/*.parquet') valid = Dataset('merlin/valid/*.parquet') print('Train dataset shape: {}, Validation: {}'.format(train.to_ddf().compute().shape, valid.to_ddf().compute().shape)) experiment = Experiment(api_key=os.getenv('COMET_API_KEY'), project_name=self.COMET_PROJECT_NAME) self.comet_experiment_key = experiment.get_key() experiment.add_tag(current.pathspec) experiment.log_parameters(self.hypers) user_schema = train.schema.select_by_tag(Tags.USER) user_inputs = mm.InputBlockV2(user_schema) query = mm.Encoder(user_inputs, mm.MLPBlock([128, 64])) item_schema = train.schema.select_by_tag(Tags.ITEM) item_inputs = mm.InputBlockV2(item_schema) candidate = mm.Encoder(item_inputs, mm.MLPBlock([128, 64])) model = mm.TwoTowerModelV2(query, candidate) opt = tf.keras.optimizers.Adagrad(learning_rate=self.hypers['LEARNING_RATE']) model.compile(optimizer=opt, run_eagerly=False, metrics=[mm.RecallAt(int(self.TOP_K)), mm.NDCGAt(int(self.TOP_K))]) model.fit(train, validation_data=valid, batch_size=self.hypers['BATCH_SIZE'], epochs=int(self.N_EPOCHS)) self.metrics = model.evaluate(valid, batch_size=1024, return_dict=True) print('\n\n====> Eval results: {}\n\n'.format(self.metrics)) model_hash = str(hashlib.md5(self.hyper_string.encode('utf-8')).hexdigest()) self.model_path = 'merlin/model{}/'.format(model_hash) model.save(self.model_path) print('Model saved!') self.next(self.join_runs) def get_items_topk_recommender_model(self, train_dataset, model, k: int): from merlin.models.utils.dataset import unique_rows_by_features from merlin.schema.tags import Tags candidate_features = unique_rows_by_features(train_dataset, Tags.ITEM, Tags.ITEM_ID) topk_model = model.to_top_k_encoder(candidate_features, k=k, batch_size=128) topk_model.compile(run_eagerly=False) return topk_model def join_runs(self, inputs): self.model_paths = {inp.hyper_string: inp.model_path for inp in inputs} self.experiment_keys = {inp.hyper_string: inp.comet_experiment_key for inp in inputs} self.results_from_runs = {inp.hyper_string: inp.metrics[self.VALIDATION_METRIC] for inp in inputs} print('Current results: {}'.format(self.results_from_runs)) (self.best_model, self_best_result) = sorted(self.results_from_runs.items(), key=(lambda x: x[1]), reverse=True)[0] print('Best model is: {}, best path is {}'.format(self.best_model, self.model_paths[self.best_model])) self.final_model_path = self.model_paths[self.best_model] self.item_id_2_meta = inputs[0].item_id_2_meta self.id_2_item_id = inputs[0].id_2_item_id self.id_2_user_id = inputs[0].id_2_user_id self.magicdir = inputs[0].magicdir self.experiment_key = self.experiment_keys[self.best_model] self.next(self.model_testing) def prepare_predictions_for_comet_panel(self, h_m_shoppers, best_predictions, item_id_2_meta, api_key, experiment_key): from comet_ml import ExistingExperiment n_shoppers = 10 predictions_to_log = [] for shopper in h_m_shoppers[:n_shoppers]: cnt_predictions = best_predictions.get(shopper, None) if (not cnt_predictions): continue for p in cnt_predictions['items']: product_type = (item_id_2_meta[p]['product_group_name'] if (p in item_id_2_meta) else 'NO_GROUP') predictions_to_log.append({'user_id': shopper, 'product_id': p, 'product_type': product_type, 'score': 1.0}) experiment = ExistingExperiment(api_key=api_key, experiment_key=experiment_key) experiment.log_asset_data(predictions_to_log, name='predictions.json') return predictions_to_log def load_merlin_model(self, dataset, path): import tensorflow as tf import merlin.models.tf as mm loaded_model = tf.keras.models.load_model(path) _ = loaded_model(mm.sample_batch(dataset, batch_size=128, include_targets=False)) print('Model re-loaded!') return loaded_model (vars={'EN_BATCH': os.getenv('EN_BATCH')}) _decorator(batch(memory=24000, image='public.ecr.aws/outerbounds/merlin-reasonable-scale:22.11-latest'), flag=os.getenv('EN_BATCH')) def model_testing(self): from merlin.io.dataset import Dataset import merlin.models.tf as mm from merlin.schema import Tags test = Dataset('merlin/test/*.parquet') loaded_model = self.load_merlin_model(test, self.final_model_path) topk_rec_model = self.get_items_topk_recommender_model(test, loaded_model, k=int(self.TOP_K)) test_loader = mm.Loader(test, batch_size=1024, transform=mm.ToTarget(test.schema, Tags.ITEM_ID)) self.test_metrics = topk_rec_model.evaluate(test_loader, batch_size=1024, return_dict=True) print('\n\n====> Test results: {}\n\n'.format(self.test_metrics)) self.next(self.saving_predictions) (vars={'EN_BATCH': os.getenv('EN_BATCH'), 'COMET_API_KEY': os.getenv('COMET_API_KEY')}) _decorator(batch(image='public.ecr.aws/outerbounds/merlin-reasonable-scale:22.11-latest'), flag=os.getenv('EN_BATCH')) (libraries={'requests': '2.28.1', 'comet-ml': '3.26.0'}) def saving_predictions(self): from merlin.io.dataset import Dataset import merlin.models.tf as mm train = Dataset('merlin/train/*.parquet') test = Dataset('merlin/test/*.parquet') loaded_model = self.load_merlin_model(test, self.final_model_path) topk_rec_model = self.get_items_topk_recommender_model(train, loaded_model, k=int(self.TOP_K)) test_dataset = mm.Loader(test, batch_size=1024, shuffle=False) self.raw_predictions = topk_rec_model.predict(test_dataset)[1] n_rows = self.raw_predictions.shape[0] self.target_shoppers = test_dataset.data.to_ddf().compute()['customer_id'] print('Inspect the shopper object for debugging...{}'.format(type(self.target_shoppers))) assert (n_rows == len(self.target_shoppers)) self.h_m_shoppers = [str(self.id_2_user_id[_]) for _ in self.target_shoppers.to_numpy().tolist()] print('Example target shoppers: ', self.h_m_shoppers[:3]) self.target_items = test_dataset.data.to_ddf().compute()['article_id'] print('Example target items: ', self.target_items[:3]) self.best_predictions = self.serialize_predictions(self.h_m_shoppers, self.id_2_item_id, self.raw_predictions, self.target_items, n_rows) print('Example target predictions', self.best_predictions[self.h_m_shoppers[0]]) self.prepare_predictions_for_comet_panel(self.h_m_shoppers, self.best_predictions, self.item_id_2_meta, os.getenv('COMET_API_KEY'), self.experiment_key) print(n_rows, len(self.best_predictions)) self.next(self.export_to_app) def serialize_predictions(self, h_m_shoppers, id_2_item_id, raw_predictions, target_items, n_rows): sku_convert = (lambda x: [str(id_2_item_id[_]) for _ in x]) predictions = {} for _ in range(n_rows): cnt_user = h_m_shoppers[_] cnt_raw_preds = raw_predictions[_].tolist() cnt_target = target_items[_] if (cnt_user not in predictions): predictions[cnt_user] = {'items': sku_convert(cnt_raw_preds), 'target': sku_convert([cnt_target])[0]} return predictions def export_to_app(self): if (not (os.environ.get('EXPORT_TO_APP', None) == '1')): print('Skipping exporting data to the CLIP-based Streamlit app.') else: import pandas as pd from app_utils import encode_image import torch from transformers import CLIPProcessor, CLIPModel rows = [] max_preds = 100 for (shopper, preds) in self.best_predictions.items(): target_item = preds['target'] target_img_url = self.item_id_2_meta[target_item]['s3_url'] top_pred = preds['items'][0] predicted_img_url = self.item_id_2_meta[top_pred]['s3_url'] if ((not target_img_url) or (not predicted_img_url)): continue new_row = {'user_id': shopper, 'target_item': target_item, 'predicted_item': top_pred, 'target_image_url': target_img_url, 'predicted_image_url': predicted_img_url, 'product_type': self.item_id_2_meta[target_item]['product_group_name']} rows.append(new_row) if (len(rows) >= max_preds): break df = pd.DataFrame(rows) assert (len(df) == max_preds) device = ('cuda' if torch.cuda.is_available() else 'cpu') model = CLIPModel.from_pretrained('openai/clip-vit-base-patch32').to(device) processor = CLIPProcessor.from_pretrained('openai/clip-vit-base-patch32') img_vectors = [] for img in list(df['target_image_url']): cnt_vector = encode_image(model, processor, img, device) img_vectors.append(cnt_vector[0]) df['image_vectors'] = img_vectors self.prediction_df = df self.next(self.cache_predictions) def cache_predictions(self): if (not bool(int(os.getenv('SAVE_TO_CACHE')))): print('Skipping deployment') else: print('Caching predictions in DynamoDB') import boto3 dynamodb = boto3.resource('dynamodb', region_name='us-west-2') table = dynamodb.Table(self.DYNAMO_TABLE) data = [{'userId': user, 'recs': json.dumps(recs)} for (user, recs) in self.best_predictions.items()] data.append({'userId': 'no_user', 'recs': json.dumps(['test_rec_{}'.format(_) for _ in range(int(self.TOP_K))])}) with table.batch_writer() as writer: for item in data: writer.put_item(Item=item) print('Predictions are all cached in DynamoDB') self.next(self.end) def end(self): print('All done\n\nSee you, recSys cowboy\n') return
def split_data_slice(data, output_file, slice_id, days_offset, days_train, days_test): data_start = datetime.fromtimestamp(data.Time.min(), timezone.utc) data_end = datetime.fromtimestamp(data.Time.max(), timezone.utc) print('Full data set {}\n\tEvents: {}\n\tSessions: {}\n\tItems: {}\n\tSpan: {} / {}'.format(slice_id, len(data), data.SessionId.nunique(), data.ItemId.nunique(), data_start.isoformat(), data_end.isoformat())) start = (datetime.fromtimestamp(data.Time.min(), timezone.utc) + datetime.timedelta(days_offset)) middle = (start + datetime.timedelta(days_train)) end = (middle + datetime.timedelta(days_test)) session_max_times = data.groupby('SessionId').Time.max() greater_start = session_max_times[(session_max_times >= start.timestamp())].index lower_end = session_max_times[(session_max_times <= end.timestamp())].index data_filtered = data[np.in1d(data.SessionId, greater_start.intersection(lower_end))] print('Slice data set {}\n\tEvents: {}\n\tSessions: {}\n\tItems: {}\n\tSpan: {} / {} / {}'.format(slice_id, len(data_filtered), data_filtered.SessionId.nunique(), data_filtered.ItemId.nunique(), start.date().isoformat(), middle.date().isoformat(), end.date().isoformat())) session_max_times = data_filtered.groupby('SessionId').Time.max() sessions_train = session_max_times[(session_max_times < middle.timestamp())].index sessions_test = session_max_times[(session_max_times >= middle.timestamp())].index train = data[np.in1d(data.SessionId, sessions_train)] print('Train set {}\n\tEvents: {}\n\tSessions: {}\n\tItems: {}\n\tSpan: {} / {}'.format(slice_id, len(train), train.SessionId.nunique(), train.ItemId.nunique(), start.date().isoformat(), middle.date().isoformat())) train.to_csv((((output_file + '_train_full.') + str(slice_id)) + '.txt'), sep='\t', index=False) test = data[np.in1d(data.SessionId, sessions_test)] test = test[np.in1d(test.ItemId, train.ItemId)] tslength = test.groupby('SessionId').size() test = test[np.in1d(test.SessionId, tslength[(tslength >= 2)].index)] print('Test set {}\n\tEvents: {}\n\tSessions: {}\n\tItems: {}\n\tSpan: {} / {} \n\n'.format(slice_id, len(test), test.SessionId.nunique(), test.ItemId.nunique(), middle.date().isoformat(), end.date().isoformat())) test.to_csv((((output_file + '_test.') + str(slice_id)) + '.txt'), sep='\t', index=False)
class MetricList(): def __init__(self, metrics): assert isinstance(metrics, dict), "'metrics' must be a dictionary of callables" self.metrics = metrics self.results = {key: 0.0 for key in self.metrics.keys()} def __call__(self, y_out, y_batch): for (key, value) in self.metrics.items(): self.results[key] += value(y_out, y_batch) def reset(self): self.results = {key: 0.0 for key in self.metrics.keys()} def get_results(self, normalize=False): assert (isinstance(normalize, bool) or isinstance(normalize, Number)), "'normalize' must be boolean or a number" if (not normalize): return self.results else: return {key: (value / normalize) for (key, value) in self.results.items()}
def TorchComplexMul(v1_complex, v2_complex): (v1_real, v1_imag) = v1_complex.chunk(2, dim=(- 1)) (v2_real, v2_imag) = v2_complex.chunk(2, dim=(- 1)) return torch.cat((((v1_real * v2_real) - (v1_imag * v2_imag)), ((v1_real * v2_imag) + (v1_imag * v2_real))), dim=(- 1))
def mock_k8s_client(): k8s_client = k8sClient.singleton_instance('default') k8s_client.get_custom_resource = _get_training_job k8s_client.get_pod = _get_pod k8s_client.list_namespaced_pod = mock_list_namespaced_pod k8s_client.create_custom_resource = mock.MagicMock(return_value=True) k8s_client.delete_custom_resource = mock.MagicMock(return_value=True) k8s_client.create_pod = mock.MagicMock(return_value=True) k8s_client.delete_pod = mock.MagicMock(return_value=True) k8s_client.create_service = mock.MagicMock(return_value=True) k8s_client.get_service = mock.MagicMock(return_value=False)
def convert_examples_to_features(examples, label_list, max_seq_length, tokenizer, output_mode, cls_token_at_end=False, pad_on_left=False, cls_token='[CLS]', sep_token='[SEP]', pad_token=0, sequence_a_segment_id=0, sequence_b_segment_id=1, cls_token_segment_id=1, pad_token_segment_id=0, mask_padding_with_zero=True, do_lower_case=False, is_multi_choice=True): label_map = {label: i for (i, label) in enumerate(label_list)} num_labels = len(label_list) if is_multi_choice: features = [[]] else: features = [] for (ex_index, example) in enumerate(examples): if do_lower_case: example.text_a = example.text_a.lower() example.text_b = example.text_b.lower() example.text_c = example.text_c.lower() if ((ex_index % 10000) == 0): logger.info(('Writing example %d of %d' % (ex_index, len(examples)))) tokens_a = tokenizer.tokenize(example.text_a) tokens_b = None tokens_c = None if (example.text_b and example.text_c): tokens_b = tokenizer.tokenize(example.text_b) tokens_c = tokenizer.tokenize(example.text_c) _truncate_seq_tuple(tokens_a, tokens_b, tokens_c, (max_seq_length - 4)) elif (example.text_b and (not example.text_c)): tokens_b = tokenizer.tokenize(example.text_b) _truncate_seq_pair(tokens_a, tokens_b, (max_seq_length - 3)) elif (len(tokens_a) > (max_seq_length - 2)): tokens_a = tokens_a[:(max_seq_length - 2)] tokens = (tokens_a + [sep_token]) segment_ids = ([sequence_a_segment_id] * len(tokens)) if tokens_c: tokens_b += ([sep_token] + tokens_c) if tokens_b: tokens += (tokens_b + [sep_token]) segment_ids += ([sequence_b_segment_id] * (len(tokens_b) + 1)) if cls_token_at_end: tokens = (tokens + [cls_token]) segment_ids = (segment_ids + [cls_token_segment_id]) else: tokens = ([cls_token] + tokens) segment_ids = ([cls_token_segment_id] + segment_ids) input_ids = tokenizer.convert_tokens_to_ids(tokens) input_mask = ([(1 if mask_padding_with_zero else 0)] * len(input_ids)) padding_length = (max_seq_length - len(input_ids)) if pad_on_left: input_ids = (([pad_token] * padding_length) + input_ids) input_mask = (([(0 if mask_padding_with_zero else 1)] * padding_length) + input_mask) segment_ids = (([pad_token_segment_id] * padding_length) + segment_ids) else: input_ids = (input_ids + ([pad_token] * padding_length)) input_mask = (input_mask + ([(0 if mask_padding_with_zero else 1)] * padding_length)) segment_ids = (segment_ids + ([pad_token_segment_id] * padding_length)) assert (len(input_ids) == max_seq_length) assert (len(input_mask) == max_seq_length) assert (len(segment_ids) == max_seq_length) if (output_mode in ['classification', 'multi-choice']): label_id = label_map[example.label] elif (output_mode == 'regression'): label_id = float(example.label) else: raise KeyError(output_mode) if is_multi_choice: features[(- 1)].append(InputFeatures(input_ids=input_ids, input_mask=input_mask, segment_ids=segment_ids, label_id=label_id)) if (len(features[(- 1)]) == num_labels): features.append([]) else: features.append(InputFeatures(input_ids=input_ids, input_mask=input_mask, segment_ids=segment_ids, label_id=label_id)) if is_multi_choice: if (len(features[(- 1)]) == 0): features = features[:(- 1)] return features
class ErnieForMaskedLM(metaclass=DummyObject): _backends = ['torch'] def __init__(self, *args, **kwargs): requires_backends(self, ['torch'])
def get_bernoulli_vae_schema(config): return [{'type': 'flatten'}, {'type': 'bernoulli-likelihood', 'num_z_channels': config['num_z_channels'], 'logit_net': {'type': 'mlp', 'activation': 'tanh', 'hidden_channels': config['logit_net']}, 'q_coupler': get_q_coupler_config(config, flattened=True)}]
def build_fake_yaml(): fake_yaml = '\n model:\n name: fake_yaml\n framework: tensorflow\n inputs: input\n device: cpu\n quantization:\n model_wise:\n weight:\n granularity: per_tensor\n scheme: sym\n dtype: int8\n algorithm: minmax\n evaluation:\n accuracy:\n metric:\n topk: 1\n tuning:\n strategy:\n name: basic\n accuracy_criterion:\n relative: 0.1\n exit_policy:\n performance_only: True\n workspace:\n path: saved\n ' y = yaml.load(fake_yaml, Loader=yaml.SafeLoader) with open('fake_yaml.yaml', 'w', encoding='utf-8') as f: yaml.dump(y, f) f.close()
def test_set_reactivity(): assert (flow().mut_settings.reactivity_mode == ReactivityMode.BATCH) run_cell(f'%flow reactivity {ReactivityMode.INCREMENTAL.value}') assert (flow().mut_settings.reactivity_mode == ReactivityMode.INCREMENTAL) run_cell(f'%flow reactivity {ReactivityMode.BATCH.value}') assert (flow().mut_settings.reactivity_mode == ReactivityMode.BATCH)
class Residual(nn.Module): def __init__(self, do_batchnorm, c, **kw): super().__init__() self.res1 = ConvBN(do_batchnorm, c, c, **kw) self.res2 = ConvBN(do_batchnorm, c, c, **kw) def forward(self, x): return (x + F.relu(self.res2(self.res1(x)))) def prep_finetune(self, iid, c, **kw): layers = [self.res1, self.res2] return itertools.chain.from_iterable([l.prep_finetune(iid, c, c, **kw) for l in layers])
_module() class DistSamplerSeedHook(Hook): def before_epoch(self, runner): if hasattr(runner.data_loader.sampler, 'set_epoch'): runner.data_loader.sampler.set_epoch(runner.epoch) elif hasattr(runner.data_loader.batch_sampler.sampler, 'set_epoch'): runner.data_loader.batch_sampler.sampler.set_epoch(runner.epoch)
def init_yolov3(args, device): import torch from models.yolo import Model from utils.google_utils import attempt_download from utils.torch_utils import intersect_dicts, torch_distributed_zero_first log.info('Loading yolov3.pt weights.') hyp = args.yolo_hyp() with torch_distributed_zero_first(args.global_rank): attempt_download('yolov3.pt') ckpt = torch.load('yolov3.pt', map_location=device) if hyp.get('anchors'): ckpt['model'].yaml['anchors'] = round(hyp['anchors']) net = Model((args.cfg or ckpt['model'].yaml), ch=3, nc=args.nc).to(device) exclude = (['anchor'] if (args.cfg or hyp.get('anchors')) else []) state_dict = ckpt['model'].float().state_dict() state_dict = intersect_dicts(state_dict, net.state_dict(), exclude=exclude) net.load_state_dict(state_dict, strict=False) net.to(device) return net
class ConditionalBatchNorm2d(nn.Module): def __init__(self, num_features, num_classes, eps=0.0001, momentum=0.1): super().__init__() self.num_features = num_features self.bn = nn.BatchNorm2d(num_features, affine=False, eps=eps, momentum=momentum) self.gamma_embed = SpectralNorm(nn.Linear(num_classes, num_features, bias=False)) self.beta_embed = SpectralNorm(nn.Linear(num_classes, num_features, bias=False)) def forward(self, x, y): out = self.bn(x) gamma = (self.gamma_embed(y) + 1) beta = self.beta_embed(y) out = ((gamma.view((- 1), self.num_features, 1, 1) * out) + beta.view((- 1), self.num_features, 1, 1)) return out
def deeplabv3plus_pvtv2(num_classes=1, output_stride=8, pretrained_backbone=True): return _segm_pvtv2('deeplabv3plus', 'pvtv2', num_classes, output_stride=output_stride, pretrained_backbone=pretrained_backbone)
def test_eddington_differentpotentials_dMdE_integral(): pots = [potential.PlummerPotential(amp=2.3, b=1.3), potential.PowerSphericalPotentialwCutoff(amp=1.3, alpha=1.9, rc=1.2)] tols = [1e-06 for pot in pots] for (pot, tol) in zip(pots, tols): dfh = eddingtondf(pot=pot) check_dMdE_integral(dfh, tol) return None
def set_visible_area(point_dict, axes): min_x = .0 min_y = .0 max_x = (- .0) max_y = (- .0) for (id, point) in dict_utils.get_item_iterator(point_dict): min_x = min(point.x, min_x) min_y = min(point.y, min_y) max_x = max(point.x, max_x) max_y = max(point.y, max_y) axes.set_aspect('equal', adjustable='box') axes.set_xlim([(min_x - 10), (max_x + 10)]) axes.set_ylim([(min_y - 10), (max_y + 10)])
def prototype_test(): state = prototype_state() state['train_dialogues'] = './tests/data/ttrain.dialogues.pkl' state['test_dialogues'] = './tests/data/ttest.dialogues.pkl' state['valid_dialogues'] = './tests/data/tvalid.dialogues.pkl' state['dictionary'] = './tests/data/ttrain.dict.pkl' state['save_dir'] = './tests/models/' state['max_grad_steps'] = 20 state['initialize_from_pretrained_word_embeddings'] = False state['pretrained_word_embeddings_file'] = './tests/data/MT_WordEmb.pkl' state['fix_pretrained_word_embeddings'] = False state['valid_freq'] = 50 state['prefix'] = 'testmodel_' state['updater'] = 'adam' state['maxout_out'] = False state['deep_out'] = True state['deep_dialogue_input'] = True state['utterance_encoder_gating'] = 'GRU' state['dialogue_encoder_gating'] = 'GRU' state['utterance_decoder_gating'] = 'GRU' state['bidirectional_utterance_encoder'] = True state['direct_connection_between_encoders_and_decoder'] = True state['bs'] = 5 state['sort_k_batches'] = 1 state['use_nce'] = False state['decoder_bias_type'] = 'all' state['qdim_encoder'] = 15 state['qdim_decoder'] = 5 state['sdim'] = 10 state['rankdim'] = 10 return state
def get_processor_name(): if (platform.system() == 'Windows'): return platform.processor() elif (platform.system() == 'Darwin'): os.environ['PATH'] = ((os.environ['PATH'] + os.pathsep) + '/usr/sbin') command = 'sysctl -n machdep.cpu.brand_string' return subprocess.check_output(command).strip() elif (platform.system() == 'Linux'): command = "cat /proc/cpuinfo | grep 'model name' -m 1" name = subprocess.check_output(command, shell=True).strip() return str(name, 'utf-8')
def main(): args = get_argument() if args.resnet: import torchvision.models as models model = models.resnet18(pretrained=True) model = ProbModel(model) else: model = mobilenet_v2('modeling/classification/mobilenetv2_1.0-f2a8633.pth.tar') model = ProbModel(model) model.eval() if args.quantize: data = torch.ones((4, 3, 224, 224)) if args.distill_range: import copy model_original = copy.deepcopy(model) model_original.eval() transformer = TorchTransformer() transformer._build_graph(model_original, data, [QuantMeasure]) graph = transformer.log.getGraph() bottoms = transformer.log.getBottoms() data_distill = getDistilData(model_original, 'imagenet', args.dis_batch_size, bn_merged=False, num_batch=args.dis_num_batch, gpu=True, value_range=[(- 2.), 2.64], size=[224, 224], early_break_factor=(1.2 if args.resnet else 0.5)) transformer = TorchTransformer() module_dict = {} if args.distill_range: module_dict[1] = [(torch.nn.Conv2d, QConv2d), (torch.nn.Linear, QLinear)] else: module_dict[1] = [(torch.nn.Conv2d, QuantNConv2d), (torch.nn.Linear, QuantNLinear)] if (args.relu or args.equalize): module_dict[0] = [(torch.nn.ReLU6, torch.nn.ReLU)] (model, transformer) = switch_layers(model, transformer, data, module_dict, ignore_layer=[QuantMeasure], quant_op=True) graph = transformer.log.getGraph() bottoms = transformer.log.getBottoms() if args.distill_range: targ_layer = [QConv2d, QLinear] else: targ_layer = [QuantNConv2d, QuantNLinear] set_layer_bits(graph, args.bits_weight, args.bits_activation, args.bits_bias, targ_layer) model = merge_batchnorm(model, graph, bottoms, targ_layer) if (args.equalize or args.distill_range): res = create_relation(graph, bottoms, targ_layer, delete_single=False) if args.equalize: cross_layer_equalization(graph, res, targ_layer, visualize_state=False, converge_thres=2e-07, signed=True) if args.clip_weight: clip_weight(graph, range_clip=[(- 15), 15], targ_type=targ_layer) if args.correction: bias_correction(graph, bottoms, targ_layer, bits_weight=args.bits_weight, signed=True) if args.distill_range: set_update_stat(model, [QuantMeasure], True) model = update_quant_range(model.cuda(), data_distill, graph, bottoms) set_update_stat(model, [QuantMeasure], False) else: set_quant_minmax(graph, bottoms) torch.cuda.empty_cache() module_dict = {} if args.distill_range: module_dict[1] = [(QConv2d, torch.nn.Conv2d), (QLinear, torch.nn.Linear)] else: module_dict[1] = [(QuantNConv2d, torch.nn.Conv2d), (QuantNLinear, torch.nn.Linear)] (model, transformer) = switch_layers(model, transformer, data, module_dict, ignore_layer=[QuantMeasure], quant_op=False) graph = transformer.log.getGraph() bottoms = transformer.log.getBottoms() x = torch.rand(1, 3, 224, 224) torch_out = torch.onnx._export(model, x, 'model.onnx', export_params=True) os.system('python3 -m onnxsim model.onnx model-sim.onnx') os.system('rm model.onnx') cur_path = os.path.abspath(os.getcwd()) os.system('mv model-sim.onnx {}'.format(os.path.join(args.ncnn_build, 'tools/onnx', 'model-sim.onnx'))) os.chdir(os.path.join(args.ncnn_build, 'tools/onnx')) os.system('./onnx2ncnn model-sim.onnx model.param model.bin') lines = [line.strip() for line in open('model.param', 'r')] with open('model.param', 'w') as ww: for (idx, line) in enumerate(lines): if ((idx == 2) and ('input' in line.lower())): line += ' 0=224 1=224 2=3' ww.write((line + '\n')) if (not os.path.exists(os.path.join(cur_path, 'modeling/ncnn'))): os.makedirs(os.path.join(cur_path, 'modeling/ncnn')) os.system('rm model-sim.onnx') if args.quantize: os.system('mv model.param {}'.format(os.path.join(args.ncnn_build, 'tools/quantize', 'model.param'))) os.system('mv model.bin {}'.format(os.path.join(args.ncnn_build, 'tools/quantize', 'model.bin'))) os.chdir(os.path.join(args.ncnn_build, 'tools/quantize')) os.system('./ncnn2table --param=model.param --bin=model.bin --images={} --output=model_int8_channel.table --mean={},{},{} --norm={},{},{} --size=224,224 --thread=2'.format(args.image_path, (0.485 * 255), (0.456 * 255), (0.406 * 255), (1 / (0.229 * 255)), (1 / (0.224 * 255)), (1 / (0.225 * 255)))) table_old = [line.strip() for line in open('model_int8_channel.table', 'r')] table_new = [] count = 0 for ii in range(2): for idx in graph: if (type(graph[idx]) in [torch.nn.Conv2d, torch.nn.Linear]): if (ii == 0): mi = float(torch.min(graph[idx].weight)) ma = float(torch.max(graph[idx].weight)) else: mi = float(torch.min(graph[idx].quant.running_min)) ma = float(torch.max(graph[idx].quant.running_max)) scale = (128.0 / max(abs(ma), abs(mi))) if (ii == 0): table_new.append(' '.join((table_old[count].split(' ')[0:1] + ([str(scale)] * graph[idx].weight.shape[0])))) else: table_new.append(' '.join((table_old[count].split(' ')[0:1] + [str(scale)]))) count += 1 with open('model_int8_tensor.table', 'w') as ww: for line in table_new: ww.write((line + '\n')) os.system('./ncnn2int8 model.param model.bin model_int8.param model_int8.bin model_int8_tensor.table') lines = [line.strip() for line in open('model_int8.param', 'r')] os.system('cp model_int8.param {}'.format(os.path.join(cur_path, args.param))) os.system('cp model_int8.bin {}'.format(os.path.join(cur_path, args.bin))) os.system('cp model_int8_tensor.table {}'.format(os.path.join(cur_path, args.table))) else: os.system('mv model.param {}'.format(os.path.join(cur_path, args.param))) os.system('mv model.bin {}'.format(os.path.join(cur_path, args.bin))) os.chdir(cur_path) line = ' '.join([l.strip() for l in open(args.param, 'r')][(- 1)].split()).split(' ')[1] print(('=' * 100)) print("Target layer name '{}'".format(line)) print(('=' * 100))
class AlbertTokenizerFast(PreTrainedTokenizerFast): vocab_files_names = VOCAB_FILES_NAMES pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES slow_tokenizer_class = AlbertTokenizer def __init__(self, vocab_file=None, tokenizer_file=None, do_lower_case=True, remove_space=True, keep_accents=False, bos_token='[CLS]', eos_token='[SEP]', unk_token='<unk>', sep_token='[SEP]', pad_token='<pad>', cls_token='[CLS]', mask_token='[MASK]', **kwargs): mask_token = (AddedToken(mask_token, lstrip=True, rstrip=False, normalized=False) if isinstance(mask_token, str) else mask_token) super().__init__(vocab_file, tokenizer_file=tokenizer_file, do_lower_case=do_lower_case, remove_space=remove_space, keep_accents=keep_accents, bos_token=bos_token, eos_token=eos_token, unk_token=unk_token, sep_token=sep_token, pad_token=pad_token, cls_token=cls_token, mask_token=mask_token, **kwargs) self.do_lower_case = do_lower_case self.remove_space = remove_space self.keep_accents = keep_accents self.vocab_file = vocab_file self.can_save_slow_tokenizer = (False if (not self.vocab_file) else True) 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 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 self.can_save_slow_tokenizer): raise ValueError('Your fast tokenizer does not have the necessary information to save the vocabulary for a slow tokenizer.') 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)): copyfile(self.vocab_file, out_vocab_file) return (out_vocab_file,)
class HRModule(nn.Module): def __init__(self, num_branches, blocks, num_blocks, in_channels, num_channels, multiscale_output=True, with_cp=False, conv_cfg=None, norm_cfg=dict(type='BN')): super(HRModule, self).__init__() self._check_branches(num_branches, num_blocks, in_channels, num_channels) self.in_channels = in_channels self.num_branches = num_branches self.multiscale_output = multiscale_output self.norm_cfg = norm_cfg self.conv_cfg = conv_cfg self.with_cp = with_cp self.branches = self._make_branches(num_branches, blocks, num_blocks, num_channels) self.fuse_layers = self._make_fuse_layers() self.relu = nn.ReLU(inplace=False) def _check_branches(self, num_branches, num_blocks, in_channels, num_channels): if (num_branches != len(num_blocks)): error_msg = f'NUM_BRANCHES({num_branches}) != NUM_BLOCKS({len(num_blocks)})' raise ValueError(error_msg) if (num_branches != len(num_channels)): error_msg = f'NUM_BRANCHES({num_branches}) != NUM_CHANNELS({len(num_channels)})' raise ValueError(error_msg) if (num_branches != len(in_channels)): error_msg = f'NUM_BRANCHES({num_branches}) != NUM_INCHANNELS({len(in_channels)})' raise ValueError(error_msg) def _make_one_branch(self, branch_index, block, num_blocks, num_channels, stride=1): downsample = None if ((stride != 1) or (self.in_channels[branch_index] != (num_channels[branch_index] * block.expansion))): downsample = nn.Sequential(build_conv_layer(self.conv_cfg, self.in_channels[branch_index], (num_channels[branch_index] * block.expansion), kernel_size=1, stride=stride, bias=False), build_norm_layer(self.norm_cfg, (num_channels[branch_index] * block.expansion))[1]) layers = [] layers.append(block(self.in_channels[branch_index], num_channels[branch_index], stride, downsample=downsample, with_cp=self.with_cp, norm_cfg=self.norm_cfg, conv_cfg=self.conv_cfg)) self.in_channels[branch_index] = (num_channels[branch_index] * block.expansion) for i in range(1, num_blocks[branch_index]): layers.append(block(self.in_channels[branch_index], num_channels[branch_index], with_cp=self.with_cp, norm_cfg=self.norm_cfg, conv_cfg=self.conv_cfg)) return nn.Sequential(*layers) def _make_branches(self, num_branches, block, num_blocks, num_channels): branches = [] for i in range(num_branches): branches.append(self._make_one_branch(i, block, num_blocks, num_channels)) return nn.ModuleList(branches) def _make_fuse_layers(self): if (self.num_branches == 1): return None num_branches = self.num_branches in_channels = self.in_channels fuse_layers = [] num_out_branches = (num_branches if self.multiscale_output else 1) for i in range(num_out_branches): fuse_layer = [] for j in range(num_branches): if (j > i): fuse_layer.append(nn.Sequential(build_conv_layer(self.conv_cfg, in_channels[j], in_channels[i], kernel_size=1, stride=1, padding=0, bias=False), build_norm_layer(self.norm_cfg, in_channels[i])[1], nn.Upsample(scale_factor=(2 ** (j - i)), mode='nearest'))) elif (j == i): fuse_layer.append(None) else: conv_downsamples = [] for k in range((i - j)): if (k == ((i - j) - 1)): conv_downsamples.append(nn.Sequential(build_conv_layer(self.conv_cfg, in_channels[j], in_channels[i], kernel_size=3, stride=2, padding=1, bias=False), build_norm_layer(self.norm_cfg, in_channels[i])[1])) else: conv_downsamples.append(nn.Sequential(build_conv_layer(self.conv_cfg, in_channels[j], in_channels[j], kernel_size=3, stride=2, padding=1, bias=False), build_norm_layer(self.norm_cfg, in_channels[j])[1], nn.ReLU(inplace=False))) fuse_layer.append(nn.Sequential(*conv_downsamples)) fuse_layers.append(nn.ModuleList(fuse_layer)) return nn.ModuleList(fuse_layers) def forward(self, x): if (self.num_branches == 1): return [self.branches[0](x[0])] for i in range(self.num_branches): x[i] = self.branches[i](x[i]) x_fuse = [] for i in range(len(self.fuse_layers)): y = 0 for j in range(self.num_branches): if (i == j): y += x[j] else: y += self.fuse_layers[i][j](x[j]) x_fuse.append(self.relu(y)) return x_fuse
def create_model(bert_config, is_training, input_ids, input_mask, P_mask, A_mask, B_mask, segment_ids, labels, num_labels, use_one_hot_embeddings): model = modeling.BertModel(config=bert_config, is_training=is_training, input_ids=input_ids, input_mask=input_mask, token_type_ids=segment_ids, use_one_hot_embeddings=use_one_hot_embeddings) all_out = model.get_sequence_output() hidden_size = all_out.shape[(- 1)].value _P_mask = tf.cast(P_mask, tf.float32) _A_mask = tf.cast(A_mask, tf.float32) _B_mask = tf.cast(B_mask, tf.float32) _P_mask_ = tf.broadcast_to(_P_mask, shape=(tf.shape(all_out)[2], tf.shape(all_out)[0], tf.shape(all_out)[1])) P_mask_ = tf.transpose(_P_mask_, perm=[1, 2, 0]) _A_mask_ = tf.broadcast_to(_A_mask, shape=(tf.shape(all_out)[2], tf.shape(all_out)[0], tf.shape(all_out)[1])) A_mask_ = tf.transpose(_A_mask_, perm=[1, 2, 0]) _B_mask_ = tf.broadcast_to(_B_mask, shape=(tf.shape(all_out)[2], tf.shape(all_out)[0], tf.shape(all_out)[1])) B_mask_ = tf.transpose(_B_mask_, perm=[1, 2, 0]) P_ = tf.multiply(all_out, P_mask_) P = tf.reduce_sum(P_, axis=1) A_ = tf.multiply(all_out, A_mask_) A = tf.reduce_sum(A_, axis=1) B_ = tf.multiply(all_out, B_mask_) B = tf.reduce_sum(B_, axis=1) PA = tf.multiply(P, A) PB = tf.multiply(P, B) PP = tf.multiply(P, P) AB = tf.multiply(A, B) N = tf.subtract(PP, AB) AB_weights = tf.get_variable('AB_weights', [1, hidden_size], initializer=tf.truncated_normal_initializer(stddev=0.02)) N_weights = tf.get_variable('N_weights', [1, hidden_size], initializer=tf.truncated_normal_initializer(stddev=0.02)) if is_training: AB_weights = tf.nn.dropout(AB_weights, keep_prob=0.9) N_weights = tf.nn.dropout(N_weights, keep_prob=0.9) A_out = tf.matmul(PA, AB_weights, transpose_b=True) B_out = tf.matmul(PB, AB_weights, transpose_b=True) N_out = tf.matmul(N, N_weights, transpose_b=True) output_bias = tf.get_variable('output_bias', [num_labels], initializer=tf.zeros_initializer()) with tf.variable_scope('loss'): logits = tf.concat([A_out, B_out, N_out], axis=1) logits = tf.nn.bias_add(logits, output_bias) probabilities = tf.nn.softmax(logits, axis=(- 1)) log_probs = tf.nn.log_softmax(logits, axis=(- 1)) one_hot_labels = tf.one_hot(labels, depth=num_labels, dtype=tf.float32) per_example_loss = (- tf.reduce_sum((one_hot_labels * log_probs), axis=(- 1))) loss = tf.reduce_mean(per_example_loss) return (loss, per_example_loss, logits, probabilities)
def checkpoint(step, epoch): model_out_path = 'models/{}/GFN_epoch_{}.pkl'.format(step, epoch) torch.save(model, model_out_path) print('===>Checkpoint saved to {}'.format(model_out_path))
class ounoise(): def __init__(self, std, action_dim, mean=0, theta=0.15, dt=0.01, x0=None): self.std = std self.mean = mean self.action_dim = action_dim self.theta = theta self.dt = dt self.x0 = x0 def reset(self): self.x_prev = (self.x0 if (self.x0 is not None) else np.zeros(self.action_dim)) def noise(self): x = ((self.x_prev + ((self.theta * (self.mean - self.x_prev)) * self.dt)) + ((self.std * np.sqrt(self.dt)) * np.random.normal(size=self.action_dim))) self.x_prev = x return x
def get_games_from_file(filename): pgn = open(filename, errors='ignore') offsets = [] while True: offset = pgn.tell() headers = chess.pgn.read_headers(pgn) if (headers is None): break offsets.append(offset) n = len(offsets) print(f'found {n} games') games = [] for offset in offsets: pgn.seek(offset) games.append(chess.pgn.read_game(pgn)) return games
class ThreeCarsHighSpeedCollision(Scenario): def init_scene(self, prefix, settings=None, spectator_tr=None): super().init_scene(prefix, settings, spectator_tr) blueprint_library = self.world.get_blueprint_library() vehicle00_tr = carla.Transform(carla.Location(110, (- 255), 0.05), carla.Rotation(yaw=180)) vehicle00 = self.world.spawn_actor(blueprint_library.filter('prius')[0], vehicle00_tr) vehicle01_tr = carla.Transform(carla.Location(53, (- 257), 0.0), carla.Rotation(yaw=0)) vehicle01 = self.world.spawn_actor(blueprint_library.filter('a2')[0], vehicle01_tr) vehicle02_tr = carla.Transform(carla.Location(85, (- 230), 0.04), carla.Rotation(yaw=(- 90))) vehicle02 = self.world.spawn_actor(blueprint_library.filter('lincoln')[0], vehicle02_tr) self.wait(1) vehicle00.set_target_velocity(carla.Vector3D((- 30), 0, 0)) vehicle01.set_target_velocity(carla.Vector3D((+ 30), 0, 0)) vehicle02.set_target_velocity(carla.Vector3D(0, (- 30), 0)) self.add_actor(vehicle00, 'Car') self.add_actor(vehicle01, 'Car') self.add_actor(vehicle02, 'Car') self.wait(1)
def main(): parser = ArgumentParser() parser.add_argument('pcd', help='Point cloud file') parser.add_argument('config', help='Config file') parser.add_argument('checkpoint', help='Checkpoint file') parser.add_argument('--device', default='cuda:0', help='Device used for inference') parser.add_argument('--score-thr', type=float, default=0.0, help='bbox score threshold') parser.add_argument('--out-dir', type=str, default='demo', help='dir to save results') parser.add_argument('--show', action='store_true', help='show online visuliaztion results') parser.add_argument('--snapshot', action='store_true', help='whether to save online visuliaztion results') args = parser.parse_args() model = init_model(args.config, args.checkpoint, device=args.device) (result, data) = inference_detector(model, args.pcd) show_result_meshlab(data, result, args.out_dir, args.score_thr, show=args.show, snapshot=args.snapshot, task='det')
class RNNLearnerState(NamedTuple): params: Params opt_states: OptStates key: chex.PRNGKey env_state: LogEnvState timestep: TimeStep dones: Done hstates: HiddenStates
def _ent_in_context_at_k(guess_item, gold_item, k): titles = eval_downstream.get_gold_titles(gold_item) if ('provenance' in guess_item['output'][0]): provenance = guess_item['output'][0]['provenance'] for i in range(0, min(k, len(provenance))): if ('text' in provenance[i]): normalized_text = eval_downstream.normalize_answer(provenance[i]['text']) for t in titles: if (eval_downstream.normalize_answer(t) in normalized_text): return 1 return 0
class AVATAR_PT_MotionPanel(bpy.types.Panel): bl_idname = 'AVATAR_PT_MotionPanel' bl_label = 'Motion' bl_space_type = 'VIEW_3D' bl_region_type = 'UI' bl_category = 'Avatar' bpy.types.Object.bvh_offset = IntProperty(name='Offset', description='Start motion offset', default=0, min=0, max=250) bpy.types.Object.bvh_start_origin = BoolProperty(name='Origin', description='Start at origin', default=False) def draw(self, context): layout = self.layout obj = context.object wm = context.window_manager layout.operator('avt.set_rest_pose', text='Reset pose') layout.prop(context.scene, 'skel_rig', text='') layout.operator('avt.load_bvh', text='Load BVH')
class StatefulContainer(object): def __init__(self): self._state = dict() self._factories = dict() def add_factory(self, name, factory: Callable[([], Any)]): self._factories[name] = factory def merge_state_dict(self, state_dict: Dict[(str, Any)]): self._state.update(state_dict) def state_dict(self) -> Dict[(str, Any)]: return self._state def __getattr__(self, name): if ((name not in self._state) and (name in self._factories)): self._state[name] = self._factories[name]() if (name in self._state): return self._state[name] raise AttributeError(f'Task state has no factory for attribute {name}')
class MultiLoop(object): no_resolve_ = (str, set) class _multi_loop_container(object): def __init__(self, item, level=0): self.item = item self.level = level def multi_loop(self, method, *args, **kwargs): kwargs = kwargs.copy() descend = kwargs.pop('loop_descend', 1) no_resolve = kwargs.pop('no_resolve', self.no_resolve_) no_resolve += kwargs.pop('no_resolve_add', tuple()) kwargs_new = kwargs.copy() args_new = list(args) if (descend <= 0): return [method(*args_new, **kwargs_new)] kwargs_new['loop_descend'] = descend kwargs_new['no_resolve'] = no_resolve result = [] for (iv, v) in enumerate(args): if isinstance(v, no_resolve): continue level = 1 if isinstance(v, self._multi_loop_container): if (v.level == descend): continue level = (v.level + 1) v = v.item if isinstance(v, dict): if (len(v) <= 1): continue for (k, i) in v.items(): args_new[iv] = {k: i} result += self.multi_loop(method, *args_new, **kwargs_new) return result if isinstance(v, Iterable): for i in v: args_new[iv] = self._multi_loop_container(i, level) result += self.multi_loop(method, *args_new, **kwargs_new) return result for (kw, v) in kwargs.items(): if isinstance(v, no_resolve): continue level = 1 if isinstance(v, self._multi_loop_container): if (v.level == descend): continue level = (v.level + 1) v = v.item if isinstance(v, dict): if (len(v) <= 1): continue for (k, i) in v.items(): kwargs_new[kw] = {k: i} result += self.multi_loop(method, *args_new, **kwargs_new) return result if isinstance(v, Iterable): for i in v: kwargs_new[kw] = self._multi_loop_container(i, level) result += self.multi_loop(method, *args_new, **kwargs_new) return result for (iv, v) in enumerate(args_new): if isinstance(v, self._multi_loop_container): args_new[iv] = v.item for (kw, v) in kwargs_new.items(): if isinstance(v, self._multi_loop_container): kwargs_new[kw] = v.item return [method(*args_new, **kwargs_new)] def clean(kwargs, extra=None): kw = kwargs.copy() if (extra is not None): if isinstance(extra, str): extra = (extra,) else: extra = tuple() extra += ('loop_descend', 'no_resolve', 'no_resolve_add') for x in extra: kw.pop(x, None) return kw
class LSTMState(object): def __init__(self, states): self.states = states def from_pytorch(cls, states): (hs, cs) = states (_, bs, d) = hs.shape hs = hs.view((- 1), 2, bs, d) cs = cs.view((- 1), 2, bs, d) nl = hs.shape[0] states = [(h.sum(dim=0), c.sum(dim=0)) for (h, c) in zip(hs.unbind(dim=0), cs.unbind(dim=0))] return LSTMState(states) def stack(cls, iterator_states, dim): nl = len(iterator_states[0]) hs = [list() for _ in range(nl)] cs = [list() for _ in range(nl)] for states in iterator_states: for (i, state) in enumerate(states.states): (h, c) = state hs[i].append(h) cs[i].append(c) states = list() for i in range(nl): h = torch.stack(hs[i], dim) c = torch.stack(cs[i], dim) states.append((h, c)) return LSTMState(states) def zero_state(cls, num_layers, shape): states = list() for _ in range(num_layers): h = get_zeros(*shape) c = get_zeros(*shape) states.append((h, c)) return LSTMState(states) def shape(self): return self.states[0][0].shape def clone(self): new_states = [(s[0].clone(), s[1].clone()) for s in self.states] return LSTMState(new_states) def dim(self): return self.states[0][0].dim() def unsqueeze(self, dim): new_states = [(s[0].unsqueeze(dim), s[1].unsqueeze(dim)) for s in self.states] return LSTMState(new_states) def view(self, *sizes): new_states = [(s[0].view(*sizes), s[1].view(*sizes)) for s in self.states] return LSTMState(new_states) def unbind(self, dim): n = self.states[0][0].shape[dim] ret = [list() for _ in range(n)] for s in self.states: (h, c) = s hs = h.unbind(dim) cs = c.unbind(dim) for (i, (h, c)) in enumerate(zip(hs, cs)): ret[i].append((h, c)) ret = tuple((LSTMState(s) for s in ret)) return ret def expand(self, *sizes): new_states = [(s[0].expand(*sizes), s[1].expand(*sizes)) for s in self.states] return LSTMState(new_states) def contiguous(self): states = list() for s in self.states: h = s[0].contiguous() c = s[1].contiguous() states.append((h, c)) return LSTMState(states) def __len__(self): return len(self.states) def detach_(self): for s in self.states: s[0].detach_() s[1].detach_() def size(self): return self.states[0][0].size() def cat(self, other, dim): states = list() for (s1, s2) in zip(self.states, other.states): h = torch.cat([s1[0], s2[0]], dim) c = torch.cat([s1[1], s2[1]], dim) states.append((h, c)) return LSTMState(states) def __mul__(self, other): states = list() for s in self.states: h = (s[0] * other) c = (s[1] * other) states.append((h, c)) return LSTMState(states) def __add__(self, other): states = list() for (s1, s2) in zip(self.states, other.states): h = (s1[0] + s2[0]) c = (s1[1] + s2[1]) states.append((h, c)) return LSTMState(states) def get_output(self): return self.states[(- 1)][0] def get(self, ind): return self.states[ind] def __getitem__(self, key): states = list() for s in self.states: h = s[0][key] c = s[1][key] states.append((h, c)) return LSTMState(states) def __setitem__(self, key, item): for (s1, s2) in zip(self.states, item.states): s1[0][key] = s2[0] s1[1][key] = s2[1]
class GatherOperation(Function): def forward(ctx, features: torch.Tensor, idx: torch.Tensor) -> torch.Tensor: assert features.is_contiguous() assert idx.is_contiguous() (B, npoint) = idx.size() (_, C, N) = features.size() output = torch.cuda.FloatTensor(B, C, npoint) pointnet2.gather_points_wrapper(B, C, N, npoint, features, idx, output) ctx.for_backwards = (idx, C, N) return output def backward(ctx, grad_out): (idx, C, N) = ctx.for_backwards (B, npoint) = idx.size() grad_features = Variable(torch.cuda.FloatTensor(B, C, N).zero_()) grad_out_data = grad_out.data.contiguous() pointnet2.gather_points_grad_wrapper(B, C, N, npoint, grad_out_data, idx, grad_features.data) return (grad_features, None)
class SegformerDropPath(nn.Module): def __init__(self, drop_prob: Optional[float]=None) -> None: super().__init__() self.drop_prob = drop_prob def forward(self, hidden_states: torch.Tensor) -> torch.Tensor: return drop_path(hidden_states, self.drop_prob, self.training) def extra_repr(self) -> str: return 'p={}'.format(self.drop_prob)
def _do_apply_self_mask(m): if (not g_options.use_self_mask): return m self_mask = _get_self_mask(m) return ((m * (1 - self_mask)) + ((- 10) * self_mask))
def test_summarize_weighted(model, X, w): d1 = model.distributions[0] d2 = model.distributions[1] model.summarize(X, sample_weight=w) assert_array_almost_equal(model._xw_sum, [0., 4.173105, 4.912965, 4.113657], 4) assert_array_almost_equal(model._xw_starts_sum, [0.136405, 3.163595], 4) assert_array_almost_equal(model._xw_ends_sum, [0.876271, 2.423729], 4) assert_array_almost_equal(d1._w_sum, [5.049643, 5.049643, 5.049643], 4) assert_array_almost_equal(d1._xw_sum, [8.834015, 5.17916, 0.], 4) assert_array_almost_equal(d2._w_sum, [11.450351, 11.450351, 11.450351], 4) assert_array_almost_equal(d2._xw_sum, [18.86598, 12.320832, 21.093086], 4)
def data_preparation(args): dataset_path = path.join('data', (('data_3d_' + args.dataset) + '.npz')) if (args.dataset == 'h36m'): from common.h36m_dataset import Human36mDataset, TEST_SUBJECTS dataset = Human36mDataset(dataset_path) if args.s1only: subjects_train = ['S1'] else: subjects_train = ['S1', 'S5', 'S6', 'S7', 'S8'] subjects_test = TEST_SUBJECTS else: raise KeyError('Invalid dataset') print('==> Loading 3D data...') dataset = read_3d_data(dataset) print('==> Loading 2D detections...') keypoints = create_2d_data(path.join('data', (((('data_2d_' + args.dataset) + '_') + args.keypoints) + '.npz')), dataset) action_filter = (None if (args.actions == '*') else args.actions.split(',')) if (action_filter is not None): action_filter = map((lambda x: dataset.define_actions(x)[0]), action_filter) print('==> Selected actions: {}'.format(action_filter)) stride = args.downsample (poses_train, poses_train_2d, actions_train, cams_train) = fetch(subjects_train, dataset, keypoints, action_filter, stride) (poses_valid, poses_valid_2d, actions_valid, cams_valid) = fetch(subjects_test, dataset, keypoints, action_filter, stride) train_det2d3d_loader = DataLoader(PoseDataSet(poses_train, poses_train_2d, actions_train, cams_train), batch_size=args.batch_size, shuffle=True, num_workers=args.num_workers, pin_memory=True) train_gt2d3d_loader = DataLoader(PoseDataSet(poses_train, poses_train_2d, actions_train, cams_train), batch_size=args.batch_size, shuffle=True, num_workers=args.num_workers, pin_memory=True) valid_loader = DataLoader(PoseDataSet(poses_valid, poses_valid_2d, actions_valid, cams_valid), batch_size=args.batch_size, shuffle=False, num_workers=args.num_workers, pin_memory=True) target_2d_loader = DataLoader(PoseTarget(poses_train_2d), batch_size=args.batch_size, shuffle=True, num_workers=args.num_workers, pin_memory=True) target_3d_loader = DataLoader(PoseTarget(poses_train), batch_size=args.batch_size, shuffle=True, num_workers=args.num_workers, pin_memory=True) mpi3d_npz = np.load('data_extra/test_set/test_3dhp.npz') tmp = mpi3d_npz mpi3d_loader = DataLoader(PoseBuffer([tmp['pose3d']], [tmp['pose2d']]), batch_size=args.batch_size, shuffle=False, num_workers=args.num_workers, pin_memory=True) return {'dataset': dataset, 'train_det2d3d_loader': train_det2d3d_loader, 'train_gt2d3d_loader': train_gt2d3d_loader, 'target_2d_loader': target_2d_loader, 'target_3d_loader': target_3d_loader, 'H36M_test': valid_loader, 'mpi3d_loader': mpi3d_loader, 'action_filter': action_filter, 'subjects_test': subjects_test, 'keypoints': keypoints}
class PointNet2FPModule(nn.Module): def __init__(self, *, mlp: List[int], bn: bool=True, use_paconv=False, args=None): super().__init__() self.use_paconv = use_paconv if self.use_paconv: self.mlp = paconv.SharedPAConv(mlp, bn=bn, config=args) else: self.mlp = block.SharedMLP(mlp, bn=bn) def forward(self, unknown: torch.Tensor, known: torch.Tensor, unknow_feats: torch.Tensor, known_feats: torch.Tensor) -> torch.Tensor: if (known is not None): (dist, idx) = pointops.nearestneighbor(unknown, known) dist_recip = (1.0 / (dist + 1e-08)) norm = torch.sum(dist_recip, dim=2, keepdim=True) weight = (dist_recip / norm) interpolated_feats = pointops.interpolation(known_feats, idx, weight) else: interpolated_feats = known_feats.expand(*known_feats.size()[0:2], unknown.size(1)) if (unknow_feats is not None): new_features = torch.cat([interpolated_feats, unknow_feats], dim=1) else: new_features = interpolated_feats return self.mlp(new_features.unsqueeze((- 1))).squeeze((- 1))
class Exp(MyExp): def __init__(self): super(Exp, self).__init__() self.num_classes = 1 self.depth = 0.33 self.width = 0.5 self.exp_name = os.path.split(os.path.realpath(__file__))[1].split('.')[0] self.train_ann = 'train.json' self.val_ann = 'train.json' self.input_size = (608, 1088) self.test_size = (608, 1088) self.random_size = (12, 26) self.max_epoch = 80 self.print_interval = 20 self.eval_interval = 5 self.test_conf = 0.001 self.nmsthre = 0.7 self.no_aug_epochs = 10 self.basic_lr_per_img = (0.001 / 64.0) self.warmup_epochs = 1 def get_data_loader(self, batch_size, is_distributed, no_aug=False): from yolox.data import MOTDataset, TrainTransform, YoloBatchSampler, DataLoader, InfiniteSampler, MosaicDetection dataset = MOTDataset(data_dir=os.path.join(get_yolox_datadir(), 'mix_det'), json_file=self.train_ann, name='', img_size=self.input_size, preproc=TrainTransform(rgb_means=(0.485, 0.456, 0.406), std=(0.229, 0.224, 0.225), max_labels=500)) dataset = MosaicDetection(dataset, mosaic=(not no_aug), img_size=self.input_size, preproc=TrainTransform(rgb_means=(0.485, 0.456, 0.406), std=(0.229, 0.224, 0.225), max_labels=1000), degrees=self.degrees, translate=self.translate, scale=self.scale, shear=self.shear, perspective=self.perspective, enable_mixup=self.enable_mixup) self.dataset = dataset if is_distributed: batch_size = (batch_size // dist.get_world_size()) sampler = InfiniteSampler(len(self.dataset), seed=(self.seed if self.seed else 0)) batch_sampler = YoloBatchSampler(sampler=sampler, batch_size=batch_size, drop_last=False, input_dimension=self.input_size, mosaic=(not no_aug)) dataloader_kwargs = {'num_workers': self.data_num_workers, 'pin_memory': True} dataloader_kwargs['batch_sampler'] = batch_sampler train_loader = DataLoader(self.dataset, **dataloader_kwargs) return train_loader def get_eval_loader(self, batch_size, is_distributed, testdev=False): from yolox.data import MOTDataset, ValTransform valdataset = MOTDataset(data_dir=os.path.join(get_yolox_datadir(), 'mot'), json_file=self.val_ann, img_size=self.test_size, name='train', preproc=ValTransform(rgb_means=(0.485, 0.456, 0.406), std=(0.229, 0.224, 0.225))) if is_distributed: batch_size = (batch_size // dist.get_world_size()) sampler = torch.utils.data.distributed.DistributedSampler(valdataset, shuffle=False) else: sampler = torch.utils.data.SequentialSampler(valdataset) dataloader_kwargs = {'num_workers': self.data_num_workers, 'pin_memory': True, 'sampler': sampler} dataloader_kwargs['batch_size'] = batch_size val_loader = torch.utils.data.DataLoader(valdataset, **dataloader_kwargs) return val_loader def get_evaluator(self, batch_size, is_distributed, testdev=False): from yolox.evaluators import COCOEvaluator val_loader = self.get_eval_loader(batch_size, is_distributed, testdev=testdev) evaluator = COCOEvaluator(dataloader=val_loader, img_size=self.test_size, confthre=self.test_conf, nmsthre=self.nmsthre, num_classes=self.num_classes, testdev=testdev) return evaluator
def load_file_list(path=None, regx='\\.npz', printable=True): if (path == False): path = os.getcwd() file_list = os.listdir(path) return_list = [] for (idx, f) in enumerate(file_list): if re.search(regx, f): return_list.append(f) if printable: print(('Match file list = %s' % return_list)) print(('Number of files = %d' % len(return_list))) return return_list
class ElementWiseUnaryOp(UnaryOpBase): def __init__(self): super().__init__() self.inp_ranks = [rank_all()] self.out_ranks = [rank_all()] def type_transfer(self, input_shapes: List[AbsTensor]) -> List[AbsTensor]: SanityCheck.eq(len(input_shapes), 1) return [input_shapes[0]] def deduct_inp_ranks_and_dtype(self, out_abs_tensor: List[AbsTensor]) -> List[Tuple[(int, DType)]]: return [(out_abs_tensor[0].ndims, out_abs_tensor[0].dtype)]
class CoarseAlign(): def __init__(self, nbScale, nbIter, tolerance, transform, minSize, segId, segFg, scaleR=2, imageNet=True, segNet=True): self.nbIter = nbIter self.tolerance = tolerance resnet_feature_layers = ['conv1', 'bn1', 'relu', 'maxpool', 'layer1', 'layer2', 'layer3'] if imageNet: resNetfeat = models.resnet50(pretrained=True) else: resNetfeat = resnet50() featPth = '../../model/pretrained/resnet50_moco.pth' param = torch.load(featPth) state_dict = {k.replace('module.', ''): v for (k, v) in param['model'].items()} msg = 'Loading pretrained model from {}'.format(featPth) print(msg) resNetfeat.load_state_dict(state_dict) resnet_module_list = [getattr(resNetfeat, l) for l in resnet_feature_layers] last_layer_idx = resnet_feature_layers.index('layer3') self.net = torch.nn.Sequential(*resnet_module_list[:(last_layer_idx + 1)]) self.net.cuda() self.net.eval() if segNet: self.segNet = segEval.SegNet('../../model/pretrained/ade20k_resnet50dilated_encoder.pth', '../../model/pretrained/ade20k_resnet50dilated_decoder.pth', segId, segFg) normalize = transforms.Normalize(mean=[0.485, 0.456, 0.406], std=[0.229, 0.224, 0.225]) self.toTensor = transforms.ToTensor() self.preproc = transforms.Compose([transforms.ToTensor(), normalize]) if (transform == 'Affine'): self.Transform = outil.Affine self.nbPoint = 3 else: self.Transform = outil.Homography self.nbPoint = 4 self.strideNet = 16 self.minSize = minSize if (nbScale == 1): self.scaleList = [1] else: self.scaleList = (np.linspace(scaleR, 1, ((nbScale // 2) + 1)).tolist() + np.linspace(1, (1 / scaleR), ((nbScale // 2) + 1)).tolist()[1:]) print(self.scaleList) torch.cuda.empty_cache() def ResizeMinSize(self, I, minSize): (w, h) = I.size ratio = min((w / float(minSize)), (h / float(minSize))) (new_w, new_h) = (int(round((w / ratio))), int(round((h / ratio)))) (new_w, new_h) = (((new_w // self.strideNet) * self.strideNet), ((new_h // self.strideNet) * self.strideNet)) (ratioW, ratioH) = ((new_w / float(w)), (new_h / float(h))) Iresize = I.resize((new_w, new_h), resample=Image.LANCZOS) return Iresize def setPair(self, Is_org, It_org): with torch.no_grad(): IsList = [] for i in range(len(self.scaleList)): IsList.append(self.ResizeMinSize(Is_org, int((self.minSize * self.scaleList[i])))) self.Is = IsList[(len(self.scaleList) // 2)] self.IsTensor = self.toTensor(self.Is).unsqueeze(0).cuda() featsMultiScale = [] WMultiScale = [] HMultiScale = [] for i in range(len(self.scaleList)): feat = F.normalize(self.net(self.preproc(IsList[i]).unsqueeze(0).cuda())) (Ws, Hs) = outil.getWHTensor(feat) featsMultiScale.append(feat.contiguous().view(1024, (- 1))) WMultiScale.append(Ws) HMultiScale.append(Hs) torch.cuda.empty_cache() featsMultiScale = torch.cat(featsMultiScale, dim=1) WMultiScale = torch.cat(WMultiScale) HMultiScale = torch.cat(HMultiScale) torch.cuda.empty_cache() self.It = self.ResizeMinSize(It_org, self.minSize) self.ItTensor = self.toTensor(self.It).unsqueeze(0).cuda() featt = F.normalize(self.net(self.preproc(self.It).unsqueeze(0).cuda())) (Wt, Ht) = outil.getWHTensor(featt) (WtInt, HtInt) = outil.getWHTensor_Int(featt) (self.W2, self.H2) = (featt.size()[2], featt.size()[3]) featt = featt.contiguous().view(1024, (- 1)) (index1, index2) = outil.mutualMatching(featsMultiScale, featt) self.W1MutualMatch = WMultiScale[index1] self.H1MutualMatch = HMultiScale[index1] self.W2MutualMatch = Wt[index2] self.H2MutualMatch = Ht[index2] self.W2MutualMatchInt = WtInt[index2] self.H2MutualMatchInt = HtInt[index2] def skyFromSeg(self, path): return self.segNet.getSky(path) def getCoarse(self, Mt): with torch.no_grad(): MtExtend = (1 - Mt).astype(np.float32) MtExtend = torch.from_numpy(MtExtend).cuda().unsqueeze(0).unsqueeze(0) MtTensor = F.interpolate(input=MtExtend, size=(self.W2, self.H2), mode='bilinear') MtTensor = (MtTensor > 0.5).squeeze() validMutualMatch = MtTensor[(self.W2MutualMatchInt, self.H2MutualMatchInt)] ones = torch.cuda.FloatTensor(self.W1MutualMatch[validMutualMatch].size(0)).fill_(1) match1 = torch.cat((self.H1MutualMatch[validMutualMatch].unsqueeze(1), self.W1MutualMatch[validMutualMatch].unsqueeze(1), ones.unsqueeze(1)), dim=1) match2 = torch.cat((self.H2MutualMatch[validMutualMatch].unsqueeze(1), self.W2MutualMatch[validMutualMatch].unsqueeze(1), ones.unsqueeze(1)), dim=1) if (len(match1) < self.nbPoint): return None (bestParam, _, indexInlier, _) = outil.RANSAC(self.nbIter, match1, match2, self.tolerance, self.nbPoint, self.Transform) if (bestParam is None): return None else: return bestParam.astype(np.float32)
def visual_image_MT(vis, mask_train, pred_train1, mask_val, pred_val1, pred_val2): vis.heatmap(mask_train, win='train_mask', opts=dict(title='Train Mask', colormap='Viridis')) vis.heatmap(pred_train1, win='train_pred1', opts=dict(title='Train Pred', colormap='Viridis')) vis.heatmap(mask_val, win='val_mask', opts=dict(title='Val Mask', colormap='Viridis')) vis.heatmap(pred_val1, win='val_pred1', opts=dict(title='Val Pred1', colormap='Viridis')) vis.heatmap(pred_val2, win='val_pred2', opts=dict(title='Val Pred2', colormap='Viridis'))
def evaluate(y_true, y_pred_proba, debug=False): max_threshold = (- 1) max_f1 = 0 max_recall = 0 max_precision = 0 mac_acc = 0 for THRESHOLD in range(50, 51): THRESHOLD = (THRESHOLD / 100) y_pred_thr = [(1 if (x >= THRESHOLD) else 0) for x in y_pred_proba] f1 = f1_score(y_true, y_pred_thr, average='macro') recall = recall_score(y_true, y_pred_thr) precision = precision_score(y_true, y_pred_thr) acc = accuracy_score(y_true, y_pred_thr) if debug: print('THRESHOLD: {:.3f} \tF1: {:.8f} \tRecall: {:.8f} \tPrecision: {:.8f}'.format(THRESHOLD, f1, recall, precision)) if (f1 > max_f1): max_f1 = f1 max_recall = recall max_precision = precision mac_acc = acc max_threshold = THRESHOLD print('##MAX## \nTHRESHOLD: {:.3f} \tF1: {:.8f} \tRecall: {:.8f} \tPrec: {:.8f} \tAcc: {:.8f}'.format(max_threshold, max_f1, max_recall, max_precision, mac_acc)) return (max_f1, max_recall, max_precision, mac_acc)
class ResUnetSkipConnectionBlock(nn.Module): def __init__(self, outer_nc, inner_nc, input_nc=None, submodule=None, outermost=False, innermost=False, norm_layer=nn.BatchNorm2d, use_dropout=False): super(ResUnetSkipConnectionBlock, self).__init__() self.outermost = outermost use_bias = (norm_layer == nn.InstanceNorm2d) if (input_nc is None): input_nc = outer_nc downconv = nn.Conv2d(input_nc, inner_nc, kernel_size=3, stride=2, padding=1, bias=use_bias) res_downconv = [ResidualBlock(inner_nc, norm_layer), ResidualBlock(inner_nc, norm_layer)] res_upconv = [ResidualBlock(outer_nc, norm_layer), ResidualBlock(outer_nc, norm_layer)] downrelu = nn.ReLU(True) uprelu = nn.ReLU(True) if (norm_layer != None): downnorm = norm_layer(inner_nc) upnorm = norm_layer(outer_nc) if outermost: upsample = nn.Upsample(scale_factor=2, mode='nearest') upconv = nn.Conv2d((inner_nc * 2), outer_nc, kernel_size=3, stride=1, padding=1, bias=use_bias) down = ([downconv, downrelu] + res_downconv) up = [upsample, upconv] model = ((down + [submodule]) + up) elif innermost: upsample = nn.Upsample(scale_factor=2, mode='nearest') upconv = nn.Conv2d(inner_nc, outer_nc, kernel_size=3, stride=1, padding=1, bias=use_bias) down = ([downconv, downrelu] + res_downconv) if (norm_layer == None): up = ([upsample, upconv, uprelu] + res_upconv) else: up = ([upsample, upconv, upnorm, uprelu] + res_upconv) model = (down + up) else: upsample = nn.Upsample(scale_factor=2, mode='nearest') upconv = nn.Conv2d((inner_nc * 2), outer_nc, kernel_size=3, stride=1, padding=1, bias=use_bias) if (norm_layer == None): down = ([downconv, downrelu] + res_downconv) up = ([upsample, upconv, uprelu] + res_upconv) else: down = ([downconv, downnorm, downrelu] + res_downconv) up = ([upsample, upconv, upnorm, uprelu] + res_upconv) if use_dropout: model = (((down + [submodule]) + up) + [nn.Dropout(0.5)]) else: model = ((down + [submodule]) + up) self.model = nn.Sequential(*model) def forward(self, x): if self.outermost: return self.model(x) else: return torch.cat([x, self.model(x)], 1)
def match_api(A: str, B: str, num=50, equal_type=1, fuzz=True, neq_dir='output/neq', fail_dir='output/fail', success_dir='output/success', err_dir='output/err', bug_dir='output/bug'): def verify_wrapper(api_A, api_B, indices, neq_dir, fail_dir, success_dir, err_dir, bug_dir, index=None, value=None): for _ in range(10): res = verify(api_A, api_B, num, equal_type, fuzz, indices, neq_dir, fail_dir, success_dir, err_dir, bug_dir, strict_mode=True) if (ResultType.NOT_EQUIVALENT in res): return False elif (ResultType.SUCCESS in res): return True return False os.makedirs(neq_dir, exist_ok=True) os.makedirs(fail_dir, exist_ok=True) os.makedirs(success_dir, exist_ok=True) os.makedirs(err_dir, exist_ok=True) os.makedirs(bug_dir, exist_ok=True) api_A = TorchAPI(A) api_B = TorchAPI(B) results = [] if ((len(api_A.arg_defs) == 0) or (len(api_B.arg_defs) == 0) or (api_A.is_class != api_B.is_class)): return (None, []) indices = match_argument(api_A.arg_defs, api_B.arg_defs) source_matched_indices = [p[0] for p in indices] if (len(source_matched_indices) < len(api_A.arg_defs)): for i in range(len(api_A.arg_defs)): if (i not in source_matched_indices): if api_A.arg_defs[i].is_optional: api_A.arg_defs[i].ignore = True else: return (None, []) target_matched_indices = [p[1] for p in indices] if (len(target_matched_indices) < len(api_B.arg_defs)): for i in range(len(api_B.arg_defs)): if ((i not in target_matched_indices) and (not api_B.arg_defs[i].is_optional)): return (None, []) if fuzz: results = verify(api_A, api_B, num, equal_type, fuzz, indices, neq_dir, fail_dir, success_dir, err_dir, bug_dir) else: results = verify_wrapper(api_A, api_B, indices, neq_dir, fail_dir, success_dir, err_dir, bug_dir) return (results, indices)
def _extract_archive(file_path, path='.', archive_format='auto'): if (archive_format is None): return False if (archive_format == 'auto'): archive_format = ['tar', 'zip'] if isinstance(archive_format, six.string_types): archive_format = [archive_format] for archive_type in archive_format: if (archive_type == 'tar'): open_fn = tarfile.open is_match_fn = tarfile.is_tarfile elif (archive_type == 'zip'): open_fn = zipfile.ZipFile is_match_fn = zipfile.is_zipfile else: raise Exception('Invalid archive type.') if is_match_fn(file_path): with open_fn(file_path) as archive: try: archive.extractall(path) except (tarfile.TarError, RuntimeError, KeyboardInterrupt): if os.path.exists(path): if os.path.isfile(path): os.remove(path) else: shutil.rmtree(path) raise return True return False
class Blip2Model(nn.Module): def __init__(self, args: Namespace): super(Blip2Model, self).__init__() self.args = args (self.model, self.vis_processors, _) = load_model_and_preprocess(name='blip2_t5', model_type='pretrain_flant5xxl', is_eval=True, device=args.device) def forward(self, q: str, image_input) -> list: question = f'Question: {q} Answer:' images = [self.vis_processors['eval'](im) for im in image_input] images = torch.stack(images).to(self.args.device) answer = self.model.generate({'image': images, 'prompt': ([question] * len(images))}) return [a.lower() for a in answer]
class ContinuousMLPPolicy(Policy): def __init__(self, env_spec, name='ContinuousMLPPolicy', hidden_sizes=(64, 64), hidden_nonlinearity=tf.nn.relu, hidden_w_init=tf.initializers.glorot_uniform(seed=deterministic.get_tf_seed_stream()), hidden_b_init=tf.zeros_initializer(), output_nonlinearity=tf.nn.tanh, output_w_init=tf.initializers.glorot_uniform(seed=deterministic.get_tf_seed_stream()), output_b_init=tf.zeros_initializer(), layer_normalization=False): super().__init__(name, env_spec) action_dim = env_spec.action_space.flat_dim self._hidden_sizes = hidden_sizes self._hidden_nonlinearity = hidden_nonlinearity self._hidden_w_init = hidden_w_init self._hidden_b_init = hidden_b_init self._output_nonlinearity = output_nonlinearity self._output_w_init = output_w_init self._output_b_init = output_b_init self._layer_normalization = layer_normalization self._obs_dim = env_spec.observation_space.flat_dim self.model = MLPModel(output_dim=action_dim, name='MLPModel', hidden_sizes=hidden_sizes, hidden_nonlinearity=hidden_nonlinearity, hidden_w_init=hidden_w_init, hidden_b_init=hidden_b_init, output_nonlinearity=output_nonlinearity, output_w_init=output_w_init, output_b_init=output_b_init, layer_normalization=layer_normalization) self._initialize() def _initialize(self): state_input = tf.compat.v1.placeholder(tf.float32, shape=(None, self._obs_dim)) with tf.compat.v1.variable_scope(self.name) as vs: self._variable_scope = vs outputs = self.model.build(state_input).outputs self._f_prob = tf.compat.v1.get_default_session().make_callable(outputs, feed_list=[state_input]) def input_dim(self): return self._obs_dim def get_action_sym(self, obs_var, name=None): with tf.compat.v1.variable_scope(self._variable_scope): return self.model.build(obs_var, name=name).outputs def get_action(self, observation): (actions, agent_infos) = self.get_actions([observation]) action = actions[0] return (action, {k: v[0] for (k, v) in agent_infos.items()}) def get_actions(self, observations): observations = self.observation_space.flatten_n(observations) actions = self._f_prob(observations) actions = self.action_space.unflatten_n(actions) return (actions, dict()) def get_regularizable_vars(self): trainable = self.get_trainable_vars() return [var for var in trainable if (('hidden' in var.name) and ('kernel' in var.name))] def vectorized(self): return True def clone(self, name): new_policy = self.__class__(name=name, env_spec=self._env_spec, hidden_sizes=self._hidden_sizes, hidden_nonlinearity=self._hidden_nonlinearity, hidden_w_init=self._hidden_w_init, hidden_b_init=self._hidden_b_init, output_nonlinearity=self._output_nonlinearity, output_w_init=self._output_w_init, output_b_init=self._output_b_init, layer_normalization=self._layer_normalization) new_policy.model.parameters = self.model.parameters return new_policy def __getstate__(self): new_dict = super().__getstate__() del new_dict['_f_prob'] return new_dict def __setstate__(self, state): super().__setstate__(state) self._initialize()
class DatasetLossGame(StochasticCooperativeGame): def __init__(self, extension, data, labels, loss, groups=None): self.extension = extension self.loss = loss self.N = len(data) assert (len(labels) == self.N) self.exogenous = (data, labels) num_features = data.shape[1] if (groups is None): self.players = num_features self.groups_matrix = None else: inds_list = [] for group in groups: inds_list += list(group) assert np.all((np.sort(inds_list) == np.arange(num_features))) self.players = len(groups) self.groups_matrix = np.zeros((len(groups), num_features), dtype=bool) for (i, group) in enumerate(groups): self.groups_matrix[(i, group)] = True def __call__(self, S, U): if (U is None): U = self.sample(len(S)) (x, y) = U if (self.loss is utils.crossentropyloss): if ((y.ndim == 1) or (y.shape[1] == 1)): if np.issubdtype(y.dtype, np.floating): y = y.astype(int) if (self.groups_matrix is not None): S = np.matmul(S, self.groups_matrix) return (- self.loss(self.extension(x, S), y))
def mixres50_w234a234(**kwargs): return ResNet(Bottleneck, qm.MixActivConv2d, [3, 4, 6, 3], wbits=[2, 3, 4], abits=[2, 3, 4], share_weight=True, **kwargs)
def _get_local_ip(): try: s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM) s.connect(('8.8.8.8', 80)) ip = s.getsockname()[0] finally: s.close() return ip
class GradientReceiver(Receiver): def receive_notify(self, solver: 'Solver', message): if (not (Signal.TRAIN_PIPE_END in message)): return for netnode in solver.netnodes: if (not netnode.require_no_grad): model = netnode.net total_norm = 0 for p in model.parameters(): param_norm = p.grad.data.norm(2) total_norm += (param_norm.item() ** 2) total_norm = (total_norm ** (1.0 / 2)) assert isinstance(solver.receivers[0], SummaryWriter) solver.summary_receiver.add_scalar('gradient/total_norm', total_norm, solver.global_step)
def _test_update(inertia): x1 = torch.tensor([1.0, 1.4, 1.8, (- 1.1), 0.0]) x2 = torch.tensor([2.2, 8.2, 0.1, 105.2, 0.0]) y = ((x1 * inertia) + (x2 * (1 - inertia))) _update_parameter(x1, x2, inertia=inertia) assert_array_almost_equal(x1, y)
def train(args, train_dataset, model, tokenizer, train_highway=False): if (args.local_rank in [(- 1), 0]): tb_writer = SummaryWriter() args.train_batch_size = (args.per_gpu_train_batch_size * max(1, args.n_gpu)) train_sampler = (RandomSampler(train_dataset) if (args.local_rank == (- 1)) else DistributedSampler(train_dataset)) train_dataloader = DataLoader(train_dataset, sampler=train_sampler, batch_size=args.train_batch_size) if (args.max_steps > 0): t_total = args.max_steps args.num_train_epochs = ((args.max_steps // (len(train_dataloader) // args.gradient_accumulation_steps)) + 1) else: t_total = ((len(train_dataloader) // args.gradient_accumulation_steps) * args.num_train_epochs) no_decay = ['bias', 'LayerNorm.weight'] if train_highway: optimizer_grouped_parameters = [{'params': [p for (n, p) in model.named_parameters() if (('highway' in n) and (not any(((nd in n) for nd in no_decay))))], 'weight_decay': args.weight_decay}, {'params': [p for (n, p) in model.named_parameters() if (('highway' in n) and any(((nd in n) for nd in no_decay)))], 'weight_decay': 0.0}] else: optimizer_grouped_parameters = [{'params': [p for (n, p) in model.named_parameters() if (('highway' not in n) and (not any(((nd in n) for nd in no_decay))))], 'weight_decay': args.weight_decay}, {'params': [p for (n, p) in model.named_parameters() if (('highway' not in n) and any(((nd in n) for nd in no_decay)))], 'weight_decay': 0.0}] optimizer = AdamW(optimizer_grouped_parameters, lr=args.learning_rate, eps=args.adam_epsilon) scheduler = get_linear_schedule_with_warmup(optimizer, num_warmup_steps=args.warmup_steps, num_training_steps=t_total) if args.fp16: try: from apex import amp except ImportError: raise ImportError('Please install apex from to use fp16 training.') (model, optimizer) = amp.initialize(model, optimizer, opt_level=args.fp16_opt_level) if (args.n_gpu > 1): model = torch.nn.DataParallel(model) if (args.local_rank != (- 1)): model = torch.nn.parallel.DistributedDataParallel(model, device_ids=[args.local_rank], output_device=args.local_rank, find_unused_parameters=True) logger.info('***** Running training *****') logger.info(' Num examples = %d', len(train_dataset)) logger.info(' Num Epochs = %d', args.num_train_epochs) logger.info(' Instantaneous batch size per GPU = %d', args.per_gpu_train_batch_size) logger.info(' Total train batch size (w. parallel, distributed & accumulation) = %d', ((args.train_batch_size * args.gradient_accumulation_steps) * (torch.distributed.get_world_size() if (args.local_rank != (- 1)) else 1))) logger.info(' Gradient Accumulation steps = %d', args.gradient_accumulation_steps) logger.info(' Total optimization steps = %d', t_total) global_step = 0 (tr_loss, logging_loss) = (0.0, 0.0) model.zero_grad() train_iterator = trange(int(args.num_train_epochs), desc='Epoch', disable=(args.local_rank not in [(- 1), 0])) set_seed(args) for _ in train_iterator: epoch_iterator = tqdm(train_dataloader, desc='Iteration', disable=(args.local_rank not in [(- 1), 0])) for (step, batch) in enumerate(epoch_iterator): model.train() batch = tuple((t.to(args.device) for t in batch)) inputs = {'input_ids': batch[0], 'attention_mask': batch[1], 'labels': batch[3]} if (args.model_type != 'distilbert'): inputs['token_type_ids'] = (batch[2] if (args.model_type in ['bert', 'xlnet']) else None) inputs['train_highway'] = train_highway outputs = model(**inputs) loss = outputs[0] if (args.n_gpu > 1): loss = loss.mean() if (args.gradient_accumulation_steps > 1): loss = (loss / args.gradient_accumulation_steps) if args.fp16: with amp.scale_loss(loss, optimizer) as scaled_loss: scaled_loss.backward() else: loss.backward() tr_loss += loss.item() if (((step + 1) % args.gradient_accumulation_steps) == 0): if args.fp16: torch.nn.utils.clip_grad_norm_(amp.master_params(optimizer), args.max_grad_norm) else: torch.nn.utils.clip_grad_norm_(model.parameters(), args.max_grad_norm) optimizer.step() scheduler.step() model.zero_grad() global_step += 1 if ((args.local_rank in [(- 1), 0]) and (args.logging_steps > 0) and ((global_step % args.logging_steps) == 0)): if ((args.local_rank == (- 1)) and args.evaluate_during_training): results = evaluate(args, model, tokenizer) for (key, value) in results.items(): tb_writer.add_scalar('eval_{}'.format(key), value, global_step) tb_writer.add_scalar('lr', scheduler.get_lr()[0], global_step) tb_writer.add_scalar('loss', ((tr_loss - logging_loss) / args.logging_steps), global_step) logging_loss = tr_loss if ((args.local_rank in [(- 1), 0]) and (args.save_steps > 0) and ((global_step % args.save_steps) == 0)): output_dir = os.path.join(args.output_dir, 'checkpoint-{}'.format(global_step)) if (not os.path.exists(output_dir)): os.makedirs(output_dir) model_to_save = (model.module if hasattr(model, 'module') else model) model_to_save.save_pretrained(output_dir) torch.save(args, os.path.join(output_dir, 'training_args.bin')) logger.info('Saving model checkpoint to %s', output_dir) if ((args.max_steps > 0) and (global_step > args.max_steps)): epoch_iterator.close() break if ((args.max_steps > 0) and (global_step > args.max_steps)): train_iterator.close() break if (args.local_rank in [(- 1), 0]): tb_writer.close() return (global_step, (tr_loss / global_step))
def custom_loss(y_pred, y_true, model_name): if ('stochastic' in model_name): return KL_loss(y_pred, ((Coeff_Energy * y_true[0]) + (Coeff_Latency * y_true[1]))) return torch.sum(((y_pred - y_true) ** 2))
class Statement(SliceableMixin): _TEXT_REPR_MAX_LENGTH: int = 70 _stmts_by_ts: Dict[(Timestamp, List['Statement'])] = {} _stmts_by_id: Dict[(IdType, List['Statement'])] = {} def __init__(self, stmt_node: ast.stmt, frame: Optional[FrameType]=None, timestamp: Optional[Timestamp]=None, prev_stmt: Optional['Statement']=None, override: bool=False) -> None: self.stmt_node: ast.stmt = stmt_node self.frame: Optional[FrameType] = frame self._timestamp = (timestamp or Timestamp.current()) self._finished: bool = False self.override: bool = override self.prev_stmt = prev_stmt self.class_scope: Optional[Namespace] = None self.lambda_call_point_deps_done_once = False self.node_id_for_last_call: Optional[int] = None self._stmt_contains_cascading_reactive_rval: Optional[bool] = None self.dynamic_parents: Dict[(IdType, Set[Symbol])] = {} self.dynamic_children: Dict[(IdType, Set[Symbol])] = {} self.static_parents: Dict[(IdType, Set[Symbol])] = {} self.static_children: Dict[(IdType, Set[Symbol])] = {} def current(cls) -> 'Statement': return cls.at_timestamp(Timestamp.current()) def id(self) -> IdType: return self.stmt_id def timestamp(self) -> Timestamp: return self._timestamp def prev(self) -> Optional['Statement']: return self.prev_stmt def text(self) -> str: if (isinstance(self.stmt_node, ast.Assign) and (self.stmt_node.lineno == max((getattr(nd, 'lineno', self.stmt_node.lineno) for nd in ast.walk(self.stmt_node))))): components = [] for node in (self.stmt_node.targets + [self.stmt_node.value]): components.append(astunparse.unparse(node).strip()) components[(- 1)] = self._strip_tuple_parens(node, components[(- 1)]) return ' = '.join(components).strip() else: return astunparse.unparse(self.stmt_node).strip() def _strip_tuple_parens(node: ast.AST, text: str) -> str: if (isinstance(node, (ast.BinOp, ast.Tuple)) and (len(text) >= 2) and (text[0] == '(') and (text[(- 1)] == ')')): return text[1:(- 1)] else: return text def create_and_track(cls, stmt_node: ast.stmt, frame: Optional[FrameType]=None, timestamp: Optional[Timestamp]=None, override: bool=False) -> 'Statement': stmt_id = id(stmt_node) prev_stmt = (cls.from_id(stmt_id) if cls.has_id(stmt_id) else None) stmt = cls(stmt_node, frame=frame, timestamp=timestamp, prev_stmt=prev_stmt, override=override) if (override and cls._stmts_by_ts.get(timestamp)): prev = cls.at_timestamp(timestamp) all_with_prev_id = cls._stmts_by_id.pop(prev.id) assert (len(all_with_prev_id) == 1) assert (not cls.has_id(stmt.stmt_id)) cls._stmts_by_ts[stmt.timestamp] = [stmt] cls._stmts_by_id[stmt.stmt_id] = [stmt] for _ in SlicingContext.iter_slicing_contexts(): for cid in list(prev.children.keys()): cls.from_id(cid).replace_parent_edges(prev, stmt) for pid in list(prev.parents.keys()): cls.from_id(pid).replace_child_edges(prev, stmt) else: cls._stmts_by_ts.setdefault(stmt.timestamp, []).append(stmt) cls._stmts_by_id.setdefault(stmt.stmt_id, []).append(stmt) with static_slicing_context(): for (parent, syms) in flow().stmt_deferred_static_parents.get(stmt.timestamp, {}).items(): stmt.add_parent_edges(parent, syms) flow().stmt_deferred_static_parents.pop(stmt.timestamp, None) return stmt def is_module_stmt(self) -> bool: return (tracer().parent_stmt_by_id.get(self.stmt_id) is None) def is_outer_stmt(self) -> bool: return pyc.is_outer_stmt(self.stmt_id) def is_initial_frame_stmt(self) -> bool: return tracer().is_initial_frame_stmt(self.stmt_id) def clear(cls): cls._stmts_by_ts = {} def at_timestamp(cls, ts: TimestampOrCounter, stmt_num: Optional[int]=None) -> 'Statement': assert (isinstance(ts, Timestamp) or (stmt_num is not None)) if isinstance(ts, Timestamp): ts_to_use = ts else: ts_to_use = Timestamp(ts, stmt_num) return cls._stmts_by_ts[ts_to_use][0] def from_id(cls, stmt_id: IdType) -> 'Statement': return cls._stmts_by_id[stmt_id][0] def from_id_nullable(cls, stmt_id: IdType) -> Optional['Statement']: return cls._stmts_by_id.get(stmt_id, [None])[0] def has_id(cls, stmt_id: IdType) -> bool: return (len(cls._stmts_by_id.get(stmt_id, [])) > 0) def all_at_timestamp(cls, ts: Timestamp) -> List['Statement']: return cls._stmts_by_ts.get(ts, []) def containing_cell(self) -> 'Cell': return cells().at_timestamp(self.timestamp) def lineno(self) -> int: return self.stmt_node.lineno def finished(self) -> bool: return self._finished def stmt_id(self) -> int: return id(self.stmt_node) def __str__(self): return self.text def __repr__(self): return f'<{self.__class__.__name__}[ts={self.timestamp},text={repr(self.text[:self._TEXT_REPR_MAX_LENGTH])}]>' def slice(self, blacken: bool=True, seed_only: bool=False, format_type: Optional[Type[FormatType]]=None) -> Slice: return self.format_slice(blacken=blacken, seed_only=seed_only, format_type=format_type) def code(self, blacken: bool=True, format_type: Optional[Type[FormatType]]=None) -> Slice: return self.format_slice(blacken=blacken, seed_only=True, format_type=format_type) def to_function(self, *args, **kwargs): return self.code().to_function(*args, **kwargs) def stmt_contains_cascading_reactive_rval(self) -> bool: if (self._stmt_contains_cascading_reactive_rval is None): self._stmt_contains_cascading_reactive_rval = stmt_contains_cascading_reactive_rval(self.stmt_node) return self._stmt_contains_cascading_reactive_rval def _contains_lval(self) -> bool: return stmt_contains_lval(self.stmt_node) def get_post_call_scope(self, call_frame: FrameType) -> Scope: old_scope = tracer().cur_frame_original_scope if isinstance(self.stmt_node, ast.ClassDef): pending_ns = Namespace.make_child_namespace(old_scope, self.stmt_node.name) tracer().pending_class_namespaces.append(pending_ns) return pending_ns if isinstance(self.stmt_node, (ast.FunctionDef, ast.AsyncFunctionDef)): func_name = self.stmt_node.name else: func_name = None func_sym = tracer().calling_symbol if ((func_sym is None) or (func_sym.call_scope is None)): func_sym = flow().statement_to_func_sym.get(id(self.stmt_node), None) if (func_sym is None): return old_scope if (func_sym.call_scope is None): msg = ('got non-function symbol %s for name %s' % (func_sym.full_path, func_name)) if flow().is_dev_mode: raise TypeError(msg) else: logger.warning(msg) return old_scope if (not self.finished): prev_call_scope = func_sym.call_scope new_call_scope = prev_call_scope.parent_scope.make_child_scope(func_sym.name) if (prev_call_scope.symtab is not None): new_call_scope.symtab = prev_call_scope.symtab func_sym.call_scope = new_call_scope func_sym.create_symbols_for_call_args(call_frame) return func_sym.call_scope def _handle_reactive_store(target: ast.AST) -> None: try: symbol_ref = SymbolRef(target) reactive_seen = False blocking_seen = False for resolved in symbol_ref.gen_resolved_symbols(tracer().cur_frame_original_scope, only_yield_final_symbol=False, yield_all_intermediate_symbols=True, inherit_reactivity=False, yield_in_reverse=True): if resolved.is_blocking: blocking_seen = True if (resolved.is_reactive and (not blocking_seen)): if resolved.is_cascading_reactive: flow().updated_deep_reactive_symbols.add(resolved.sym) else: flow().updated_deep_reactive_symbols_last_cell.add(resolved.sym) reactive_seen = True if ((not resolved.is_live) and resolved.atom.is_cascading_reactive): resolved.sym.bump_cascading_reactive_cell_num() if resolved.is_last: resolved.sym.refresh() if (reactive_seen and (not blocking_seen)): if resolved.is_cascading_reactive: flow().updated_reactive_symbols.add(resolved.sym) else: flow().updated_reactive_symbols_last_cell.add(resolved.sym) if (blocking_seen and (resolved.sym not in flow().updated_symbols)): flow().blocked_reactive_timestamps_by_symbol[resolved.sym] = flow().cell_counter() except TypeError: return def _handle_assign_target_for_deps(self, target: ast.AST, deps: Set[Symbol], maybe_fixup_literal_namespace: bool=False) -> None: try: (scope, name, obj, is_subscript, excluded_deps) = tracer().resolve_store_data_for_target(target, self.frame) except KeyError: if flow().is_dev_mode: logger.warning('keyerror for %s', (astunparse.unparse(target) if isinstance(target, ast.AST) else target)) return if (isinstance(target, ast.Name) and getattr(obj, '__name__', '').startswith(Slice.FUNC_PREFIX)): obj.__name__ = target.id subscript_vals_to_use = [is_subscript] if scope.is_namespace_scope: namespace = cast(Namespace, scope) for (modname, classname) in DUPED_ATTRSUB_CLASSES: module = sys.modules.get(modname) if (module is None): continue clazz = getattr(module, classname, None) if (clazz is None): continue if (isinstance(namespace.obj, clazz) and (name in namespace.obj.columns)): subscript_vals_to_use.append((not is_subscript)) break for subscript_val in subscript_vals_to_use: upserted = scope.upsert_symbol_for_name(name, obj, (deps - excluded_deps), self.stmt_node, is_subscript=subscript_val, symbol_node=target, is_cascading_reactive=self.stmt_contains_cascading_reactive_rval) logger.info('sym %s upserted to scope %s has parents %s', upserted, scope, upserted.parents) self._handle_reactive_store(target) if maybe_fixup_literal_namespace: namespace_for_upsert = flow().namespaces.get(id(obj), None) if ((namespace_for_upsert is not None) and namespace_for_upsert.is_anonymous): namespace_for_upsert.scope_name = str(name) namespace_for_upsert.parent_scope = scope def _handle_store_target_tuple_unpack_from_deps(self, target: Union[(ast.List, ast.Tuple)], deps: Set[Symbol]) -> None: for inner_target in target.elts: if isinstance(inner_target, (ast.List, ast.Tuple)): self._handle_store_target_tuple_unpack_from_deps(inner_target, deps) else: self._handle_assign_target_for_deps(inner_target, deps) def _handle_starred_store_target(self, target: ast.Starred, inner_deps: List[Optional[Symbol]]) -> None: try: (scope, name, obj, is_subscript, _) = tracer().resolve_store_data_for_target(target, self.frame) except KeyError as e: logger.info('Exception: %s', e) return ns = flow().namespaces.get(id(obj), None) if (ns is None): ns = Namespace(obj, str(name), parent_scope=scope) for (i, inner_dep) in enumerate(inner_deps): deps = (set() if (inner_dep is None) else {inner_dep}) ns.upsert_symbol_for_name(i, inner_dep.obj, deps, self.stmt_node, is_subscript=True, is_cascading_reactive=self.stmt_contains_cascading_reactive_rval) scope.upsert_symbol_for_name(name, obj, set(), self.stmt_node, is_subscript=is_subscript, symbol_node=target, is_cascading_reactive=self.stmt_contains_cascading_reactive_rval) self._handle_reactive_store(target.value) def _handle_store_target_tuple_unpack_from_namespace(self, target: Union[(ast.List, ast.Tuple)], rhs_namespace: Namespace, extra_deps: Set[Symbol]) -> None: saved_starred_node: Optional[ast.Starred] = None saved_starred_deps = [] for ((i, inner_dep), (_, inner_target)) in match_container_obj_or_namespace_with_literal_nodes(rhs_namespace, target): if isinstance(inner_target, ast.Starred): saved_starred_node = inner_target saved_starred_deps.append(inner_dep) continue if (inner_dep is None): inner_deps = set() else: inner_deps = {inner_dep} inner_deps |= extra_deps if isinstance(inner_target, (ast.List, ast.Tuple)): inner_namespace = flow().namespaces.get(inner_dep.obj_id, None) if ((inner_namespace is None) or (inner_namespace.obj is None)): self._handle_store_target_tuple_unpack_from_deps(inner_target, inner_deps) else: self._handle_store_target_tuple_unpack_from_namespace(inner_target, inner_namespace, extra_deps) else: self._handle_assign_target_for_deps(inner_target, inner_deps, maybe_fixup_literal_namespace=True) if (saved_starred_node is not None): self._handle_starred_store_target(saved_starred_node, saved_starred_deps) def _handle_store_target(self, target: ast.AST, value: ast.AST, skip_namespace_check: bool=False) -> None: if isinstance(target, (ast.List, ast.Tuple)): rhs_namespace = (None if skip_namespace_check else flow().namespaces.get(id(tracer().saved_assign_rhs_obj), None)) if ((rhs_namespace is None) or (rhs_namespace.obj is None)): self._handle_store_target_tuple_unpack_from_deps(target, resolve_rval_symbols(value)) else: extra_deps: Set[Symbol] = set() if isinstance(value, ast.Call): extra_deps |= resolve_rval_symbols(value) self._handle_store_target_tuple_unpack_from_namespace(target, rhs_namespace, extra_deps) else: self._handle_assign_target_for_deps(target, resolve_rval_symbols(value), maybe_fixup_literal_namespace=True) def _handle_store(self, node: Union[(ast.Assign, ast.For)]) -> None: if isinstance(node, ast.Assign): for target in node.targets: self._handle_store_target(target, node.value) elif isinstance(node, ast.For): self._handle_store_target(node.target, node.iter, skip_namespace_check=True) else: raise TypeError(('node type not supported for node: %s' % ast.dump(node))) def _handle_delete(self) -> None: assert isinstance(self.stmt_node, ast.Delete) for target in self.stmt_node.targets: try: (scope, obj, name, is_subscript) = tracer().resolve_del_data_for_target(target) scope.delete_data_symbol_for_name(name, is_subscript=is_subscript) except KeyError as e: logger.info('got key error: %s', e) def _make_lval_data_symbols(self) -> None: if isinstance(self.stmt_node, (ast.Assign, ast.For)): self._handle_store(self.stmt_node) else: self._make_lval_data_symbols_old() def _make_lval_data_symbols_old(self) -> None: symbol_edges = get_symbol_edges(self.stmt_node) should_overwrite = (not isinstance(self.stmt_node, ast.AugAssign)) is_function_def = isinstance(self.stmt_node, (ast.FunctionDef, ast.AsyncFunctionDef)) is_class_def = isinstance(self.stmt_node, ast.ClassDef) is_import = isinstance(self.stmt_node, (ast.Import, ast.ImportFrom)) if (is_function_def or is_class_def): assert (len(symbol_edges) == 1) for (target, dep_node) in symbol_edges: rval_deps = resolve_rval_symbols(dep_node) if is_import: dep_node_as_alias = cast(ast.alias, dep_node) if isinstance(self.stmt_node, ast.ImportFrom): module = (sys.modules.get(f'{self.stmt_node.module}.{dep_node_as_alias.name}') or sys.modules.get(self.stmt_node.module)) if (self.frame.f_locals is shell().user_ns): for alias in self.stmt_node.names: if (alias.name == '*'): flow().starred_import_modules.add(module.__name__) break else: module = sys.modules.get(dep_node_as_alias.name) if (module not in (None, builtins)): module_sym = tracer().create_if_not_exists_module_symbol(module, self.stmt_node, is_load=False, is_named=(dep_node_as_alias.asname is None)) if (module_sym is not None): rval_deps.update(flow().aliases.get(module_sym.obj_id, set())) target_as_str = cast(str, target) if ((target_as_str == '*') or ('.' in target_as_str)): continue logger.info('create edges from %s to %s', rval_deps, target) if is_class_def: assert (self.class_scope is not None) class_ref = self.frame.f_locals[cast(ast.ClassDef, self.stmt_node).name] self.class_scope.obj = class_ref flow().namespaces[id(class_ref)] = self.class_scope try: (scope, name, obj, is_subscript, excluded_deps) = tracer().resolve_store_data_for_target(target, self.frame) scope.upsert_symbol_for_name(name, obj, (rval_deps - excluded_deps), self.stmt_node, overwrite=should_overwrite, is_subscript=is_subscript, is_function_def=is_function_def, is_import=is_import, class_scope=self.class_scope, propagate=(not isinstance(self.stmt_node, ast.For)), symbol_node=(target if isinstance(target, ast.AST) else None), is_cascading_reactive=self.stmt_contains_cascading_reactive_rval) if isinstance(self.stmt_node, (ast.FunctionDef, ast.ClassDef, ast.AsyncFunctionDef, ast.Import, ast.ImportFrom)): self._handle_reactive_store(self.stmt_node) elif isinstance(target, ast.AST): self._handle_reactive_store(target) except KeyError: if flow().is_dev_mode: logger.warning('keyerror for %s', (astunparse.unparse(target) if isinstance(target, ast.AST) else target)) except ImportError: raise except Exception as e: logger.warning('exception while handling store: %s', e) if flow().is_test: raise e def handle_dependencies(self) -> None: for external_call in tracer().external_calls: logger.info('external call: %s', external_call) external_call._handle_impl() if self._contains_lval(): self._make_lval_data_symbols() elif isinstance(self.stmt_node, ast.Delete): self._handle_delete() else: resolve_rval_symbols(self.stmt_node) def mark_finished(self) -> None: self._finished = True self.frame = None def finished_execution_hook(self) -> None: if self._finished: return self.handle_dependencies() with tracer().dataflow_tracing_disabled(): for sym in list(tracer().this_stmt_updated_symbols): passing_watchpoints = sym.watchpoints(sym.obj, position=(flow().get_position(self.frame)[0], self.lineno), symbol_name=sym.readable_name) if passing_watchpoints: flow().active_watchpoints.append((passing_watchpoints, sym)) self.mark_finished()
def test_subscript_dependency_fp(): run_cell('lst = [0, 1, 2]') run_cell('x = 5') run_cell('y = x + lst[0]') run_cell('lst[1] = 10') run_cell('logging.info(y)') assert_not_detected('y depends only on unchanged lst[0] and not on changed lst[1]')
class ImageSize(PyClassnameExceptionRaiser): def __init__(self, *args, **kwargs): if ((len(args) == 0) and (len(kwargs) == 0)): self.x = 0 self.y = 0 self.z = 0 self.c = 0 self.t = 0 return missing_keys = get_missing_keys(kwargs.keys(), get_required_dimension_keys()) if missing_keys: self.raise_creating_clex('Missing arguments: {}'.format(','.join(missing_keys))) self.x = self._parse_int(kwargs, 'x') self.y = self._parse_int(kwargs, 'y') self.z = self._parse_int(kwargs, 'z') self.c = self._parse_int(kwargs, 'c') self.t = self._parse_int(kwargs, 't') def _parse_int(self, dict, key): try: return int(dict[key]) except ValueError: raise PyImarisWriterException('Error parsing int from parameter {} : "{}"'.format(key, dict[key])) def __str__(self): return '{} (x: {} y: {} z: {} c: {} t: {})'.format(self.__class__.__name__, self.x, self.y, self.z, self.c, self.t) def __truediv__(self, other): return ImageSize(x=(self.x // other.x), y=(self.y // other.y), z=(self.z // other.z), c=(self.c // other.c), t=(self.t // other.t))
def clip_random(image, min_shape): (img_height, img_width) = (tf.shape(image)[0], tf.shape(image)[1]) (min_height, min_width) = (min_shape[0], min_shape[1]) height = tf.cond(tf.math.greater(img_height, min_height), (lambda : tf.random.uniform([], min_height, img_height, dtype=tf.int32)), (lambda : img_height)) width = tf.cond(tf.math.greater(img_width, min_width), (lambda : tf.random.uniform([], min_width, img_width, dtype=tf.int32)), (lambda : img_width)) image = tf.image.random_crop(image, size=(height, width, tf.shape(image)[(- 1)])) return image
def main(): parser = prepare_parser() config = vars(parser.parse_args()) print(config) run(config)
def metrics_mean(l): metrics = {} for e in l: for metric_name in e: if (metric_name not in metrics): metrics[metric_name] = [] metrics[metric_name].append(e[metric_name]) for metric_name in metrics: metrics[metric_name] = np.mean(np.array(metrics[metric_name])) return metrics
def get_frame(discr, gen, dc_vars, device=None, discr_src=None): if (type(dc_vars) is not edic): dc_vars = edic(dc_vars) shape_x = (dc_vars['shape_x'] if ('shape_x' in dc_vars) else (dc_vars['dim_x'],)) shape_s = (discr.shape_s if hasattr(discr, 'shape_s') else (dc_vars['dim_s'],)) shape_v = (discr.shape_v if hasattr(discr, 'shape_v') else (dc_vars['dim_v'],)) std_v1x = (discr.std_v1x if hasattr(discr, 'std_v1x') else dc_vars['qstd_v']) std_s1vx = (discr.std_s1vx if hasattr(discr, 'std_s1vx') else dc_vars['qstd_s']) std_s1x = (discr.std_s1x if hasattr(discr, 'std_s1x') else dc_vars['qstd_s']) mode = dc_vars['mode'] if mode.startswith('svgm'): q_args_stem = (discr.v1x, std_v1x, discr.s1vx, std_s1vx) elif mode.startswith('svae'): q_args_stem = (discr.s1x, std_s1x) else: return None if ((mode == 'svgm-da2') and (discr_src is not None)): q_args = ((discr_src.v1x, (discr_src.std_v1x if hasattr(discr_src, 'std_v1x') else dc_vars['qstd_v']), discr_src.s1vx, (discr_src.std_s1vx if hasattr(discr_src, 'std_s1vx') else dc_vars['qstd_s'])) + q_args_stem) elif ((mode == 'svae-da2') and (discr_src is not None)): q_args = ((discr_src.s1x, (discr_src.std_s1x if hasattr(discr_src, 'std_s1x') else dc_vars['qstd_s'])) + q_args_stem) elif (mode in MODES_TWIST): q_args = (((None,) * len(q_args_stem)) + q_args_stem) else: q_args = (q_args_stem + ((None,) * len(q_args_stem))) if mode.startswith('svgm'): frame = SemVar(shape_s, shape_v, shape_x, dc_vars['dim_y'], gen.x1sv, dc_vars['pstd_x'], discr.y1s, *q_args, *dc_vars.sublist(['mu_s', 'sig_s', 'mu_v', 'sig_v', 'corr_sv']), (mode in MODES_DA), *dc_vars.sublist(['src_mvn_prior', 'tgt_mvn_prior']), device) elif mode.startswith('svae'): frame = SupVAE(shape_s, shape_x, dc_vars['dim_y'], gen.x1s, dc_vars['pstd_x'], discr.y1s, *q_args, *dc_vars.sublist(['mu_s', 'sig_s']), (mode in MODES_DA), *dc_vars.sublist(['src_mvn_prior', 'tgt_mvn_prior']), device) return frame
class ChatBotBasedSudokuSolver(object): def __init__(self, llm_agent) -> None: self.llm_agent = llm_agent self.parser = LLMReplyParserForSudoku() def generate_prompt(self, init_board): (role, msg_content) = ('user', self._generate_message_content(init_board)) msgs = self.llm_agent.compose_messages([role], [msg_content]) return msgs def run(self, sudoku_puzzle_instance_str): sudoku_puzzle_instance_str = self._rectify_sudoku_puzzle_instance_str(sudoku_puzzle_instance_str) prompt = self.generate_prompt(sudoku_puzzle_instance_str) llm_reply = self.llm_agent.get_reply(prompt) (success, solution) = self.parser.parse_llm_reply(llm_reply) if (not success): print('Failed to extract solution from the reply') return (False, None) (success, json_obj) = utils.extract_json_from_text_string('{{"rows": {} }}'.format(sudoku_puzzle_instance_str)) if (not success): return (False, None) (success, init_board) = self.parser.extract_sudoku_board(json_obj) if (not success): raise 'Invalid initial board: {}'.format(sudoku_puzzle_instance_str) result = RuleBasedSudokuStateChecker.check_sudoku_board(init_board, solution) return (result.is_valid, result.rows) def _rectify_sudoku_puzzle_instance_str(self, sudoku_puzzle_instance_str): sudoku_puzzle_instance_str = sudoku_puzzle_instance_str.replace("'", '') sudoku_puzzle_instance_str = sudoku_puzzle_instance_str.replace('"', '') cell_allowed = ['1', '2', '3', '4', '5', '6', '7', '8', '9', '*'] for cell in cell_allowed: sudoku_puzzle_instance_str = sudoku_puzzle_instance_str.replace(cell, (('"' + cell) + '"')) return sudoku_puzzle_instance_str
def main(argv): with open(FLAGS.config, 'r') as f: args = AttrDict(yaml.safe_load(f)) logdir = 'logs/exp' for (k, v) in args.items(): if (k == 'ref'): logdir += f"_{v.split('/')[(- 1)].split('.')[0]}" else: logdir += f'_{v}' demo_traj = jnp.array(np.load(args.ref)) demo_len = demo_traj.shape[0] args.ep_len = min(args.ep_len, demo_len) args.cycle_len = min(args.get('cycle_len', demo_len), demo_len) args.ep_len_eval = min(args.get('ep_len_eval', demo_len), demo_len) train_env = envs.get_environment(env_name='humanoid_mimic_train', system_config=args.system_config, reference_traj=demo_traj, obs_type=args.get('obs_type', 'timestamp'), cyc_len=args.cycle_len, total_length=args.ep_len_eval, rollout_length=args.ep_len, early_termination=args.get('early_termination', False), demo_replay_mode=args.demo_replay_mode, err_threshold=args.threshold, replay_rate=args.get('replay_rate', 0.05), reward_scaling=args.get('reward_scaling', 1.0), rot_weight=args.rot_weight, vel_weight=args.vel_weight, ang_weight=args.ang_weight) eval_env = envs.get_environment(env_name='humanoid_mimic', system_config=args.system_config, reference_traj=demo_traj, obs_type=args.get('obs_type', 'timestamp'), cyc_len=args.cycle_len, rot_weight=args.rot_weight, vel_weight=args.vel_weight, ang_weight=args.ang_weight) with metrics.Writer(logdir) as writer: (make_inference_fn, params, _) = dmm.train(seed=args.seed, environment=train_env, eval_environment=eval_env, episode_length=(args.ep_len - 1), eval_episode_length=(args.ep_len_eval - 1), num_envs=args.num_envs, num_eval_envs=args.num_eval_envs, learning_rate=args.lr, num_evals=(args.max_it + 1), max_gradient_norm=args.max_grad_norm, network_factory=functools.partial(apg_networks.make_apg_networks, hidden_layer_sizes=(512, 256)), normalize_observations=args.normalize_observations, save_dir=logdir, progress_fn=writer.write_scalars, use_linear_scheduler=args.use_lr_scheduler, truncation_length=args.get('truncation_length', None))
def test_getitem(): np.random.seed(1) torch.manual_seed(1) point_cloud_range = [(- 50), (- 50), (- 5), 50, 50, 3] file_client_args = dict(backend='disk') class_names = ['car', 'truck', 'trailer', 'bus', 'construction_vehicle', 'bicycle', 'motorcycle', 'pedestrian', 'traffic_cone', 'barrier'] pipeline = [dict(type='LoadPointsFromFile', coord_type='LIDAR', load_dim=5, use_dim=5, file_client_args=file_client_args), dict(type='LoadPointsFromMultiSweeps', sweeps_num=9, use_dim=[0, 1, 2, 3, 4], file_client_args=file_client_args, pad_empty_sweeps=True, remove_close=True, test_mode=True), dict(type='LoadAnnotations3D', with_bbox_3d=True, with_label_3d=True), dict(type='GlobalRotScaleTrans', rot_range=[(- 0.3925), 0.3925], scale_ratio_range=[0.95, 1.05], translation_std=[0, 0, 0]), dict(type='RandomFlip3D', flip_ratio_bev_horizontal=0.5), dict(type='PointsRangeFilter', point_cloud_range=point_cloud_range), dict(type='ObjectRangeFilter', point_cloud_range=point_cloud_range), dict(type='ObjectNameFilter', classes=class_names), dict(type='PointShuffle'), dict(type='DefaultFormatBundle3D', class_names=class_names), dict(type='Collect3D', keys=['points', 'gt_bboxes_3d', 'gt_labels_3d'])] input_modality = dict(use_lidar=True, use_camera=False, use_radar=False, use_map=False, use_external=False) dataset_cfg = dict(type='CBGSDataset', dataset=dict(type='NuScenesDataset', data_root='tests/data/nuscenes', ann_file='tests/data/nuscenes/nus_info.pkl', pipeline=pipeline, classes=class_names, modality=input_modality, test_mode=False, use_valid_flag=True, box_type_3d='LiDAR')) nus_dataset = build_dataset(dataset_cfg) assert (len(nus_dataset) == 20) data = nus_dataset[0] assert (data['img_metas'].data['flip'] is True) assert (data['img_metas'].data['pcd_horizontal_flip'] is True) assert (data['points']._data.shape == (537, 5)) data = nus_dataset[2] assert (data['img_metas'].data['flip'] is False) assert (data['img_metas'].data['pcd_horizontal_flip'] is False) assert (data['points']._data.shape == (901, 5))
def numseconds_to_numsamples(numseconds, sample_rate): candidate = int((numseconds * sample_rate)) log2 = np.log2(candidate) out_value = int((2 ** np.round(log2))) assert (out_value != 0), 'The inputs given gave an output value of 0. This is not acceptable.' return out_value
def add_interactive_args(parser): group = parser.add_argument_group('Interactive') gen_parser_from_dataclass(group, InteractiveConfig())
def PrintCategories(): sys.stderr.write(''.join(((' %s\n' % cat) for cat in _ERROR_CATEGORIES))) sys.exit(0)
class SparseBottleneck(nn.Module): def __init__(self, in_planes, growth_rate, sparsity=0.5, sparse_func='reg'): super(SparseBottleneck, self).__init__() assert (sparse_func in models.sparse_func_dict) self.bn1 = nn.BatchNorm2d(in_planes) self.conv1 = nn.Conv2d(in_planes, (4 * growth_rate), kernel_size=1, bias=False) self.bn2 = nn.BatchNorm2d((4 * growth_rate)) self.conv2 = nn.Conv2d((4 * growth_rate), growth_rate, kernel_size=3, padding=1, bias=False) self.sparse1 = models.sparse_func_dict[sparse_func](sparsity) self.sparse2 = models.sparse_func_dict[sparse_func](sparsity) def forward(self, x): out = self.conv1(self.sparse1(self.bn1(x))) out = self.conv2(self.sparse2(self.bn2(out))) out = torch.cat([out, x], 1) return out
class _FC_Layers(nn.Module): def __init__(self, opt): super(_FC_Layers, self).__init__() self.opt = opt self.classifier = nn.Sequential(nn.Linear(4096, 4096), nn.ReLU(True), nn.Dropout(), nn.Linear(4096, 4096), nn.ReLU(True), nn.Dropout(), nn.Linear(4096, self.opt.num_classes)) self._initialize_weights() def _initialize_weights(self): for m in self.modules(): if isinstance(m, nn.Conv2d): n = ((m.kernel_size[0] * m.kernel_size[1]) * m.out_channels) m.weight.data.normal_(0, math.sqrt((2.0 / n))) if (m.bias is not None): m.bias.data.zero_() elif isinstance(m, nn.ConvTranspose2d): m.weight.data.normal_(0.0, 0.02) elif isinstance(m, nn.BatchNorm2d): m.weight.data.fill_(1) m.bias.data.zero_() elif isinstance(m, nn.Linear): n = m.weight.size(1) m.weight.data.normal_(0, 0.01) m.bias.data.zero_() def forward(self, input): return self.classifier(input)
class PUSlice(object): num_classes = 32 inputchannel = 1 def __init__(self, data_dir, normlizetype): self.data_dir = data_dir self.normlizetype = normlizetype def data_preprare(self, test=False): if (len(os.path.basename(self.data_dir).split('.')) == 2): with open(self.data_dir, 'rb') as fo: list_data = pickle.load(fo, encoding='bytes') else: list_data = get_files(self.data_dir, test) with open(os.path.join(self.data_dir, 'PUSlice.pkl'), 'wb') as fo: pickle.dump(list_data, fo) if test: test_dataset = dataset(list_data=list_data, test=True, transform=None) return test_dataset else: data_pd = pd.DataFrame({'data': list_data[0], 'label': list_data[1]}) (train_pd, val_pd) = train_test_split(data_pd, test_size=0.2, random_state=40, stratify=data_pd['label']) train_dataset = dataset(list_data=train_pd, transform=data_transforms('train', self.normlizetype)) val_dataset = dataset(list_data=val_pd, transform=data_transforms('val', self.normlizetype)) return (train_dataset, val_dataset)
def glue_eval_data_collator(dataset: Dataset, batch_size: int): batch_idx = np.arange(len(dataset)) steps_per_epoch = math.ceil((len(dataset) / batch_size)) batch_idx = np.array_split(batch_idx, steps_per_epoch) for idx in batch_idx: batch = dataset[idx] batch = {k: np.array(v) for (k, v) in batch.items()} (yield batch)
class TestQuantization(unittest.TestCase): def tearDownClass(self): shutil.rmtree('./saved_results', ignore_errors=True) def test_int8_huggingface_model(self): from neural_compressor.utils.load_huggingface import OptimizedModel model_name_or_path = 'Intel/distilbert-base-uncased-finetuned-sst-2-english-int8-static' tokenizer = transformers.AutoTokenizer.from_pretrained(model_name_or_path) model = OptimizedModel.from_pretrained(model_name_or_path, from_tf=bool(('.ckpt' in model_name_or_path)), config=None, cache_dir=None, revision=None, use_auth_token=None) stat = model.state_dict() self.assertTrue((stat['classifier.module._packed_params.dtype'] == torch.qint8)) from huggingface_hub import hf_hub_download resolved_weights_file = hf_hub_download(repo_id=model_name_or_path, filename='pytorch_model.bin') q_config = torch.load(resolved_weights_file)['best_configure'] inc_model = Model(model) inc_model.q_config = q_config save_for_huggingface_upstream(inc_model, tokenizer, 'saved_results') load_model = OptimizedModel.from_pretrained('saved_results')
def run_clpr_train(input_doc): input_doc = filter_feats(input_doc, load=True) print('Finished Feature selection') input_doc = add_embeddings(input_doc) clpr_feats = [] for (idx, l) in enumerate(input_doc._.Labels): if (l == 1): clpr_feats.append(input_doc._.Features[idx]) input_doc._.CLPR_Features = clpr_feats print('Added Embeddings') adu_model = fit_clpr_model(input_doc) print('Fit CLPR Model') return adu_model
_flax class FlaxDDIMSchedulerTest(FlaxSchedulerCommonTest): scheduler_classes = (FlaxDDIMScheduler,) forward_default_kwargs = (('num_inference_steps', 50),) def get_scheduler_config(self, **kwargs): config = {'num_train_timesteps': 1000, 'beta_start': 0.0001, 'beta_end': 0.02, 'beta_schedule': 'linear'} config.update(**kwargs) return config def full_loop(self, **config): scheduler_class = self.scheduler_classes[0] scheduler_config = self.get_scheduler_config(**config) scheduler = scheduler_class(**scheduler_config) state = scheduler.create_state() (key1, key2) = random.split(random.PRNGKey(0)) num_inference_steps = 10 model = self.dummy_model() sample = self.dummy_sample_deter state = scheduler.set_timesteps(state, num_inference_steps) for t in state.timesteps: residual = model(sample, t) output = scheduler.step(state, residual, t, sample) sample = output.prev_sample state = output.state (key1, key2) = random.split(key2) return sample def check_over_configs(self, time_step=0, **config): kwargs = dict(self.forward_default_kwargs) num_inference_steps = kwargs.pop('num_inference_steps', None) for scheduler_class in self.scheduler_classes: (sample, _) = self.dummy_sample residual = (0.1 * sample) scheduler_config = self.get_scheduler_config(**config) scheduler = scheduler_class(**scheduler_config) state = scheduler.create_state() with tempfile.TemporaryDirectory() as tmpdirname: scheduler.save_config(tmpdirname) (new_scheduler, new_state) = scheduler_class.from_pretrained(tmpdirname) if ((num_inference_steps is not None) and hasattr(scheduler, 'set_timesteps')): state = scheduler.set_timesteps(state, num_inference_steps) new_state = new_scheduler.set_timesteps(new_state, num_inference_steps) elif ((num_inference_steps is not None) and (not hasattr(scheduler, 'set_timesteps'))): kwargs['num_inference_steps'] = num_inference_steps output = scheduler.step(state, residual, time_step, sample, **kwargs).prev_sample new_output = new_scheduler.step(new_state, residual, time_step, sample, **kwargs).prev_sample assert (jnp.sum(jnp.abs((output - new_output))) < 1e-05), 'Scheduler outputs are not identical' def test_from_save_pretrained(self): kwargs = dict(self.forward_default_kwargs) num_inference_steps = kwargs.pop('num_inference_steps', None) for scheduler_class in self.scheduler_classes: (sample, _) = self.dummy_sample residual = (0.1 * sample) scheduler_config = self.get_scheduler_config() scheduler = scheduler_class(**scheduler_config) state = scheduler.create_state() with tempfile.TemporaryDirectory() as tmpdirname: scheduler.save_config(tmpdirname) (new_scheduler, new_state) = scheduler_class.from_pretrained(tmpdirname) if ((num_inference_steps is not None) and hasattr(scheduler, 'set_timesteps')): state = scheduler.set_timesteps(state, num_inference_steps) new_state = new_scheduler.set_timesteps(new_state, num_inference_steps) elif ((num_inference_steps is not None) and (not hasattr(scheduler, 'set_timesteps'))): kwargs['num_inference_steps'] = num_inference_steps output = scheduler.step(state, residual, 1, sample, **kwargs).prev_sample new_output = new_scheduler.step(new_state, residual, 1, sample, **kwargs).prev_sample assert (jnp.sum(jnp.abs((output - new_output))) < 1e-05), 'Scheduler outputs are not identical' def check_over_forward(self, time_step=0, **forward_kwargs): kwargs = dict(self.forward_default_kwargs) kwargs.update(forward_kwargs) num_inference_steps = kwargs.pop('num_inference_steps', None) for scheduler_class in self.scheduler_classes: (sample, _) = self.dummy_sample residual = (0.1 * sample) scheduler_config = self.get_scheduler_config() scheduler = scheduler_class(**scheduler_config) state = scheduler.create_state() with tempfile.TemporaryDirectory() as tmpdirname: scheduler.save_config(tmpdirname) (new_scheduler, new_state) = scheduler_class.from_pretrained(tmpdirname) if ((num_inference_steps is not None) and hasattr(scheduler, 'set_timesteps')): state = scheduler.set_timesteps(state, num_inference_steps) new_state = new_scheduler.set_timesteps(new_state, num_inference_steps) elif ((num_inference_steps is not None) and (not hasattr(scheduler, 'set_timesteps'))): kwargs['num_inference_steps'] = num_inference_steps output = scheduler.step(state, residual, time_step, sample, **kwargs).prev_sample new_output = new_scheduler.step(new_state, residual, time_step, sample, **kwargs).prev_sample assert (jnp.sum(jnp.abs((output - new_output))) < 1e-05), 'Scheduler outputs are not identical' def test_scheduler_outputs_equivalence(self): def set_nan_tensor_to_zero(t): return t.at[(t != t)].set(0) def recursive_check(tuple_object, dict_object): if isinstance(tuple_object, (List, Tuple)): for (tuple_iterable_value, dict_iterable_value) in zip(tuple_object, dict_object.values()): recursive_check(tuple_iterable_value, dict_iterable_value) elif isinstance(tuple_object, Dict): for (tuple_iterable_value, dict_iterable_value) in zip(tuple_object.values(), dict_object.values()): recursive_check(tuple_iterable_value, dict_iterable_value) elif (tuple_object is None): return else: self.assertTrue(jnp.allclose(set_nan_tensor_to_zero(tuple_object), set_nan_tensor_to_zero(dict_object), atol=1e-05), msg=f'Tuple and dict output are not equal. Difference: {jnp.max(jnp.abs((tuple_object - dict_object)))}. Tuple has `nan`: {jnp.isnan(tuple_object).any()} and `inf`: {jnp.isinf(tuple_object)}. Dict has `nan`: {jnp.isnan(dict_object).any()} and `inf`: {jnp.isinf(dict_object)}.') kwargs = dict(self.forward_default_kwargs) num_inference_steps = kwargs.pop('num_inference_steps', None) for scheduler_class in self.scheduler_classes: scheduler_config = self.get_scheduler_config() scheduler = scheduler_class(**scheduler_config) state = scheduler.create_state() (sample, _) = self.dummy_sample residual = (0.1 * sample) if ((num_inference_steps is not None) and hasattr(scheduler, 'set_timesteps')): state = scheduler.set_timesteps(state, num_inference_steps) elif ((num_inference_steps is not None) and (not hasattr(scheduler, 'set_timesteps'))): kwargs['num_inference_steps'] = num_inference_steps outputs_dict = scheduler.step(state, residual, 0, sample, **kwargs) if ((num_inference_steps is not None) and hasattr(scheduler, 'set_timesteps')): state = scheduler.set_timesteps(state, num_inference_steps) elif ((num_inference_steps is not None) and (not hasattr(scheduler, 'set_timesteps'))): kwargs['num_inference_steps'] = num_inference_steps outputs_tuple = scheduler.step(state, residual, 0, sample, return_dict=False, **kwargs) recursive_check(outputs_tuple[0], outputs_dict.prev_sample) def test_step_shape(self): kwargs = dict(self.forward_default_kwargs) num_inference_steps = kwargs.pop('num_inference_steps', None) for scheduler_class in self.scheduler_classes: scheduler_config = self.get_scheduler_config() scheduler = scheduler_class(**scheduler_config) state = scheduler.create_state() (sample, _) = self.dummy_sample residual = (0.1 * sample) if ((num_inference_steps is not None) and hasattr(scheduler, 'set_timesteps')): state = scheduler.set_timesteps(state, num_inference_steps) elif ((num_inference_steps is not None) and (not hasattr(scheduler, 'set_timesteps'))): kwargs['num_inference_steps'] = num_inference_steps output_0 = scheduler.step(state, residual, 0, sample, **kwargs).prev_sample output_1 = scheduler.step(state, residual, 1, sample, **kwargs).prev_sample self.assertEqual(output_0.shape, sample.shape) self.assertEqual(output_0.shape, output_1.shape) def test_timesteps(self): for timesteps in [100, 500, 1000]: self.check_over_configs(num_train_timesteps=timesteps) def test_steps_offset(self): for steps_offset in [0, 1]: self.check_over_configs(steps_offset=steps_offset) scheduler_class = self.scheduler_classes[0] scheduler_config = self.get_scheduler_config(steps_offset=1) scheduler = scheduler_class(**scheduler_config) state = scheduler.create_state() state = scheduler.set_timesteps(state, 5) assert jnp.equal(state.timesteps, jnp.array([801, 601, 401, 201, 1])).all() def test_betas(self): for (beta_start, beta_end) in zip([0.0001, 0.001, 0.01, 0.1], [0.002, 0.02, 0.2, 2]): self.check_over_configs(beta_start=beta_start, beta_end=beta_end) def test_schedules(self): for schedule in ['linear', 'squaredcos_cap_v2']: self.check_over_configs(beta_schedule=schedule) def test_time_indices(self): for t in [1, 10, 49]: self.check_over_forward(time_step=t) def test_inference_steps(self): for (t, num_inference_steps) in zip([1, 10, 50], [10, 50, 500]): self.check_over_forward(time_step=t, num_inference_steps=num_inference_steps) def test_variance(self): scheduler_class = self.scheduler_classes[0] scheduler_config = self.get_scheduler_config() scheduler = scheduler_class(**scheduler_config) state = scheduler.create_state() assert (jnp.sum(jnp.abs((scheduler._get_variance(state, 0, 0) - 0.0))) < 1e-05) assert (jnp.sum(jnp.abs((scheduler._get_variance(state, 420, 400) - 0.14771))) < 1e-05) assert (jnp.sum(jnp.abs((scheduler._get_variance(state, 980, 960) - 0.3246))) < 1e-05) assert (jnp.sum(jnp.abs((scheduler._get_variance(state, 0, 0) - 0.0))) < 1e-05) assert (jnp.sum(jnp.abs((scheduler._get_variance(state, 487, 486) - 0.00979))) < 1e-05) assert (jnp.sum(jnp.abs((scheduler._get_variance(state, 999, 998) - 0.02))) < 1e-05) def test_full_loop_no_noise(self): sample = self.full_loop() result_sum = jnp.sum(jnp.abs(sample)) result_mean = jnp.mean(jnp.abs(sample)) assert (abs((result_sum - 172.0067)) < 0.01) assert (abs((result_mean - 0.223967)) < 0.001) def test_full_loop_with_set_alpha_to_one(self): sample = self.full_loop(set_alpha_to_one=True, beta_start=0.01) result_sum = jnp.sum(jnp.abs(sample)) result_mean = jnp.mean(jnp.abs(sample)) if (jax_device == 'tpu'): assert (abs((result_sum - 149.8409)) < 0.01) assert (abs((result_mean - 0.1951)) < 0.001) else: assert (abs((result_sum - 149.8295)) < 0.01) assert (abs((result_mean - 0.1951)) < 0.001) def test_full_loop_with_no_set_alpha_to_one(self): sample = self.full_loop(set_alpha_to_one=False, beta_start=0.01) result_sum = jnp.sum(jnp.abs(sample)) result_mean = jnp.mean(jnp.abs(sample)) if (jax_device == 'tpu'): pass else: assert (abs((result_sum - 149.0784)) < 0.01) assert (abs((result_mean - 0.1941)) < 0.001) def test_prediction_type(self): for prediction_type in ['epsilon', 'sample', 'v_prediction']: self.check_over_configs(prediction_type=prediction_type)
def combine_columns(df: pd.DataFrame, columns: List[str], separator: str=' / ') -> pd.DataFrame: def _pad_percentage(f: float) -> str: return '{:5.1f}'.format(f).replace(' ', '\\phantom{0}') out_df = df.xs(columns[0], axis=1, level=(- 1)).applymap(_pad_percentage) for col in columns[1:]: out_df += (separator + df.xs(col, axis=1, level=(- 1)).applymap(_pad_percentage)) return out_df
class Parallelize(): def __init__(self, benchmark: Benchmark, num_workers: int=4): self.benchmark = benchmark self.num_workers = num_workers def run_single_job(self, pipeline_class: type, config: blocks.PipelineConfig, filepath: Path, description: Text) -> Annotation: idx_process = (int(current_process().name.split('-')[1]) - 1) pipeline = pipeline_class(config) progress = TQDMProgressBar(description, leave=False, position=idx_process, do_close=True) return self.benchmark.run_single(pipeline, filepath, progress) def __call__(self, pipeline_class: type, config: blocks.PipelineConfig, metric: Optional[BaseMetric]=None) -> Union[(pd.DataFrame, List[Annotation])]: audio_file_paths = self.benchmark.get_file_paths() num_audio_files = len(audio_file_paths) try: torch.multiprocessing.set_start_method('spawn') except RuntimeError: pass freeze_support() pool = Pool(processes=self.num_workers, initargs=(RLock(),), initializer=tqdm.set_lock) arg_list = [(pipeline_class, config, filepath, f'Streaming {filepath.stem} ({(i + 1)}/{num_audio_files})') for (i, filepath) in enumerate(audio_file_paths)] jobs = [pool.apply_async(self.run_single_job, args=args) for args in arg_list] pool.close() predictions = [job.get() for job in jobs] metric = (pipeline_class.suggest_metric() if (metric is None) else metric) return self.benchmark.evaluate(predictions, metric)
def get_images(directory, img_ext): assert os.path.isdir(directory) image_paths = glob.glob((directory + f'/**/*{img_ext}'), recursive=True) for path in image_paths: if (path.split(os.sep)[(- 2)] not in ['damaged_jpeg', 'jpeg_write']): (yield path)