code
stringlengths
101
5.91M
class DotProductAttention(nn.Module): def __init__(self, dim: int, scale: bool=True) -> None: super(DotProductAttention, self).__init__() if scale: self.sqrt_dim = np.sqrt(dim) else: self.sqrt_dim = 1 def forward(self, query: torch.FloatTensor, key: torch.FloatTensor, value: torch.FloatTensor, mask: Optional[torch.FloatTensor]=None) -> Tuple[(torch.FloatTensor, torch.FloatTensor)]: score = (torch.matmul(query, key.transpose(2, 3)) / self.sqrt_dim) if (mask is not None): score.masked_fill_(mask, (- 10000.0)) attn = F.softmax(score, (- 1)) if (len(query.size()) == 3): context = torch.bmm(attn, value) else: context = torch.matmul(attn, value) return (context, attn)
def _binary_erosion(input, structure, iterations, mask, output, border_value, origin, invert, brute_force): try: iterations = operator.index(iterations) except TypeError: raise TypeError('iterations parameter should be an integer') input = numpy.asarray(input) if numpy.iscomplexobj(input): raise TypeError('Complex type not supported') if (structure is None): structure = generate_binary_structure(input.ndim, 1) else: structure = numpy.asarray(structure, dtype=bool) if (structure.ndim != input.ndim): raise RuntimeError('structure and input must have same dimensionality') if (not structure.flags.contiguous): structure = structure.copy() if (numpy.prod(structure.shape, axis=0) < 1): raise RuntimeError('structure must not be empty') if (mask is not None): mask = numpy.asarray(mask) if (mask.shape != input.shape): raise RuntimeError('mask and input must have equal sizes') origin = _ni_support._normalize_sequence(origin, input.ndim) cit = _center_is_true(structure, origin) if isinstance(output, numpy.ndarray): if numpy.iscomplexobj(output): raise TypeError('Complex output type not supported') else: output = bool output = _ni_support._get_output(output, input) if (iterations == 1): _nd_image.binary_erosion(input, structure, mask, output, border_value, origin, invert, cit, 0) return output elif (cit and (not brute_force)): (changed, coordinate_list) = _nd_image.binary_erosion(input, structure, mask, output, border_value, origin, invert, cit, 1) structure = structure[tuple(([slice(None, None, (- 1))] * structure.ndim))] for ii in range(len(origin)): origin[ii] = (- origin[ii]) if (not (structure.shape[ii] & 1)): origin[ii] -= 1 if (mask is not None): mask = numpy.asarray(mask, dtype=numpy.int8) if (not structure.flags.contiguous): structure = structure.copy() _nd_image.binary_erosion2(output, structure, mask, (iterations - 1), origin, invert, coordinate_list) return output else: tmp_in = numpy.empty_like(input, dtype=bool) tmp_out = output if ((iterations >= 1) and (not (iterations & 1))): (tmp_in, tmp_out) = (tmp_out, tmp_in) changed = _nd_image.binary_erosion(input, structure, mask, tmp_out, border_value, origin, invert, cit, 0) ii = 1 while ((ii < iterations) or ((iterations < 1) and changed)): (tmp_in, tmp_out) = (tmp_out, tmp_in) changed = _nd_image.binary_erosion(tmp_in, structure, mask, tmp_out, border_value, origin, invert, cit, 0) ii += 1 return output
def get_datapoints_map(influences_collections: List[Dict[(int, float)]]) -> Tuple[(List[int], Dict[(int, int)])]: possible_datapoints = [] for influences in influences_collections: possible_datapoints.extend(list(influences.keys())) possible_datapoints = sorted(set(possible_datapoints)) datapoints_map = dict(((v, k) for (k, v) in enumerate(possible_datapoints))) return (possible_datapoints, datapoints_map)
def smallest_positions(word, subword, pos=0) -> (list | bool): pos -= 1 n = len(word) res = ([None] * len(subword)) for (i, swi) in enumerate(subword): for j in range((pos + 1), (n + 1)): if (j == n): return False if (word[j] == swi): pos = j break if (pos != j): return False res[i] = pos return res
def opt_score(method, data, global_knowledge): task = 'llm-scorer' evaluator = LocalLLM(method, task=task) template = prompt_templates['opt_score'] (scores, meta_data) = ([], []) for sample in tqdm(data): target = sample['output'] local_knowledge = sample['knowledge'] prompt = 'Generate a text that is factually correct and consistent with the instruction.\n' instruction = template.format(instruction=get_instruction(sample), local_knowledge=local_knowledge, global_knowledge=retrieve_global_knowledge(target, global_knowledge)) score = evaluator.score(instruction, target, prompt) ppl = np.exp(avg_loss) scores.append((1 / ppl)) meta_data.append(f'PPL:{ppl}') return (scores, meta_data)
def test_angular_size(): from skypy.galaxies import morphology cosmology = FlatLambdaCDM(Om0=1.0, H0=70.0) scalar_radius = (1.0 * units.kpc) scalar_redshift = 1.0 angular_size = morphology.angular_size(scalar_radius, scalar_redshift, cosmology) assert np.isscalar(angular_size.value) assert angular_size.unit.is_equivalent(units.rad) radius_without_units = 1.0 with pytest.raises(units.UnitTypeError): morphology.angular_size(radius_without_units, scalar_redshift, cosmology)
def from_music21(stream: Stream, resolution: int=DEFAULT_RESOLUTION) -> Union[(Music, List[Music], Track, List[Track])]: if isinstance(stream, Opus): return from_music21_opus(stream, resolution) elif isinstance(stream, Part): return from_music21_part(stream, resolution) else: return from_music21_score(stream, resolution)
def labelled_pronoun(row): txt = row.text tokens = txt.split(' ') if ('tokens' in row.index): tokens = row.tokens if ('a_span' in row.index): start_a = row.a_span[0] end_a = row.a_span[1] start_b = row.b_span[0] end_b = row.b_span[1] start_p = row.pronoun_offset_token end_p = row.pronoun_offset_token else: start_a = (len(txt[:row.a_offset].split(' ')) - 1) end_a = ((start_a + len(row.a.split(' '))) - 1) start_b = (len(txt[:row.b_offset].split(' ')) - 1) end_b = ((start_b + len(row.b.split(' '))) - 1) start_p = (len(txt[:row.pronoun_offset].split(' ')) - 1) end_p = ((start_p + len(row.pronoun.split(' '))) - 1) clusters = [[[start_a, end_a]], [[start_b, end_b]]] if row.a_coref: clusters[0].append([start_p, end_p]) elif row.b_coref: clusters[1].append([start_p, end_p]) else: clusters.append([[start_p, end_p]]) return (tokens, clusters)
class HybridEmbedding(Layer): def __init__(self, embed_size, unfixed_embed_size, embed_dim, init='uniform', name='HybridEmbedding'): super(HybridEmbedding, self).__init__() self.init = initializations.get(init) self.unfixed_embed_size = unfixed_embed_size self.W_unfixed = self.init((embed_size, embed_dim)) self.W_fixed = self.init((embed_size, embed_dim)) self.W_fixed.name = 'HybridEmbedding_fiexed_embed_matrix' self.params = [self.W_unfixed] if (name is not None): self.set_name(name) def get_output_mask(self, X): return (T.ones_like(X) * (1 - T.eq(X, 0))) def __call__(self, X, mask_zero=False): cond = T.lt(X, self.unfixed_embed_size) out = T.switch(T.shape_padright(cond), self.W_unfixed[X], self.W_fixed[X]) if mask_zero: return (out, self.get_output_mask(X)) else: return out
def test_fbms(model): test_set = TestFBMS('../DB/FBMS/TestSet') test_loader = torch.utils.data.DataLoader(test_set, batch_size=1, num_workers=4) model.cuda() ious = [] for vos_data in test_loader: imgs = vos_data['imgs'].cuda() flows = vos_data['flows'].cuda() masks = vos_data['masks'] video_name = vos_data['video_name'][0] files = vos_data['files'] os.makedirs('outputs/FBMS_test/{}'.format(video_name), exist_ok=True) vos_out = model(imgs, flows) iou = 0 count = 0 for i in range(masks.size(1)): tv.utils.save_image(vos_out['masks'][(0, i)].float().cpu(), 'outputs/FBMS_test/{}/{}'.format(video_name, files[i][0].split('/')[(- 1)])) if (torch.sum(masks[(0, i)]) == 0): continue iou = (iou + (torch.sum((masks[(0, i)] * vos_out['masks'][(0, i)].cpu())) / torch.sum((masks[(0, i)] + vos_out['masks'][(0, i)].cpu()).clamp(0, 1)))) count = (count + 1) print('{} iou: {:.5f}'.format(video_name, (iou / count))) ious.append((iou / count)) print("total seqs' iou: {:.5f}\n".format((sum(ious) / len(ious))))
class PreTrainingConfig(Config): def __init__(self, **kwargs): self.tokenizer: transformers.PreTrainedTokenizer = kwargs.pop('tokenizer') self.stoi = self.tokenizer.get_vocab() self.special_ids = list(set(self.tokenizer.all_special_ids)) self.non_special_ids = [idx for idx in self.stoi.values() if (idx not in self.special_ids)] super().__init__(**kwargs) def cls_id(self): return self.tokenizer.cls_token_id def sep_id(self): return self.tokenizer.sep_token_id def unk_id(self): return self.tokenizer.unk_token_id def pad_id(self): return self.tokenizer.pad_token_id def mask_id(self): return self.tokenizer.mask_token_id
class Main(): def regular(self): for family in LANGS.keys(): print('regular', family) regular(family) def hall(self): for family in LANGS.keys(): print('hall', family) halluication(family) def concat(self): for family in LANGS.keys(): print('concat', family) concat_langid(family) def concat_hall(self): for family in LANGS.keys(): print('concat_hall', family) concat_halluication(family) def all(self): self.regular() self.hall() self.concat() self.concat_hall() def gen_langs(self): langs = [] for family in LANGS.keys(): langs.extend(LANGS[family]) print(' '.join(sorted(langs))) print(len(langs)) def gen_family(self): family = list(LANGS.keys()) print(' '.join(sorted(family))) print(len(family))
class BaseModel(nn.Module, ABC): name = 'base' def __init__(self): super().__init__() self.best_state_dict = None def set_best(self): self.best_state_dict = copy.deepcopy(self.state_dict()) def recover_best(self): self.load_state_dict(self.best_state_dict) def save(self, path): fname = self.get_name(path) torch.save({'kwargs': self.get_args(), 'model_state_dict': self.state_dict()}, fname) def get_args(self): pass def load(cls, path): checkpoints = cls.load_checkpoint(path) model = cls(**checkpoints['kwargs']) model.load_state_dict(checkpoints['model_state_dict']) return model def load_checkpoint(cls, path): fname = cls.get_name(path) return torch.load(fname, map_location=constants.device) def get_name(cls, path): return ('%s/model.tch' % path)
def build_dataloader(dataset, batch_size=1, num_workers=1, training=True, dist=False, persistent_workers=True): shuffle = training sampler = (DistributedSampler(dataset, shuffle=shuffle) if dist else None) if (sampler is not None): shuffle = False if training: return DataLoader(dataset, batch_size=batch_size, num_workers=num_workers, collate_fn=dataset.collate_fn, shuffle=shuffle, sampler=sampler, drop_last=True, pin_memory=True, persistent_workers=persistent_workers) else: assert (batch_size == 1) return DataLoader(dataset, batch_size=batch_size, num_workers=num_workers, collate_fn=dataset.collate_fn, shuffle=False, sampler=sampler, drop_last=False, pin_memory=True, persistent_workers=persistent_workers)
def update_edge_info(tensor_info): info = tensor_info.filter(regex='tensor|dtype|shape') info = info.to_dict('index') def format_value(meta): msg = '{0}\n {1} \n {2}' if (meta['fp32net_shape'] == meta['int8net_shape']): shape = meta['int8net_shape'] else: shape = [meta['fp32net_shape'], meta['int8net_shape']] dtype = (meta['fp32net_dtype'], meta['int8net_dtype']) return msg.format(meta['tensor'], shape, dtype) return {k: format_value(v) for (k, v) in info.items()}
def register_Ns3EpcX2Sap_methods(root_module, cls): cls.add_constructor([]) cls.add_constructor([param('ns3::EpcX2Sap const &', 'arg0')]) cls.add_static_attribute('m_maxPdcpSn', 'uint16_t const', is_const=True) return
_model def tf_efficientnet_b4(pretrained=False, **kwargs): kwargs['bn_eps'] = BN_EPS_TF_DEFAULT kwargs['pad_type'] = 'same' model = _gen_efficientnet('tf_efficientnet_b4', channel_multiplier=1.4, depth_multiplier=1.8, pretrained=pretrained, **kwargs) return model
class HMMSearchJob(ExternalSeppJob): def __init__(self, **kwargs): self.job_type = 'hmmsearch' ExternalSeppJob.__init__(self, self.job_type, **kwargs) self.hmmmodel = None self.fragments = None self.outfile = None self.elim = None self.filters = None if hasattr(sepp.config.options().hmmsearch, 'piped'): self.pipe = (sepp.config.options().hmmsearch.piped.strip().lower() == 'true') else: self.pipe = True self.results_on_temp = (not self.pipe) def setup(self, hmmmodel, fragments, output_file, elim=None, filters=True, **kwargs): self.hmmmodel = hmmmodel self.fragments = fragments self.outfile = output_file self.elim = elim self.filters = filters self._kwargs = kwargs def partial_setup_for_subproblem(self, fragments_file, subproblem, elim=None, filters=True, **kwargs): assert isinstance(subproblem, sepp.problem.SeppProblem) self.outfile = sepp.filemgr.tempfile_for_subproblem('hmmsearch.results.', subproblem) self.fragments = fragments_file if (not self.fragments): self.fake_run = True self.elim = elim self.filters = filters self._kwargs = kwargs def get_invocation(self): invoc = [self.path, '--noali', '--cpu', '1'] if (not self.pipe): invoc.extend(['-o', self.outfile]) if (self.elim is not None): invoc.extend(['-E', str(self.elim)]) if (not self.filters): invoc.extend(['--max']) if ('user_options' in self._kwargs): invoc.extend(self._kwargs['user_options'].split()) invoc.extend([self.hmmmodel, self.fragments]) return invoc def characterize_input(self): return ('model:%s, fragments:%s, elim:%s, filter:%s, output:%s' % (self.hmmmodel, self.fragments, self.elim, self.filters, ('Piped' if self.pipe else self.outfile))) def read_results(self): if self.results_on_temp: if self.fake_run: res = {} else: assert os.path.exists(self.outfile) assert (os.stat(self.outfile)[stat.ST_SIZE] != 0) with open(self.outfile, 'r') as outfile: res = self.read_results_from_temp(outfile) with open(self.outfile, 'w') as target: target.write(str(res)) return self.outfile else: if self.fake_run: res = {} elif self.pipe: outfile = self.stdoutdata.split('\n') res = self.read_results_from_temp(outfile) else: outfile = open(self.outfile, 'r') res = self.read_results_from_temp(outfile) outfile.close() return res def read_results_from_temp(self, outfile): results = {} pattern = re.compile('([^\\s]+)\\s+([^\\s]+)\\s+([^\\s]+)\\s+([^\\s]+)\\s+([^\\s]+)\\s+([^\\s]+)\\s+([^\\s]+)\\s+([^\\s]+)\\s+([^\\s]+)') start_reading = False for line in outfile: line = line.strip() if ((not start_reading) and (line.startswith('E-value') is True)): start_reading = True elif (start_reading and (line == '')): start_reading = False break elif start_reading: matches = pattern.search(line) if ((matches is not None) and (matches.group(0).find('--') == (- 1))): results[matches.group(9).strip()] = (float(matches.group(1).strip()), float(matches.group(2).strip())) return results
def init_pretrained_weights(model, model_url): pretrain_dict = model_zoo.load_url(model_url) pattern = re.compile('^(.*denselayer\\d+\\.(?:norm|relu|conv))\\.((?:[12])\\.(?:weight|bias|running_mean|running_var))$') for key in list(pretrain_dict.keys()): res = pattern.match(key) if res: new_key = (res.group(1) + res.group(2)) pretrain_dict[new_key] = pretrain_dict[key] del pretrain_dict[key] model_dict = model.state_dict() pretrain_dict = {k: v for (k, v) in pretrain_dict.items() if ((k in model_dict) and (model_dict[k].size() == v.size()))} model_dict.update(pretrain_dict) model.load_state_dict(model_dict)
def IntegralLatticeDirectSum(Lattices, return_embeddings=False): for L in Lattices: if (not isinstance(L, FreeQuadraticModule_integer_symmetric)): raise ValueError('Lattices must be a list of lattices') N = len(Lattices) dims = [L_i.dimension() for L_i in Lattices] degrees = [L_i.degree() for L_i in Lattices] degree_tot = sum(degrees) sum_degree = [sum(degrees[:i]) for i in range((N + 1))] inner_product_list = [copy(L_i.inner_product_matrix()) for L_i in Lattices] IM = matrix.block_diagonal(inner_product_list) ambient = FreeQuadraticModule(ZZ, degree_tot, inner_product_matrix=IM) basis = [matrix.block(1, 3, [matrix.zero(dims[i], sum_degree[i]), Lattices[i].basis_matrix(), matrix.zero(dims[i], (sum_degree[(- 1)] - sum_degree[(i + 1)]))]) for i in range(N)] basis_matrix = matrix.block(N, 1, basis) ipm = ambient.inner_product_matrix() direct_sum = FreeQuadraticModule_integer_symmetric(ambient=ambient, basis=basis_matrix, inner_product_matrix=ipm, already_echelonized=False) if (not return_embeddings): return direct_sum sum_dims = [sum(dims[:i]) for i in range((N + 1))] phi = [Lattices[i].hom(direct_sum.basis()[sum_dims[i]:sum_dims[(i + 1)]]) for i in range(N)] return [direct_sum, phi]
def plot_log(axs, log, info, xticks=None, yticks=None, xnbins=None, ynbins=None, groups=None, show_legends=True, swap_axes=False): import matplotlib.pyplot as plt if (axs is None): fig = plt.figure() else: fig = None if (groups is None): n_gr = len(info) groups = nm.arange(n_gr) else: n_gr = len(groups) n_col = min(5.0, nm.fix(nm.sqrt(n_gr))) if (int(n_col) == 0): n_row = 0 else: n_row = int(nm.ceil((n_gr / n_col))) n_col = int(n_col) if (xticks is None): xticks = ([None] * n_gr) if (yticks is None): yticks = ([None] * n_gr) if (xnbins is None): xnbins = ([None] * n_gr) if (ynbins is None): ynbins = ([None] * n_gr) isub = offset = 0 for (ig, (xlabel, ylabel, yscale, names, plot_kwargs)) in ordered_iteritems(info): if (ig not in groups): offset += len(names) continue if (axs is None): ax = fig.add_subplot(n_row, n_col, (isub + 1)) else: ax = axs[ig] if (not swap_axes): (xnb, ynb) = (xnbins[isub], ynbins[isub]) (xti, yti) = (xticks[isub], yticks[isub]) ax.set_yscale(yscale) for (ip, name) in enumerate(names): (xs, ys, vlines) = log[(ip + offset)] draw_data(ax, xs, ys, name, plot_kwargs[ip]) for x in vlines: ax.axvline(x, color='k', alpha=0.3) else: (xlabel, ylabel) = (ylabel, xlabel) (xti, yti) = (yticks[isub], xticks[isub]) (xnb, ynb) = (ynbins[isub], xnbins[isub]) ax.set_xscale(yscale) for (ip, name) in enumerate(names): (xs, ys, vlines) = log[(ip + offset)] draw_data(ax, xs, ys, name, plot_kwargs[ip], swap_axes=True) for x in vlines: ax.axhline(x, color='k', alpha=0.3) offset += len(names) if (xti is not None): ax.set_xticks(xti) if (yti is not None): ax.set_yticks(yti) if (xnb is not None): ax.locator_params(tight=True, axis='x', nbins=xnb) if (ynb is not None): ax.locator_params(tight=True, axis='y', nbins=ynb) if xlabel: ax.set_xlabel(xlabel) if ylabel: ax.set_ylabel(ylabel) if show_legends: ax.legend(loc='best') isub += 1 plt.tight_layout(pad=0.5)
() ('serialization-dir', type=click.Path(exists=True)) ('save-dir', type=click.Path()) def convert_allennlp_to_huggingface_model(serialization_dir: str, save_dir: str): logger.info(f'Loading allennlp data from {serialization_dir}...') model_weights_path = (Path(serialization_dir) / 'best.th') config_path = (Path(serialization_dir) / 'config.json') vocabulary_path = (Path(serialization_dir) / 'vocabulary/labels.txt') model_weights = torch.load(model_weights_path) config = json.load(open(config_path)) if (config['dataset_reader']['tokenizer']['type'] != 'pretrained_transformer'): raise ValueError('Only models that use a HuggingFace tokenizer can be converted.') huggingface_tokenizer_name = config['dataset_reader']['tokenizer']['model_name'] if (config['model']['type'] != 'span_ner'): raise ValueError('This script converts the weights of ExhaustiveNERModel (registered as `span_ner`).') if (config['model']['feature_extractor']['embedder']['type'] != 'transformers-luke'): raise ValueError('Only models that use TransformersLukeEmbedder (registered as `transformers-luke`) can be converted.') huggingface_model_name = config['model']['feature_extractor']['embedder']['model_name'] config = AutoConfig.from_pretrained(huggingface_model_name) tokenizer = AutoTokenizer.from_pretrained(huggingface_tokenizer_name) setattr(config, 'classifier_bias', ('classifier.bias' in model_weights)) setattr(config, 'num_labels', model_weights['classifier.weight'].size(0)) labels = [label.strip() for label in open(vocabulary_path, 'r')] config.id2label = {id_: label for (id_, label) in enumerate(labels)} config.label2id = {label: id_ for (id_, label) in enumerate(labels)} downstream_luke_model = LukeForEntitySpanClassification(config) huggingface_model_weights = {} for (key, w) in model_weights.items(): huggingface_model_weights[key.replace('feature_extractor.embedder.luke_model', 'luke')] = w downstream_luke_model.load_state_dict(huggingface_model_weights, strict=True) downstream_luke_model.save_pretrained(save_dir) tokenizer.save_pretrained(save_dir) logger.info(f'Saved hugging face model in {save_dir}.')
def get_bias_parameters(model): return [param for (name, param) in model.named_parameters() if ('bias' in name)]
class Recorder(): def __init__(self, scoring_num_list, record_filtered=True, prefix=''): self.elems = [] self.filtered_elems = [] self.seen_smis = set() self.record_filtered = record_filtered if self.record_filtered: self.rd_filter = RDFilter() self.scoring_num_list = scoring_num_list self.prefix = prefix self.max_size = max(scoring_num_list) self.t = 0 def __len__(self): return len(self.elems) def add_list(self, smis, scores): new_elems = [RecorderElement(smi=smi, score=score) for (smi, score) in zip(smis, scores)] new_elems = list(set(new_elems)) new_elems = list(filter((lambda elem: (elem.smi not in self.seen_smis)), new_elems)) self.seen_smis = self.seen_smis.union(smis) self.elems.extend(new_elems) self.elems = list(set(self.elems)) if (len(self.elems) > self.max_size): self.elems = sorted(self.elems, reverse=True)[:self.max_size] if self.record_filtered: filtered_new_elems = new_elems if (len(self.filtered_elems) > 0): min_filtered_elem_score = min(self.filtered_elems).score filtered_new_elems = list(filter((lambda elem: (elem.score > min_filtered_elem_score)), filtered_new_elems)) if (len(filtered_new_elems) > self.max_size): filtered_new_elems = sorted(filtered_new_elems, reverse=True)[:self.max_size] filtered_new_elems = list(filter((lambda elem: (self.rd_filter(elem.smi) > 0.5)), filtered_new_elems)) self.filtered_elems.extend(filtered_new_elems) if (len(self.filtered_elems) > self.max_size): self.filtered_elems = sorted(self.filtered_elems, reverse=True)[:self.max_size] def full(self): return ((len(self.elems) == self.max_size) and ((not self.record_filtered) or (self.record_filtered and (len(self.filtered_elems) == self.max_size)))) def evaluate(self, rd_filtered): if (not rd_filtered): scores = [elem.score for elem in sorted(self.elems, reverse=True)] else: scores = [elem.score for elem in sorted(self.filtered_elems, reverse=True)] evaluation_elemwise_scores = np.array(scores) evaluation_score = 0.0 for scoring_num in self.scoring_num_list: evaluation_score += (evaluation_elemwise_scores[:scoring_num].mean() / len(self.scoring_num_list)) return evaluation_score def log(self): score = self.evaluate(rd_filtered=False) neptune.log_metric(f'{self.prefix}eval_optimized_score', score) if self.record_filtered: filtered_score = self.evaluate(rd_filtered=True) neptune.log_metric(f'{self.prefix}eval_filtered_score', filtered_score) self.t += 1 def log_final(self): for elem in self.elems: neptune.log_text('optimized_smi', elem.smi) neptune.log_metric('optimized_score', elem.score) if self.record_filtered: for elem in self.filtered_elems: neptune.log_text('filtered_smi', elem.smi) neptune.log_metric('filtered_score', elem.score) def get_topk(self, k): self.elems = sorted(self.elems, reverse=True)[:k] return ([elem.smi for elem in self.elems], [elem.score for elem in self.elems])
def main(): with tf.Session() as sess: (model_cfg, model_outputs) = posenet.load_model(args.model, sess) output_stride = model_cfg['output_stride'] num_images = args.num_images filenames = [f.path for f in os.scandir(args.image_dir) if (f.is_file() and f.path.endswith(('.png', '.jpg')))] if (len(filenames) > num_images): filenames = filenames[:num_images] images = {f: posenet.read_imgfile(f, 1.0, output_stride)[0] for f in filenames} start = time.time() for i in range(num_images): (heatmaps_result, offsets_result, displacement_fwd_result, displacement_bwd_result) = sess.run(model_outputs, feed_dict={'image:0': images[filenames[(i % len(filenames))]]}) output = posenet.decode_multiple_poses(heatmaps_result.squeeze(axis=0), offsets_result.squeeze(axis=0), displacement_fwd_result.squeeze(axis=0), displacement_bwd_result.squeeze(axis=0), output_stride=output_stride, max_pose_detections=10, min_pose_score=0.25) print('Average FPS:', (num_images / (time.time() - start)))
def loss_lsgan_gen(dis_out_fake): gen_loss = (0.5 * ((dis_out_fake - torch.ones_like(dis_out_fake)) ** 2)) return gen_loss.mean()
class Partition10(nn.Module): LAYER_SCOPES = ['T5ForConditionalGeneration/T5Stack[encoder]/T5Block[13]/T5LayerSelfAttention[0]/T5Attention[SelfAttention]/Linear[v]', 'T5ForConditionalGeneration/T5Stack[encoder]/T5Block[13]/T5LayerSelfAttention[0]/T5Attention[SelfAttention]/Dropout[dropout]', 'T5ForConditionalGeneration/T5Stack[encoder]/T5Block[13]/T5LayerSelfAttention[0]/T5Attention[SelfAttention]/Linear[o]', 'T5ForConditionalGeneration/T5Stack[encoder]/T5Block[13]/T5LayerSelfAttention[0]/Dropout[dropout]', 'T5ForConditionalGeneration/T5Stack[encoder]/T5Block[13]/T5LayerFF[1]/T5LayerNorm[layer_norm]', 'T5ForConditionalGeneration/T5Stack[encoder]/T5Block[13]/T5LayerFF[1]/T5DenseReluDense[DenseReluDense]/Linear[wi]', 'T5ForConditionalGeneration/T5Stack[encoder]/T5Block[13]/T5LayerFF[1]/T5DenseReluDense[DenseReluDense]/Dropout[dropout]', 'T5ForConditionalGeneration/T5Stack[encoder]/T5Block[13]/T5LayerFF[1]/T5DenseReluDense[DenseReluDense]/Linear[wo]', 'T5ForConditionalGeneration/T5Stack[encoder]/T5Block[13]/T5LayerFF[1]/Dropout[dropout]', 'T5ForConditionalGeneration/T5Stack[encoder]/T5Block[14]/T5LayerSelfAttention[0]/T5LayerNorm[layer_norm]', 'T5ForConditionalGeneration/T5Stack[encoder]/T5Block[14]/T5LayerSelfAttention[0]/T5Attention[SelfAttention]/Linear[q]', 'T5ForConditionalGeneration/T5Stack[encoder]/T5Block[14]/T5LayerSelfAttention[0]/T5Attention[SelfAttention]/Linear[k]'] TENSORS = [] def __init__(self, layers, tensors, device='cuda:10'): super().__init__() for (idx, layer_scope) in enumerate(self.LAYER_SCOPES): self.add_module(f'l_{idx}', layers[layer_scope]) b = p = 0 for tensor_scope in self.TENSORS: tensor = tensors[tensor_scope] if isinstance(tensor, nn.Parameter): self.register_parameter(f'p_{p}', tensor) p += 1 else: self.register_buffer(f'b_{b}', tensor) b += 1 self.device = torch.device(device) self.input_structure = [1, 1, 1, 1, 1, 1] self.lookup = {'l_0': 'encoder.13.0.SelfAttention.v', 'l_1': 'encoder.13.0.SelfAttention.dropout', 'l_2': 'encoder.13.0.SelfAttention.o', 'l_3': 'encoder.13.0.dropout', 'l_4': 'encoder.13.1.layer_norm', 'l_5': 'encoder.13.1.DenseReluDense.wi', 'l_6': 'encoder.13.1.DenseReluDense.dropout', 'l_7': 'encoder.13.1.DenseReluDense.wo', 'l_8': 'encoder.13.1.dropout', 'l_9': 'encoder.14.0.layer_norm', 'l_10': 'encoder.14.0.SelfAttention.q', 'l_11': 'encoder.14.0.SelfAttention.k'} self.to(self.device) def forward(self, *args): (x0, x1, x2, x3, x4, x5) = unflatten(args, self.input_structure) t_0 = self.l_0(x2) t_0 = t_0.view(x3, (- 1), 32, 128) t_0 = t_0.transpose(1, 2) t_1 = x5.transpose(3, 2) t_1 = torch.matmul(x4, t_1) t_1 += x0 t_2 = t_1.float() t_2 = torch.nn.functional.softmax(t_2, dim=(- 1), _stacklevel=3, dtype=None) t_1 = t_2.type_as(t_1) t_1 = self.l_1(t_1) t_0 = torch.matmul(t_1, t_0) t_0 = t_0.transpose(1, 2) t_0 = t_0.contiguous() t_0 = t_0.view(x3, (- 1), 4096) t_0 = self.l_2(t_0) t_0 = self.l_3(t_0) t_0 = (x1 + t_0) t_1 = self.l_4(t_0) t_1 = self.l_5(t_1) t_1 = torch.nn.functional.relu(t_1, inplace=False) t_1 = self.l_6(t_1) t_1 = self.l_7(t_1) t_1 = self.l_8(t_1) t_1 = (t_0 + t_1) t_0 = self.l_9(t_1) t_2 = t_0.size() t_3 = self.l_10(t_0) t_4 = self.l_11(t_0) t_2 = t_2[0] t_3 = t_3.view(t_2, (- 1), 32, 128) t_3 = t_3.transpose(1, 2) t_4 = t_4.view(t_2, (- 1), 32, 128) t_4 = t_4.transpose(1, 2) t_4 = t_4.transpose(3, 2) t_4 = torch.matmul(t_3, t_4) t_4 += x0 return list(flatten((x0, t_1, t_0, t_2, t_4))) def state_dict(self, *args, **kwargs): return state_dict(self, *args, **kwargs) def load_state_dict(self, *args, **kwargs): return load_state_dict(self, *args, **kwargs) def named_parameters(self, *args, **kwargs): return named_parameters(self, *args, **kwargs) def named_buffers(self, *args, **kwargs): return named_buffers(self, *args, **kwargs) def cpu(self): return cpu(self) def cuda(self, device=None): return cuda(self, device=device) def to(self, *args, **kwargs): return to(self, *args, **kwargs)
def read_binary_audio(bin_data, tar_sr=None): (data, ori_sr) = sf.read(io.BytesIO(bin_data), dtype='float32') data = data.T if ((tar_sr is not None) and (ori_sr != tar_sr)): data = librosa.resample(data, ori_sr, tar_sr) else: tar_sr = ori_sr data = np.clip(data, (- 1), 1) data = (data * 32768.0) return (torch.FloatTensor(data.astype(np.float32)), tar_sr)
class BasicBlock(nn.Module): expansion = 1 def __init__(self, inplanes, planes, stride=1, downsample=None, num_group=32): super(BasicBlock, self).__init__() self.conv1 = conv3x3(inplanes, (planes * 2), stride) self.bn1 = nn.BatchNorm2d((planes * 2)) self.relu = nn.ReLU(inplace=True) self.conv2 = conv3x3((planes * 2), (planes * 2), groups=num_group) self.bn2 = nn.BatchNorm2d((planes * 2)) self.downsample = downsample self.stride = stride def forward(self, x): residual = x out = self.conv1(x) out = self.bn1(out) out = self.relu(out) out = self.conv2(out) out = self.bn2(out) if (self.downsample is not None): residual = self.downsample(x) out += residual out = self.relu(out) return out
def test_serialization(explainer_type, model, masker, data, rtol=1e-05, atol=1e-08, **kwargs): explainer_kwargs = {k: v for (k, v) in kwargs.items() if (k in ['algorithm'])} explainer_original = explainer_type(model, masker, **explainer_kwargs) shap_values_original = explainer_original(data[:1]) with tempfile.TemporaryFile() as temp_serialization_file: save_kwargs = {k: v for (k, v) in kwargs.items() if (k in ['model_saver', 'masker_saver'])} explainer_original.save(temp_serialization_file, **save_kwargs) temp_serialization_file.seek(0) load_kwargs = {k: v for (k, v) in kwargs.items() if (k in ['model_loader', 'masker_loader'])} explainer_new = explainer_type.load(temp_serialization_file, **load_kwargs) call_kwargs = {k: v for (k, v) in kwargs.items() if (k in ['max_evals'])} shap_values_new = explainer_new(data[:1], **call_kwargs) assert np.allclose(shap_values_original.base_values, shap_values_new.base_values, rtol=rtol, atol=atol) assert np.allclose(shap_values_original[0].values, shap_values_new[0].values, rtol=rtol, atol=atol) assert isinstance(explainer_original, type(explainer_new)) assert isinstance(explainer_original.masker, type(explainer_new.masker))
class BNForwardFoldingNetTest(BasePytorchTest): def __init__(self, unit_test, test_layer, fold_applied=True, add_bn=False, is_dw=False): super().__init__(unit_test, float_reconstruction_error=1e-06, val_batch_size=2) self.test_layer = test_layer self.bn_layer = nn.BatchNorm2d self.fold_applied = fold_applied self.add_bn = add_bn self.is_dw = is_dw def create_feature_network(self, input_shape): return BNForwardFoldingNet(self.test_layer, self.add_bn, self.is_dw) def get_tpc(self): return {'no_quantization': super().get_tpc()['no_quantization']} def get_core_configs(self): return {'no_quantization': super().get_core_configs()['no_quantization']} def compare(self, quantized_models, float_model, input_x=None, quantization_info=None): set_model(float_model) quant_model = quantized_models['no_quantization'] set_model(quant_model) if self.is_dw: is_bn_in_model = (sum([(type(module) is nn.Conv2d) for (name, module) in float_model.named_modules()]) == sum([(type(module) is nn.Conv2d) for (name, module) in quant_model.named_modules()])) else: is_bn_in_model = (nn.BatchNorm2d in [type(module) for (name, module) in quant_model.named_modules()]) self.unit_test.assertTrue((self.fold_applied is not is_bn_in_model)) self.unit_test.assertEqual(input_x[0].shape[0], 2, 'Expecting batch of size 2 for BN folding test.') out_float = torch_tensor_to_numpy(float_model(*input_x)) out_quant = torch_tensor_to_numpy(quant_model(*input_x)) (norm_mse, _, max_error, _) = normalized_mse(out_float, out_quant) self.unit_test.assertTrue((np.isclose(norm_mse[0], 0, atol=1e-05) or np.isclose(norm_mse[1], 0, atol=1e-05))) self.unit_test.assertTrue((np.isclose(max_error[0], 0, atol=0.0001) or np.isclose(max_error[1], 0, atol=0.0001)))
def compute_error(model_file, data_loader, cuda_on=False, name=''): model = torch.load(model_file) if cuda_on: model.cuda() model.eval() test_loss = 0 correct = 0 for (data, target) in data_loader: if cuda_on: (data, target) = (data.cuda(), target.cuda()) (data, target) = (Variable(data, volatile=True), Variable(target)) output = model(data) test_loss += F.nll_loss(output, target, size_average=False).data[0] pred = output.data.max(1, keepdim=True)[1] correct += pred.eq(target.data.view_as(pred)).cpu().sum() test_loss /= len(data_loader.dataset) print((name + 'Average loss: {:.4f}, Accuracy: {}/{} ({:.0f}%)'.format(test_loss, correct, len(data_loader.dataset), ((100.0 * correct) / len(data_loader.dataset)))))
class NeumannDirichlet(CompositeBase): def __init__(self, N, quad='GC', bc=(0, 0), domain=((- 1), 1), dtype=float, padding_factor=1, dealias_direct=False, coordinates=None, **kw): if isinstance(bc, (tuple, list)): bc = BoundaryConditions({'left': {'N': bc[0]}, 'right': {'D': bc[1]}}, domain=domain) CompositeBase.__init__(self, N, quad=quad, domain=domain, dtype=dtype, bc=bc, padding_factor=padding_factor, dealias_direct=dealias_direct, coordinates=coordinates) self._stencil = {0: 1, 1: ((- ((4 * n) + 4)) / (((2 * (n ** 2)) + (6 * n)) + 5)), 2: ((- (((2 * (n ** 2)) + (2 * n)) + 1)) / (((2 * (n ** 2)) + (6 * n)) + 5))} def boundary_condition(): return 'NeumannDirichlet' def short_name(): return 'ND'
def prepare_test(recipe_folder='tests/recipes', script_field='Script_file', hparam_field='Hparam_file', test_field='test_debug_flags', check_field='test_debug_checks', download_field='test_download', message_field='test_message', filters_fields=[], filters=[]): test_script = {} test_hparam = {} test_flag = {} test_check = {} test_download = {} test_message = {} print(f' filters_fields={filters_fields} => filters={filters}') for recipe_csvfile in os.listdir(recipe_folder): if (recipe_csvfile in __skip_list): continue print(f'Loading recipes from: {recipe_csvfile}') with open(os.path.join(recipe_folder, recipe_csvfile), newline='') as csvf: reader = csv.DictReader(csvf, delimiter=',', skipinitialspace=True) for (row_id, row) in enumerate(reader): recipe_id = f'{recipe_csvfile[:(- 4)]}_row_{(row_id + 2):02d}' if (not check_row_for_test(row, filters_fields, filters, test_field)): print(f' Skipped {recipe_id}') continue test_script[recipe_id] = row[script_field].strip() test_hparam[recipe_id] = row[hparam_field].strip() test_flag[recipe_id] = row[test_field].strip() test_check[recipe_id] = row[check_field].strip() if (download_field in row): test_download[recipe_id] = row[download_field].strip() if (message_field in row): test_message[recipe_id] = row[message_field].strip() return (test_script, test_hparam, test_flag, test_check, test_download, test_message)
def DFG_c(root_node, index_to_code, states): assignment = ['assignment_expression'] def_statement = ['init_declatator', 'pointer_declarator', 'array_declarator'] increment_statement = ['update_expression'] if_statement = ['if_statement', 'else'] for_statement = ['for_statement'] while_statement = ['while_statement'] parameter_statement = ['parameter_declaration'] do_first_statement = [] states = states.copy() if (((len(root_node.children) == 0) or (root_node.type == 'string')) and (root_node.type != 'comment')): (idx, code) = index_to_code[(root_node.start_point, root_node.end_point)] if ((root_node.type == code) or ((root_node.parent.type == 'function_declarator') and root_node)): return ([], states) elif (code in states): return ([(code, idx, 'comesFrom', [code], states[code].copy())], states) elif (root_node.type == 'identifier'): if (root_node.parent.type == 'declaration'): states[code] = [idx] return ([(code, idx, 'comesFrom', [], [])], states) return ([], states) else: return ([], states) elif (root_node.type in def_statement): if (root_node.parent.type == 'function_definition'): while ((root_node.type == 'pointer_declarator') and (root_node.child_by_field_name('declarator').type == 'pointer_declarator')): root_node = root_node.child_by_field_name('declarator') DFG = [] for child in root_node.children: if (child.type not in do_first_statement): (temp, states) = DFG_c(child, index_to_code, states) DFG += temp return (sorted(DFG, key=(lambda x: x[1])), states) name = root_node.child_by_field_name('declarator') value = root_node.child_by_field_name('value') DFG = [] if (value is None): indexs = tree_to_variable_index(name, index_to_code) for index in indexs: (idx, code) = index_to_code[index] DFG.append((code, idx, 'comesFrom', [], [])) states[code] = [idx] return (sorted(DFG, key=(lambda x: x[1])), states) else: name_indexs = tree_to_variable_index(name, index_to_code) value_indexs = tree_to_variable_index(value, index_to_code) (temp, states) = DFG_c(value, index_to_code, states) DFG += temp for index1 in name_indexs: (idx1, code1) = index_to_code[index1] for index2 in value_indexs: (idx2, code2) = index_to_code[index2] DFG.append((code1, idx1, 'comesFrom', [code2], [idx2])) states[code1] = [idx1] return (sorted(DFG, key=(lambda x: x[1])), states) elif (root_node.type in assignment): return ([], states) elif (root_node.type in increment_statement): DFG = [] indexs = tree_to_variable_index(root_node, index_to_code) for index1 in indexs: (idx1, code1) = index_to_code[index1] for index2 in indexs: (idx2, code2) = index_to_code[index2] DFG.append((code1, idx1, 'computedFrom', [code2], [idx2])) states[code1] = [idx1] return (sorted(DFG, key=(lambda x: x[1])), states) elif (root_node.type in if_statement): DFG = [] current_states = states.copy() others_states = [] flag = False tag = False if ('else' in root_node.type): tag = True for child in root_node.children: if ('else' in child.type): tag = True if ((child.type not in if_statement) and (flag is False)): (temp, current_states) = DFG_c(child, index_to_code, current_states) DFG += temp else: flag = True (temp, new_states) = DFG_c(child, index_to_code, states) DFG += temp others_states.append(new_states) others_states.append(current_states) if (tag is False): others_states.append(states) new_states = {} for dic in others_states: for key in dic: if (key not in new_states): new_states[key] = dic[key].copy() else: new_states[key] += dic[key] for key in states: if (key not in new_states): new_states[key] = states[key] else: new_states[key] += states[key] for key in new_states: new_states[key] = sorted(list(set(new_states[key]))) return (sorted(DFG, key=(lambda x: x[1])), new_states) elif (root_node.type in for_statement): DFG = [] for child in root_node.children: (temp, states) = DFG_c(child, index_to_code, states) DFG += temp flag = False for child in root_node.children: if flag: (temp, states) = DFG_c(child, index_to_code, states) DFG += temp elif (child.type == 'variable_declaration'): flag = True dic = {} for x in DFG: if ((x[0], x[1], x[2]) not in dic): dic[(x[0], x[1], x[2])] = [x[3], x[4]] else: dic[(x[0], x[1], x[2])][0] = list(set((dic[(x[0], x[1], x[2])][0] + x[3]))) dic[(x[0], x[1], x[2])][1] = sorted(list(set((dic[(x[0], x[1], x[2])][1] + x[4])))) DFG = [(x[0], x[1], x[2], y[0], y[1]) for (x, y) in sorted(dic.items(), key=(lambda t: t[0][1]))] return (sorted(DFG, key=(lambda x: x[1])), states) elif (root_node.type in while_statement): DFG = [] for i in range(2): for child in root_node.children: (temp, states) = DFG_c(child, index_to_code, states) DFG += temp dic = {} for x in DFG: if ((x[0], x[1], x[2]) not in dic): dic[(x[0], x[1], x[2])] = [x[3], x[4]] else: dic[(x[0], x[1], x[2])][0] = list(set((dic[(x[0], x[1], x[2])][0] + x[3]))) dic[(x[0], x[1], x[2])][1] = sorted(list(set((dic[(x[0], x[1], x[2])][1] + x[4])))) DFG = [(x[0], x[1], x[2], y[0], y[1]) for (x, y) in sorted(dic.items(), key=(lambda t: t[0][1]))] return (sorted(DFG, key=(lambda x: x[1])), states) elif (root_node.type in parameter_statement): child = root_node.child_by_field_name('declarator') if (not child): return ([], states) while (child.type != 'identifier'): if (child.type == 'parenthesized_declarator'): child = child.children[1] else: child = child.child_by_field_name('declarator') if (not child): return ([], states) (idx, code) = index_to_code[(child.start_point, child.end_point)] states[code] = [idx] return ([(code, idx, 'comesFrom', [], [])], states) else: DFG = [] for child in root_node.children: if (child.type not in do_first_statement): (temp, states) = DFG_c(child, index_to_code, states) DFG += temp return (sorted(DFG, key=(lambda x: x[1])), states)
def system_worker(system_id, story_generator_class, request_queue, result_queue): story_generator = story_generator_class(system_id) while True: worker_request = request_queue.get() action = worker_request['action'] if (action == 'generate'): result = story_generator.generate_response(worker_request['topic'], kw_temp=worker_request['kw_temp'], story_temp=worker_request['story_temp'], dedup=worker_request.get('dedup', None), max_len=worker_request.get('max_len', None), use_gold_titles=worker_request.get('use_gold_titles', None)) elif (action == 'generate_storyline'): result = story_generator.generate_storyline(worker_request['topic'], kw_temp=worker_request['kw_temp'], dedup=worker_request.get('dedup', None), max_len=worker_request.get('max_len', None), use_gold_titles=worker_request.get('use_gold_titles', None)) elif (action == 'collab_storyline'): result = story_generator.collab_storyline(worker_request['topic'], worker_request['storyline'], kw_temp=worker_request.get('kw_temp', None), dedup=worker_request.get('dedup', None), max_len=worker_request.get('max_len', None)) elif (action == 'generate_interactive_story'): result = story_generator.generate_interactive_story(worker_request['topic'], worker_request['storyline'], worker_request['story_sentences'], story_temp=worker_request.get('story_temp', None), only_one=worker_request['only_one']) elif (action == 'generate_story'): result = story_generator.generate_story(worker_request['topic'], worker_request['storyline'], story_temp=worker_request.get('story_temp', None)) else: result = {system_id: 'internal error'} result_queue.put(result)
def train(epoch): global lr, train_acc model.train() batch_idx = 1 total_loss = 0 correct = 0 pred = np.array([]) X_train = audio_features[(train_dep_idxs + train_non_idxs)] Y_train = audio_targets[(train_dep_idxs + train_non_idxs)] for i in range(0, X_train.shape[0], config['batch_size']): if ((i + config['batch_size']) > X_train.shape[0]): (x, y) = (X_train[i:], Y_train[i:]) else: (x, y) = (X_train[i:(i + config['batch_size'])], Y_train[i:(i + config['batch_size'])]) if config['cuda']: (x, y) = (Variable(torch.from_numpy(x).type(torch.FloatTensor), requires_grad=True).cuda(), Variable(torch.from_numpy(y)).cuda()) else: (x, y) = (Variable(torch.from_numpy(x).type(torch.FloatTensor), requires_grad=True), Variable(torch.from_numpy(y)).type(torch.FloatTensor)) optimizer.zero_grad() output = model(x) loss = criterion(output, y.view_as(output)) loss.backward() optimizer.step() batch_idx += 1 pred = np.hstack((pred, output.flatten().detach().numpy())) total_loss += loss.item() train_mae = mean_absolute_error(Y_train, pred) print('Train Epoch: {:2d}\t Learning rate: {:.4f}\t Loss: {:.4f}\t MAE: {:.4f}\t RMSE: {:.4f}\n '.format((epoch + 1), config['learning_rate'], total_loss, train_mae, np.sqrt(mean_squared_error(Y_train, pred)))) return train_mae
def miniBatchStdDev(x, subGroupSize=4): size = x.size() subGroupSize = min(size[0], subGroupSize) if ((size[0] % subGroupSize) != 0): subGroupSize = size[0] G = int((size[0] / subGroupSize)) if (subGroupSize > 1): y = x.view((- 1), subGroupSize, size[1], size[2], size[3]) y = torch.var(y, 1) y = torch.sqrt((y + 1e-08)) y = y.view(G, (- 1)) y = torch.mean(y, 1).view(G, 1) y = y.expand(G, (size[2] * size[3])).view((G, 1, 1, size[2], size[3])) y = y.expand(G, subGroupSize, (- 1), (- 1), (- 1)) y = y.contiguous().view(((- 1), 1, size[2], size[3])) else: y = torch.zeros(x.size(0), 1, x.size(2), x.size(3), device=x.device) return torch.cat([x, y], dim=1)
def get_loaders_and_models(args, device): return get_facetranslation_latent_conv_perceptual(args, device)
_LAYERS.register_module() class DropPath(nn.Module): def __init__(self, drop_prob=0.1): super(DropPath, self).__init__() self.drop_prob = drop_prob def forward(self, x): return drop_path(x, self.drop_prob, self.training)
def update_lr(optimizer, n_vals_without_improvement): global ndecs if ((ndecs == 0) and (n_vals_without_improvement > (args.early_stopping // 3))): for param_group in optimizer.param_groups: param_group['lr'] = (args.lr / 10) ndecs = 1 elif ((ndecs == 1) and (n_vals_without_improvement > ((args.early_stopping // 3) * 2))): for param_group in optimizer.param_groups: param_group['lr'] = (args.lr / 100) ndecs = 2 else: for param_group in optimizer.param_groups: param_group['lr'] = (args.lr / (10 ** ndecs))
class dotdict(dict): __getattr__ = dict.get __setattr__ = dict.__setitem__ __delattr__ = dict.__delitem__
class HfDeepSpeedConfig(): def __init__(self, config_file_or_dict): set_hf_deepspeed_config(self) dep_version_check('deepspeed') if isinstance(config_file_or_dict, dict): config = deepcopy(config_file_or_dict) elif isinstance(config_file_or_dict, str): with io.open(config_file_or_dict, 'r', encoding='utf-8') as f: config = json.load(f) else: raise ValueError('expecting either a path to a DeepSpeed config file or a pre-populated dict') self.config = config self._stage = self.get_value('zero_optimization.stage', (- 1)) self._offload = False if (self.is_zero2() or self.is_zero3()): offload_devices_valid = set(['cpu', 'nvme']) offload_devices = set([self.get_value('zero_optimization.offload_optimizer.device'), self.get_value('zero_optimization.offload_param.device')]) if (len((offload_devices & offload_devices_valid)) > 0): self._offload = True def find_config_node(self, ds_key_long): config = self.config nodes = ds_key_long.split('.') ds_key = nodes.pop() for node in nodes: config = config.get(node) if (config is None): return (None, ds_key) return (config, ds_key) def get_value(self, ds_key_long, default=None): (config, ds_key) = self.find_config_node(ds_key_long) if (config is None): return default return config.get(ds_key, default) def del_config_sub_tree(self, ds_key_long, must_exist=False): config = self.config nodes = ds_key_long.split('.') for node in nodes: parent_config = config config = config.get(node) if (config is None): if must_exist: raise ValueError(f"Can't find {ds_key_long} entry in the config: {self.config}") else: return if (parent_config is not None): parent_config.pop(node) def is_true(self, ds_key_long): value = self.get_value(ds_key_long) return (False if (value is None) else bool(value)) def is_false(self, ds_key_long): value = self.get_value(ds_key_long) return (False if (value is None) else (not bool(value))) def is_zero2(self): return (self._stage == 2) def is_zero3(self): return (self._stage == 3) def is_offload(self): return self._offload
def get_unique_graph(tops, check_types=None, none_if_empty=False): if isinstance(tops, tf_ops.Graph): return tops if (not is_iterable(tops)): raise TypeError('{} is not iterable'.format(type(tops))) if (check_types is None): check_types = (tf_ops.Operation, tf_ops.Tensor) elif (not is_iterable(check_types)): check_types = (check_types,) g = None for op in tops: if (not isinstance(op, check_types)): raise TypeError('Expected a type in ({}), got: {}'.format(', '.join([str(t) for t in check_types]), type(op))) if (g is None): g = op.graph elif (g is not op.graph): raise ValueError('Operation {} does not belong to given graph'.format(op)) if ((g is None) and (not none_if_empty)): raise ValueError("Can't find the unique graph of an empty list") return g
def register_Ns3CallbackImplBase_methods(root_module, cls): cls.add_constructor([]) cls.add_constructor([param('ns3::CallbackImplBase const &', 'arg0')]) cls.add_method('GetTypeid', 'std::string', [], is_pure_virtual=True, is_const=True, is_virtual=True) cls.add_method('IsEqual', 'bool', [param('ns3::Ptr< ns3::CallbackImplBase const >', 'other')], is_pure_virtual=True, is_const=True, is_virtual=True) cls.add_method('Demangle', 'std::string', [param('std::string const &', 'mangled')], is_static=True, visibility='protected') cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::ObjectBase*']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'void']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::WifiMacHeader const&']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::Ipv4Address']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'unsigned char']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::Ptr<ns3::Socket> ']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'bool']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::Address const&']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'unsigned int']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::Ptr<ns3::NetDevice> ']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::Ptr<ns3::Packet const> ']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'unsigned short']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::NetDevice::PacketType']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::Ipv4Header const&']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::Ptr<ns3::Ipv4> ']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::Ipv4L3Protocol::DropReason']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::Ptr<ns3::Packet> ']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::Ptr<ns3::Ipv4Route> ']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::Mac48Address']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::Ptr<ns3::MobilityModel const> ']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::WifiTxVector']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::MpduInfo']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::SignalNoiseDbm']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::WifiMacHeader const*']) cls.add_method('GetCppTypeid', 'std::string', [], is_static=True, visibility='protected', template_parameters=[u'ns3::dsr::DsrOptionSRHeader const&']) return
def istft_backward(grad_inputs, inputs, input_shapes, outputs, output_shapes, window_size, stride, fft_size, window_type='hanning', center=True, pad_mode='reflect', as_stft_backward=False): dy = grad_inputs[0] (dx_r, dx_i) = F.stft(dy, window_size, stride, fft_size, window_type, center, pad_mode, as_istft_backward=(not as_stft_backward)) return (dx_r, dx_i)
def main(): parser = argparse.ArgumentParser(description='PyTorch MNIST Conlinual leanring Example') parser.add_argument('--model', type=str, default='MLP_CL', help='Model name: MLP_CL') parser.add_argument('--bnmomentum', type=float, default=0.15, help='BN layer momentum value') parser.add_argument('--optim', type=str, default='BayesBiNN', help='Optimizer: BayesBiNN') parser.add_argument('--val-split', type=float, default=0, help='Random validation set ratio') parser.add_argument('--criterion', type=str, default='cross-entropy', help='loss funcion: square-hinge or cross-entropy') parser.add_argument('--batch-size', type=int, default=100, metavar='N', help='input batch size for training (default: 64)') parser.add_argument('--test-batch-size', type=int, default=1000, metavar='N', help='input batch size for testing (default: 1000)') parser.add_argument('--train-samples', type=int, default=1, metavar='N', help='number of Monte Carlo samples used in BayesBiNN (default: 1)') parser.add_argument('--test-samples', type=int, default=100, metavar='N', help='number of Monte Carlo samples used in evaluation for BayesBiNN (default: 1), if 0, point estimate using meanis applied, which is similar to the Bop optimizer') parser.add_argument('--epochs', type=int, default=100, metavar='N', help='number of epochs to train (default: 10)') parser.add_argument('--lr', type=float, default=0.001, metavar='LR', help='learning rate (default: 0.001)') parser.add_argument('--lr-end', type=float, default=1e-16, metavar='LR-end', help='learning rate (default: 0.01)') parser.add_argument('--lr-decay', type=float, default=0.9, metavar='LR-decay', help='learning rated decay factor for each epoch (default: 0.9)') parser.add_argument('--decay-steps', type=int, default=1, metavar='N', help='LR rate decay steps (default: 1)') parser.add_argument('--momentum', type=float, default=0.0, metavar='M', help='BayesBiNN momentum (default: 0.0)') parser.add_argument('--data-augmentation', action='store_true', default=False, help='Enable data augmentation') parser.add_argument('--log-interval', type=int, default=5000, metavar='N', help='how many batches to wait before logging training status') parser.add_argument('--save-model', action='store_true', default=False, help='For Saving the current Model') parser.add_argument('--experiment-id', type=int, default=0, help='Experiment ID for log files (int)') parser.add_argument('--no-cuda', action='store_true', default=False, help='disables CUDA training') parser.add_argument('--seed', type=int, default=1, metavar='S', help='random seed (default: 1)') parser.add_argument('--lrschedular', type=str, default='Cosine', help='Mstep,Expo,Cosine') parser.add_argument('--drop-prob', type=float, default=0.0, help='dropout rate') parser.add_argument('--trainset_scale', type=int, default=1, help='scale of training set') parser.add_argument('--lamda', type=float, default=10, metavar='lamda-init', help='initial mean value of the natural parameter lamda(default: 10)') parser.add_argument('--lamda-std', type=float, default=0, metavar='lamda-init', help='linitial std value of the natural parameter lamda(default: 0)') parser.add_argument('--temperature', type=float, default=0.01, metavar='temperature', help='initial temperature for BayesBiNN (default: 1)') parser.add_argument('--kl-reweight', type=float, default=1.0, metavar='min temperature', help='initial temperature for BayesBiNN (default: 1)') parser.add_argument('--bn-affine', type=float, default=0, metavar='bn-affine', help='whether there is bn learnable parameters, 1: learnable, 0: no (default: 1)') parser.add_argument('--num-tasks', type=int, default=5, metavar='num-tasks', help='number of tasks for continual learning') parser.add_argument('--scenario', type=str, default='class', choices=['task', 'domain', 'class']) parser.add_argument('--experiment', type=str, default='permMNIST', choices=['permMNIST', 'splitMNIST']) args = parser.parse_args() args.use_cuda = ((not args.no_cuda) and torch.cuda.is_available()) torch.manual_seed((args.seed + args.experiment_id)) np.random.seed((args.seed + args.experiment_id)) now = time.strftime('%Y_%m_%d_%H_%M_%S', time.localtime(time.time())) args.out_dir = os.path.join('./outputs', 'mnist_CL_{}_{}_lr{}_{}'.format(args.model, args.optim, args.lr, now)) os.makedirs(args.out_dir, exist_ok=True) config_save_path = os.path.join(args.out_dir, 'configs', 'config_{}.json'.format(args.experiment_id)) os.makedirs(os.path.dirname(config_save_path), exist_ok=True) with open(config_save_path, 'w') as f: json.dump(args.__dict__, f, indent=2) args.device = torch.device(('cuda' if args.use_cuda else 'cpu')) print('Running on', args.device) print('') for (key, val) in vars(args).items(): print('{}: {}'.format(key, val)) print('\n') if args.data_augmentation: transform_train = transforms.Compose([transforms.RandomCrop(32, padding=4), transforms.RandomHorizontalFlip(), transforms.ToTensor(), transforms.Normalize((0.1307,), (0.3081,))]) else: transform_train = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.1307,), (0.3081,))]) transform_test = transforms.Compose([transforms.ToTensor(), transforms.Normalize((0.1307,), (0.3081,))]) kwargs = ({'num_workers': 2, 'pin_memory': True, 'drop_last': True} if args.use_cuda else {}) train_dataset = datasets.MNIST('./data', train=True, download=True, transform=transform_train) if ((args.val_split > 0) and (args.val_split < 1)): val_dataset = datasets.MNIST('./data', train=True, download=True, transform=transform_train) num_train = len(train_dataset) indices = list(range(num_train)) split = int(np.floor((args.val_split * num_train))) np.random.shuffle(indices) (train_idx, val_idx) = (indices[split:], indices[:split]) train_sampler = SubsetRandomSampler(train_idx) val_sampler = SubsetRandomSampler(val_idx) train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=args.batch_size, sampler=train_sampler, **kwargs) val_loader = torch.utils.data.DataLoader(val_dataset, batch_size=args.batch_size, sampler=val_sampler, **kwargs) print('{} train and {} validation datapoints.'.format(len(train_loader.sampler), len(val_loader.sampler))) else: train_loader = torch.utils.data.DataLoader(train_dataset, batch_size=args.batch_size, shuffle=True, **kwargs) val_loader = None print('{} train and {} validation datapoints.'.format(len(train_loader.sampler), 0)) test_dataset = datasets.MNIST('./data', train=False, transform=transform_test) test_loader = torch.utils.data.DataLoader(test_dataset, batch_size=args.test_batch_size, shuffle=True, **kwargs) print('{} test datapoints.\n'.format(len(test_loader.sampler))) (in_features, out_features) = ((28 * 28), 10) if (args.model == 'MLP_CL'): num_units = 100 model = MLP_CL_h100(in_features, out_features, num_units, eps=0.0001, momentum=args.bnmomentum, batch_affine=(args.bn_affine == 1)) else: raise ValueError('Please select a network out of {MLP, BinaryConnect, BinaryNet}') print(model) model = model.to(args.device) if (args.optim == 'Adam'): optimizer = optim.Adam(model.parameters(), lr=args.lr) elif (args.optim == 'BayesBiNN'): effective_trainsize = (len(train_loader.sampler) * args.trainset_scale) optimizer = BayesBiNN(model, lamda_init=args.lamda, lamda_std=args.lamda_std, temperature=args.temperature, train_set_size=effective_trainsize, lr=args.lr, betas=args.momentum, num_samples=args.train_samples, reweight=args.kl_reweight) if (args.criterion == 'square-hinge'): criterion = SquaredHingeLoss() elif (args.criterion == 'cross-entropy'): criterion = nn.CrossEntropyLoss() else: raise ValueError('Please select loss criterion in {square-hinge, cross-entropy}') start = time.time() permute_list = [] for task in range(args.num_tasks): permute_list.append(torch.Tensor(np.random.permutation(784).astype(np.float64)).long()) test_results_save = [] test_results_record_save = [] for task in range(args.num_tasks): print('') print(('Current Training task %d' % task)) if (not (task == 0)): optimizer.state['prior_lamda'] = optimizer.state['lamda'] permute_state = permute_list[task] for param_group in optimizer.param_groups: param_group['lr'] = args.lr results = train_model_cl(args, model, [train_loader, test_loader], criterion, optimizer, permute_list, task) (model, optimizer, train_loss, train_acc, test_results_record) = results (test_loss, test_accuracy) = test_model_cl(args, model, test_loader, criterion, optimizer, permute=permute_state) print('test acc: {}'.format(test_accuracy)) test_acc_total = 0 for test_id in range((task + 1)): permute_state_id = permute_list[test_id] if (test_loader is not None): (test_loss, test_accuracy) = test_model_cl(args, model, test_loader, criterion, optimizer, permute=permute_state_id) test_acc_total += test_accuracy print(('## Individual Task[%d], Test Loss: %f & Test Accuracy: %f' % (test_id, test_loss, test_accuracy))) print('') test_acc_average = (test_acc_total / (task + 1)) print(('## After Task[%d],Averaged Test Accuracy: %f' % (task, test_acc_average))) test_results_save.append(test_acc_average) test_results_record_save.append(test_results_record) save_train_history_CL(args, test_results_save) array_record = np.array(test_results_record_save) np.save('CL_test_record_seed0.npy', array_record) print(test_results_save) time_total = timeSince(start) print('Task completed in {:.0f}m {:.0f}s'.format((time_total // 60), (time_total % 60)))
class TestCreationConfiguration(): max_recursion: int = 10 max_delta: int = 20 max_int: int = 2048 string_length: int = 20 bytes_length: int = 20 collection_size: int = 5 primitive_reuse_probability: float = 0.5 object_reuse_probability: float = 0.9 none_weight: float = 1 any_weight: float = 5 original_type_weight: float = 5 type_tracing_weight: float = 10 type4py_weight: float = 10 type_tracing_kept_guesses: int = 2 wrap_var_param_type_probability: float = 0.7 negate_type: float = 0.1 skip_optional_parameter_probability: float = 0.7 max_attempts: int = 1000 insertion_uut: float = 0.5 max_size: int = 100 use_random_object_for_call: float = 0.1
_fl_task(model='model', data_loader='train_loader', device='device', optimizer='optimizer') def train(model, train_loader, device, optimizer): device = 'hpu' total_loss = [] train_loader = tqdm.tqdm(train_loader, desc='train') model.train() model.to(device) criterion = nn.CrossEntropyLoss() for (inputs, targets) in train_loader: optimizer.zero_grad() outputs = model(inputs.to(device)) targets = torch.squeeze(targets, 1).long().to(device) loss = criterion(outputs, targets) total_loss.append(loss.item()) loss.backward() htcore.mark_step() optimizer.step() htcore.mark_step() return {'train_loss': np.mean(total_loss)}
class IndexedFileWriter(object): def __init__(self, path): self.f = open(path, 'wb') self.index_f = open((path + '.index'), 'wb') def append(self, record): offset = self.f.tell() self.f.write(record) self.index_f.write(struct.pack('<Q', offset)) def close(self): self.f.close() self.index_f.close()
class SPTokenizer(Tokenizer): def __init__(self, model_path) -> None: self.sp = spm.SentencePieceProcessor() self.sp.Load(model_path) def encode(self, text): return self.sp.EncodeAsPieces(text)
def weights_init(m): classname = m.__class__.__name__ if (classname.find('Linear') != (- 1)): if (hasattr(m, 'weight') and (m.weight is not None)): init_weight(m.weight) if (hasattr(m, 'bias') and (m.bias is not None)): init_bias(m.bias) elif (classname.find('AdaptiveEmbedding') != (- 1)): if hasattr(m, 'emb_projs'): for i in range(len(m.emb_projs)): if (m.emb_projs[i] is not None): nn.init.normal_(m.emb_projs[i], 0.0, args.proj_init_std) elif (classname.find('Embedding') != (- 1)): if hasattr(m, 'weight'): init_weight(m.weight) elif (classname.find('ProjectedAdaptiveLogSoftmax') != (- 1)): if (hasattr(m, 'cluster_weight') and (m.cluster_weight is not None)): init_weight(m.cluster_weight) if (hasattr(m, 'cluster_bias') and (m.cluster_bias is not None)): init_bias(m.cluster_bias) if hasattr(m, 'out_projs'): for i in range(len(m.out_projs)): if (m.out_projs[i] is not None): nn.init.normal_(m.out_projs[i], 0.0, args.proj_init_std) elif (classname.find('LayerNorm') != (- 1)): if hasattr(m, 'weight'): nn.init.normal_(m.weight, 1.0, args.init_std) if (hasattr(m, 'bias') and (m.bias is not None)): init_bias(m.bias) elif (classname.find('TransformerLM') != (- 1)): if hasattr(m, 'r_emb'): init_weight(m.r_emb) if hasattr(m, 'r_w_bias'): init_weight(m.r_w_bias) if hasattr(m, 'r_r_bias'): init_weight(m.r_r_bias) if hasattr(m, 'r_bias'): init_bias(m.r_bias)
def get_contact_map_paths(names, images_root='data/SCOPe/pdbstyle-2.06/'): cmap_paths = glob.glob((images_root + '*/*.png')) cmap_dict = {os.path.basename(path)[:7]: path for path in cmap_paths} paths = [] for name in names: if (name not in cmap_dict): name = ('d' + name[1:]) path = cmap_dict[name] paths.append(path) return paths
def check_lemmas(train_file): with open(train_file) as fin: for line in fin: line = line.strip() if ((not line) or line.startswith('#')): continue pieces = line.split('\t') word = pieces[1].lower().strip() lemma = pieces[2].lower().strip() if ((not lemma) or (lemma == '_') or (lemma == '-')): continue if (word == lemma): continue return True return False
def get_args_and_hdf5_file(output_mode, config_file_index: int): output_name = ('run_%s_%d' % (output_mode, config_file_index)) parameters = [sys.executable, 'volnet/train_volnet.py', CONFIG_FILES[config_file_index], '--volume_filenames', VOLUME_FILES, '--time_keyframes', f'{VOLUME_TIMESTEP}:{(VOLUME_TIMESTEP + 1)}:1', '--time_train', f'{VOLUME_TIMESTEP}:{(VOLUME_TIMESTEP + 1)}:1', '--time_val', f'{VOLUME_TIMESTEP}:{(VOLUME_TIMESTEP + 1)}:1', '--train:mode', 'world', '--train:samples', '256**3', '--train:batchsize', '64*64*128', '--train:sampler_importance', '0.01', '--val:copy_and_split', '--outputmode', ('%s:direct' % output_mode), '--lossmode', ('%s' % output_mode), '-l1', '1', '--lr_step', '100', '-i', '200', '--fouriercount', str(((BEST_NETWORK[0] - 4) // 2)), '--fourierstd', '1.0', '--activation', BEST_ACTIVATION, '--layers', ':'.join(([str(BEST_NETWORK[0])] * (BEST_NETWORK[1] - 1))), '--volumetric_features_resolution', str(GRID_RESOLUTION), '--volumetric_features_channels', str(GRID_CHANNELS), '--logdir', 'volnet/results/eval_world_DensityVsColorGrid/log', '--modeldir', 'volnet/results/eval_world_DensityVsColorGrid/model', '--hdf5dir', 'volnet/results/eval_world_DensityVsColorGrid/hdf5', '--name', output_name, '--save_frequency', '50'] hdf5_file = (('volnet/results/eval_world_DensityVsColorGrid/hdf5/' + output_name) + '.hdf5') return (parameters, hdf5_file)
class My_Dataset(): def __init__(self, folder, transform=None, int_class=False): self.img_paths = np.asarray(sorted([str(i) for i in sorted(p(folder).glob('**/*')) if (i.suffix in ['.jpg', '.JPG', '.jpeg', '.JPEG', '.png', '.PNG', '.tiff', '.TIFF', '.bmp', '.BMP', '.gif', '.GIF'])])) self.classes = sorted([i.name for i in p(self.img_paths[0]).parents[1].iterdir() if i.is_dir()]) self.transform = transform self.dict = dict(zip(self.classes, torch.arange(len(self.classes)))) def __getitem__(self, index): img_path = self.img_paths[index] img = Image.open(img_path) label = self.dict[p(img_path).parts[(- 2)]] if (self.transform != None): img = self.transform(img) return (img, label) def __len__(self): return len(self.img_paths) def split_(self, val_ratio, transform=None): random.seed(10) (data_paths, data_num) = self.paths_in_every_class() self.img_paths = [] new_set = deepcopy(self) new_set.img_paths = [] new_set.transform = transform for (class_name, paths) in data_paths.items(): random.shuffle(paths) sub_num = int((data_num[class_name] * val_ratio)) new_set.img_paths.extend(paths[:sub_num]) self.img_paths.extend(paths[sub_num:]) self.img_paths = np.asarray(sorted(self.img_paths)) new_set.img_paths = np.asarray(sorted(new_set.img_paths)) return new_set def paths_in_every_class(self): data_path = {i: [] for i in self.classes} data_num = {i: 0 for i in self.classes} for path in self.img_paths: class_name = p(path).parts[(- 2)] data_path[class_name].append(path) data_num[class_name] += 1 return (data_path, data_num)
class BasicBlock(nn.Module): def __init__(self, in_planes, planes, stride=1): super(BasicBlock, self).__init__() self.conv1 = nn.Conv2d(in_planes, planes, kernel_size=3, stride=stride, padding=1, bias=False) self.bn1 = nn.BatchNorm2d(planes) self.conv2 = nn.Conv2d(planes, planes, kernel_size=3, stride=1, padding=1, bias=False) self.bn2 = nn.BatchNorm2d(planes) self.shortcut = nn.Sequential() if ((stride != 1) or (in_planes != planes)): self.shortcut = nn.Sequential(nn.Conv2d(in_planes, planes, kernel_size=1, stride=stride, bias=False), nn.BatchNorm2d(planes)) self.fc1 = nn.Conv2d(planes, (planes // 16), kernel_size=1) self.fc2 = nn.Conv2d((planes // 16), planes, kernel_size=1) def forward(self, x): out = F.tanh(self.bn1(self.conv1(x))) out = self.bn2(self.conv2(out)) w = F.avg_pool2d(out, out.size(2)) w = F.tanh(self.fc1(w)) w = F.tanh(self.fc2(w)) out = (out * w) out += self.shortcut(x) out = F.tanh(out) return out
.core .usefixtures('full_pandas_dataset') def test_get_encoder_empty_rules_list(full_pandas_dataset): encoder = DatasetLabelEncoder() user_item_features = ['user_id', 'item_id'] dataset_for_fit = Dataset(feature_schema=get_features(full_pandas_dataset).subset(user_item_features), interactions=full_pandas_dataset['interactions']) encoder_from_get = encoder.fit(dataset_for_fit).get_encoder(['timestamp', 'rating']) assert (encoder_from_get is None)
class MapFreeModuleToOneStep(pAdicModuleIsomorphism): def _call_(self, x): return self.codomain()(list(x)) def _call_with_args(self, x, args=(), kwds={}): return self.codomain()(list(x), *args, **kwds)
class Model(object): def __init__(self): self.image_size = 28 self.num_channels = 1 self.num_labels = 10 self.W_conv1 = self._weight_variable([5, 5, 1, 32]) self.b_conv1 = self._bias_variable([32]) self.W_conv2 = self._weight_variable([5, 5, 32, 64]) self.b_conv2 = self._bias_variable([64]) self.W_fc1 = self._weight_variable([((7 * 7) * 64), 1024]) self.b_fc1 = self._bias_variable([1024]) self.W_fc2 = self._weight_variable([1024, 10]) self.b_fc2 = self._bias_variable([10]) def predict(self, x_image): h_conv1 = tf.nn.relu((self._conv2d(x_image, self.W_conv1) + self.b_conv1)) h_pool1 = self._max_pool_2x2(h_conv1) h_conv2 = tf.nn.relu((self._conv2d(h_pool1, self.W_conv2) + self.b_conv2)) h_pool2 = self._max_pool_2x2(h_conv2) h_pool2_flat = tf.reshape(h_pool2, [(- 1), ((7 * 7) * 64)]) h_fc1 = tf.nn.relu((tf.matmul(h_pool2_flat, self.W_fc1) + self.b_fc1)) pre_softmax = (tf.matmul(h_fc1, self.W_fc2) + self.b_fc2) return pre_softmax def _weight_variable(shape): initial = tf.truncated_normal(shape, stddev=0.1) return tf.Variable(initial) def _bias_variable(shape): initial = tf.constant(0.1, shape=shape) return tf.Variable(initial) def _conv2d(x, W): return tf.nn.conv2d(x, W, strides=[1, 1, 1, 1], padding='SAME') def _max_pool_2x2(x): return tf.nn.max_pool(x, ksize=[1, 2, 2, 1], strides=[1, 2, 2, 1], padding='SAME')
class TestShapImage(unittest.TestCase): def setUp(self) -> None: batch_size = 128 num_classes = 10 epochs = 10 (img_rows, img_cols) = (28, 28) ((x_train, y_train), (x_test, y_test)) = tf.keras.datasets.mnist.load_data() if (tf.keras.backend.image_data_format() == 'channels_first'): x_train = x_train.reshape(x_train.shape[0], 1, img_rows, img_cols) x_test = x_test.reshape(x_test.shape[0], 1, img_rows, img_cols) input_shape = (1, img_rows, img_cols) else: x_train = x_train.reshape(x_train.shape[0], img_rows, img_cols, 1) x_test = x_test.reshape(x_test.shape[0], img_rows, img_cols, 1) input_shape = (img_rows, img_cols, 1) self.x_train = Image(x_train.astype('float32'), batched=True) self.x_test = Image(x_test.astype('float32'), batched=True) self.preprocess_func = (lambda x: np.expand_dims((x.to_numpy() / 255), axis=(- 1))) x_train = self.preprocess_func(self.x_train) x_test = self.preprocess_func(self.x_test) print('x_train shape:', x_train.shape) print(x_train.shape[0], 'train samples') print(x_test.shape[0], 'test samples') y_train = tf.keras.utils.to_categorical(y_train, num_classes) y_test = tf.keras.utils.to_categorical(y_test, num_classes) model = tf.keras.models.Sequential() model.add(tf.keras.layers.Conv2D(32, kernel_size=(3, 3), activation='relu', input_shape=input_shape)) model.add(tf.keras.layers.Conv2D(64, (3, 3), activation='relu')) model.add(tf.keras.layers.MaxPooling2D(pool_size=(2, 2))) model.add(tf.keras.layers.Dropout(0.1)) model.add(tf.keras.layers.Flatten()) model.add(tf.keras.layers.Dense(128, activation='relu')) model.add(tf.keras.layers.Dropout(0.1)) model.add(tf.keras.layers.Dense(num_classes, activation='softmax')) model.compile(loss=tf.keras.losses.categorical_crossentropy, optimizer=tf.keras.optimizers.Adadelta(), metrics=['accuracy']) model.fit(x_train, y_train, batch_size=batch_size, epochs=epochs, verbose=1, validation_data=(x_test, y_test)) score = model.evaluate(x_test, y_test, verbose=0) print('Test loss:', score[0]) print('Test accuracy:', score[1]) self.model = model def test_explain(self): explainer = ShapImage(model=self.model, preprocess_function=self.preprocess_func) explanations = explainer.explain(self.x_test[0:5]) explanations.plot()
def _concatenate_with_negation(row): ones = scipy.ones(row.shape) if issparse(row): return scipy.sparse.hstack((row, (ones - row))) else: return numpy.concatenate((row, (ones - row)), int((len(row.shape) != 1)))
def pred_to_img(x, range): (range_min, range_max) = range x -= range_min if ((range_max - range_min) > 0): x /= (range_max - range_min) return tensor_to_img(x)
.core def test_wrong_type(): with pytest.raises(ValueError): next(KFolds(2, strategy='totally not query'))
def run_with(command, SAGE_BOOTSTRAP): env = dict(os.environ) env['SAGE_BOOTSTRAP'] = SAGE_BOOTSTRAP proc = subprocess.Popen([sys.executable, __file__, command], env=env, stdout=subprocess.PIPE, stderr=subprocess.PIPE) (out, err) = proc.communicate() return (out.decode('ascii'), err.decode('ascii'))
class LYTNetV2(nn.Module): def __init__(self, n_class=5, input_size=768, width_mult=1.0): super(LYTNetV2, self).__init__() input_channel = 16 last_channel = 1280 mobile_setting = [[3, 16, 16, False, 'RE', 1], [3, 64, 24, False, 'RE', 2], [3, 72, 24, False, 'RE', 1], [5, 72, 40, True, 'RE', 2], [5, 120, 40, True, 'RE', 1], [3, 240, 80, False, 'HS', 2], [3, 200, 80, False, 'HS', 1], [3, 480, 112, True, 'HS', 1], [5, 672, 160, True, 'HS', 2], [5, 960, 160, True, 'HS', 1], [3, 960, 320, False, 'RE', 1]] assert ((input_size % 32) == 0) self.last_channel = (make_divisible((last_channel * width_mult)) if (width_mult > 1.0) else last_channel) self.features = [conv_bn(3, input_channel, 2, nlin_layer=Hswish)] for (k, exp, c, se, nl, s) in mobile_setting: output_channel = make_divisible((c * width_mult)) exp_channel = make_divisible((exp * width_mult)) self.features.append(MobileBottleneck(input_channel, output_channel, k, s, exp_channel, se, nl)) input_channel = output_channel last_conv = make_divisible((960 * width_mult)) self.features.append(conv_1x1_bn(input_channel, last_conv, nlin_layer=Hswish)) self.features.append(nn.AvgPool2d(12, 9)) self.features.append(Hswish(inplace=True)) self.features.append(nn.Conv2d(last_conv, last_channel, 1, 1, 0)) self.features.append(Hswish(inplace=True)) self.features = nn.Sequential(*self.features, nn.Dropout(0.1)) self.light_classifier = nn.Sequential(nn.Linear(self.last_channel, 5), nn.Softmax()) self.direction_regression = nn.Sequential(nn.Linear(self.last_channel, 4)) self._initialize_weights() def forward(self, x): x = self.features(x) x = x.view(x.size(0), (- 1)) x1 = self.light_classifier(x) x2 = self.direction_regression(x) return (x1, x2) def _initialize_weights(self): for m in self.modules(): if isinstance(m, nn.Conv2d): if (m.bias is not None): m.bias.data.zero_() nn.init.xavier_normal_(m.weight.data) elif isinstance(m, nn.BatchNorm2d): m.weight.data.fill_(1) m.bias.data.zero_() elif isinstance(m, nn.BatchNorm1d): m.weight.data.fill_(1) m.bias.data.zero_() elif isinstance(m, nn.Linear): m.weight.data.normal_(0, 0.01)
class pAdicRingGeneric(pAdicGeneric, sage.rings.abc.pAdicRing): def is_field(self, proof=True): return False def krull_dimension(self): return 1 def _xgcd_univariate_polynomial(self, f, g): from sage.misc.stopgap import stopgap stopgap('Extended gcd computations over p-adic fields are performed using the standard Euclidean algorithm which might produce mathematically incorrect results in some cases.', 13439) base_ring = f.base_ring() fracfield = base_ring.fraction_field() f_field = f.change_ring(fracfield) g_field = g.change_ring(fracfield) xgcd = fracfield._xgcd_univariate_polynomial(f_field, g_field) lcm = base_ring(1) for f in xgcd: for i in f: lcm = i.denominator().lcm(lcm) returnlst = [] for f in xgcd: f *= lcm returnlst.append(f.change_ring(base_ring)) return tuple(returnlst) def _gcd_univariate_polynomial(self, f, g): return self._xgcd_univariate_polynomial(f, g)[0]
def _decode_inference_indices(model, sess, output_infer, output_infer_summary_prefix, inference_indices, tgt_eos, subword_option): utils.print_out((' decoding to output %s , num sents %d.' % (output_infer, len(inference_indices)))) start_time = time.time() with codecs.getwriter('utf-8')(tf.gfile.GFile(output_infer, mode='wb')) as trans_f: trans_f.write('') for decode_id in inference_indices: (nmt_outputs, infer_summary) = model.decode(sess) assert (nmt_outputs.shape[0] == 1) translation = nmt_utils.get_translation(nmt_outputs, sent_id=0, tgt_eos=tgt_eos, subword_option=subword_option) if (infer_summary is not None): image_file = ((output_infer_summary_prefix + str(decode_id)) + '.png') utils.print_out((' save attention image to %s*' % image_file)) image_summ = tf.Summary() image_summ.ParseFromString(infer_summary) with tf.gfile.GFile(image_file, mode='w') as img_f: img_f.write(image_summ.value[0].image.encoded_image_string) trans_f.write(('%s\n' % translation)) utils.print_out((translation + b'\n')) utils.print_time(' done', start_time)
class SGConv(torch_geometric.nn.SGConv): def __init__(self, dropout=0.5, **kwargs): super(SGConv, self).__init__(**kwargs) self.normalize = True self.dropout = nn.Dropout(p=dropout) def forward(self, x: Tensor, edge_index: Adj, edge_weight: OptTensor=None) -> Tensor: cache = self._cached_x if (cache is None): if self.normalize: if isinstance(edge_index, Tensor): (edge_index, edge_weight) = gcn_norm(edge_index, edge_weight, x.size(self.node_dim), False, self.add_self_loops, dtype=x.dtype) elif isinstance(edge_index, SparseTensor): edge_index = gcn_norm(edge_index, edge_weight, x.size(self.node_dim), False, self.add_self_loops, dtype=x.dtype) for k in range(self.K): x = self.propagate(edge_index, x=x, edge_weight=edge_weight, size=None) if self.cached: self._cached_x = x else: x = cache return self.lin(self.dropout(x))
class LinearModel(object): def __init__(self): pass def train_model(self, dataset_handler: inlp_dataset_handler.DatasetHandler) -> float: raise NotImplementedError def get_weights(self) -> np.ndarray: raise NotImplementedError
class DenseNet3(nn.Module): def __init__(self, depth=100, num_classes=10, growth_rate=12, reduction=0.5, bottleneck=True, dropRate=0.0): super(DenseNet3, self).__init__() in_planes = (2 * growth_rate) n = ((depth - 4) / 3) if (bottleneck == True): n = (n / 2) block = BottleneckBlock else: block = BasicBlock self.conv1 = nn.Conv2d(3, in_planes, kernel_size=3, stride=1, padding=1, bias=False) self.block1 = DenseBlock(n, in_planes, growth_rate, block, dropRate) in_planes = int((in_planes + (n * growth_rate))) self.trans1 = TransitionBlock(in_planes, int(math.floor((in_planes * reduction))), dropRate=dropRate) in_planes = int(math.floor((in_planes * reduction))) self.block2 = DenseBlock(n, in_planes, growth_rate, block, dropRate) in_planes = int((in_planes + (n * growth_rate))) self.trans2 = TransitionBlock(in_planes, int(math.floor((in_planes * reduction))), dropRate=dropRate) in_planes = int(math.floor((in_planes * reduction))) self.block3 = DenseBlock(n, in_planes, growth_rate, block, dropRate) in_planes = int((in_planes + (n * growth_rate))) self.bn1 = nn.BatchNorm2d(in_planes) self.relu = nn.ReLU(inplace=True) self.avgpool = nn.AvgPool2d(8, stride=1) self.fc = nn.Linear(in_planes, num_classes) self.in_planes = in_planes 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))) elif isinstance(m, nn.BatchNorm2d): m.weight.data.fill_(1) m.bias.data.zero_() elif isinstance(m, nn.Linear): m.bias.data.zero_() def forward(self, x): out = self.conv1(x) out = self.trans1(self.block1(out)) out = self.trans2(self.block2(out)) out = self.block3(out) out = self.relu(self.bn1(out)) out = self.avgpool(out) out = out.view((- 1), self.in_planes) feture = out return self.fc(out)
def test_no_tensorflow_metaclass_overwritten(): assert (LayerWithObservations.__bases__ == (TrackableLayer,)) assert (type(TrackableLayer) is type) assert (type(LayerWithObservations) is abc.ABCMeta)
def load_ranking(path, qrels=None): with open(path, 'r') as file: qid_to_ranked_candidate_passages = {} for line in file: try: line = line.strip().split() if (qrels is not None): if (line[0] not in qrels): continue if (len(line) == 4): qid = line[0] pid = line[1].strip() if (len(line) == 6): qid = line[0] pid = line[2].strip() if (qid not in qid_to_ranked_candidate_passages): qid_to_ranked_candidate_passages[qid] = [] qid_to_ranked_candidate_passages[qid].append(pid) except Exception: raise IOError(('"%s" is not valid format' % line)) return qid_to_ranked_candidate_passages
class AugmentedArgumentParser(argparse.ArgumentParser): def parse_args(self, args=None, namespace=None): print('parsing args...') if ((args is None) and hasattr(self, 'custom_command') and (self.custom_command is not None)): print('using custom command') print(self.custom_command) args = shlex.split(self.custom_command)[2:] return super().parse_args(args, namespace) def parse_known_args(self, args=None, namespace=None): if ((args is None) and hasattr(self, 'custom_command') and (self.custom_command is not None)): args = shlex.split(self.custom_command)[2:] return super().parse_known_args(args, namespace) def add_argument(self, *args, **kwargs): if (('type' in kwargs) and (kwargs['type'] == util.str2bool)): if ('nargs' not in kwargs): kwargs['nargs'] = '?' if ('const' not in kwargs): kwargs['const'] = True super().add_argument(*args, **kwargs)
class Errors(object): def __init__(self, msg, rtol=0.001, atol=1e-05): self.msg = msg self.errors = [] self.context = [] self.rtol = rtol self.atol = atol class ShortCircuit(Exception): pass self.exc_class = ShortCircuit def requireAlmostEqual(self, x, y, msg=None): self.almostEqualAndThen(x, y, msg, self.failWith) def checkAlmostEqual(self, x, y, msg=None): self.almostEqualAndThen(x, y, msg, self.addErr) def almostEqualAndThen(self, x, y, msg, k): if (isinstance(x, np.ndarray) and isinstance(y, np.ndarray)): try: np.testing.assert_allclose(x, y, rtol=self.rtol, atol=self.atol, equal_nan=True, verbose=True) except AssertionError as e: raise k('{}{}'.format(colonize(msg), str(e).lstrip())) else: raise RuntimeError('Unsupported almost equal test') def requireEqual(self, x, y, msg=None): self.equalAndThen(x, y, msg, self.failWith) def checkEqual(self, x, y, msg=None): self.equalAndThen(x, y, msg, self.addErr) def equalAndThen(self, x, y, msg, k): if (isinstance(x, onnx.TensorProto) and isinstance(y, onnx.TensorProto)): self.equalAndThen(x.name, y.name, msg, k) t1 = onnx.numpy_helper.to_array(x) t2 = onnx.numpy_helper.to_array(y) new_msg = "{}In embedded parameter '{}'".format(colonize(msg), x.name) self.equalAndThen(t1, t2, new_msg, k) elif (isinstance(x, np.ndarray) and isinstance(y, np.ndarray)): try: np.testing.assert_equal(x, y) except AssertionError as e: raise k('{}{}'.format(colonize(msg, ': '), str(e).lstrip())) elif (x != y): sx = str(x) sy = str(y) if ((len(sx) > 40) or (len(sy) > 40) or ('\n' in sx) or ('\n' in sy)): l = ('=' * 50) k('\n{}The value\n{}\n{}\n{}\n\ndoes not equal\n\n{}\n{}\n{}'.format(colonize(msg, ':\n'), l, sx, l, l, sy, l)) else: k('{}{} != {}'.format(colonize(msg), sx, sy)) def requireMultiLineEqual(self, x, y, msg=None): self.multiLineEqualAndThen(x, y, msg, self.failWith) def multiLineEqualAndThen(self, x, y, msg, k): if (msg is None): msg = 'Strings are not equal' if (x != y): diff = difflib.ndiff(x.splitlines(True), y.splitlines(True)) k('{}{}'.format(colonize(msg, ':\n\n'), ''.join(diff))) def addErr(self, msg): msg_w_ctx = msg for c in reversed(self.context): msg += ('\n\n * ' + '\n '.join(c.splitlines())) self.errors.append(msg) def fail(self): raise self.exc_class() def failWith(self, msg): self.addErr(msg) self.fail() def failIfErrs(self): if self.errors: self.fail() def recover(self): parent_self = self class Recover(object): def __enter__(self): pass def __exit__(self, exc_type, exc_value, traceback): if (exc_type == parent_self.exc_class): return True return Recover() def addErrCtxt(self, msg): parent_self = self class AddContext(object): def __enter__(self): parent_self.context.append(msg) def __exit__(self, exc_type, exc_value, traceback): parent_self.context.pop() return AddContext() def __enter__(self): return self def __exit__(self, exc_type, exc_value, traceback): if self.errors: errors_msg = '\n\n'.join(map((lambda x: ('ERROR: ' + x)), self.errors)) final_msg = '{}\n{}\n{}'.format(self.msg, ('-' * 70), errors_msg) raise AssertionError(final_msg) if (exc_type == self.exc_class): raise RuntimeError('ShortCircuit was raised, but no errors were recorded')
class TransformerDecoderLayer(nn.Module): def __init__(self, args, no_encoder_attn=False, add_bias_kv=False, add_zero_attn=False): super().__init__() self.plug_in_dec_self_attn = args.plug_in_dec_self_attn self.embed_dim = args.decoder_embed_dim self.dropout_module = FairseqDropout(args.dropout, module_name=self.__class__.__name__) self.quant_noise = getattr(args, 'quant_noise_pq', 0) self.quant_noise_block_size = getattr(args, 'quant_noise_pq_block_size', 8) self.cross_self_attention = getattr(args, 'cross_self_attention', False) self.self_attn = self.build_self_attention(self.embed_dim, args, add_bias_kv=add_bias_kv, add_zero_attn=add_zero_attn) self.activation_fn = utils.get_activation_fn(activation=(str(args.activation_fn) if (getattr(args, 'activation_fn', None) is not None) else 'relu')) activation_dropout_p = getattr(args, 'activation_dropout', 0) if (activation_dropout_p == 0): activation_dropout_p = getattr(args, 'relu_dropout', 0) self.activation_dropout_module = FairseqDropout(float(activation_dropout_p), module_name=self.__class__.__name__) self.normalize_before = args.decoder_normalize_before export = getattr(args, 'char_inputs', False) self.self_attn_layer_norm = LayerNorm(self.embed_dim, export=export) if no_encoder_attn: self.encoder_attn = None self.encoder_attn_layer_norm = None else: self.encoder_attn = self.build_encoder_attention(self.embed_dim, args) self.encoder_attn_layer_norm = LayerNorm(self.embed_dim, export=export) self.fc1 = self.build_fc1(self.embed_dim, args.decoder_ffn_embed_dim, self.quant_noise, self.quant_noise_block_size) self.fc2 = self.build_fc2(args.decoder_ffn_embed_dim, self.embed_dim, self.quant_noise, self.quant_noise_block_size) self.final_layer_norm = LayerNorm(self.embed_dim, export=export) self.need_attn = True self.onnx_trace = False def build_fc1(self, input_dim, output_dim, q_noise, qn_block_size): return quant_noise(nn.Linear(input_dim, output_dim), q_noise, qn_block_size) def build_fc2(self, input_dim, output_dim, q_noise, qn_block_size): return quant_noise(nn.Linear(input_dim, output_dim), q_noise, qn_block_size) def build_self_attention(self, embed_dim, args, add_bias_kv=False, add_zero_attn=False): return MultiheadAttention(embed_dim, args.decoder_attention_heads, dropout=args.attention_dropout, add_bias_kv=add_bias_kv, add_zero_attn=add_zero_attn, self_attention=(not getattr(args, 'cross_self_attention', False)), q_noise=self.quant_noise, qn_block_size=self.quant_noise_block_size) def build_encoder_attention(self, embed_dim, args): return MultiheadAttention(embed_dim, args.decoder_attention_heads, kdim=getattr(args, 'encoder_embed_dim', None), vdim=getattr(args, 'encoder_embed_dim', None), dropout=args.attention_dropout, encoder_decoder_attention=True, q_noise=self.quant_noise, qn_block_size=self.quant_noise_block_size) def prepare_for_onnx_export_(self): self.onnx_trace = True def forward(self, x, encoder_out: Optional[torch.Tensor]=None, encoder_padding_mask: Optional[torch.Tensor]=None, incremental_state: Optional[Dict[(str, Dict[(str, Optional[Tensor])])]]=None, prev_self_attn_state: Optional[List[torch.Tensor]]=None, prev_attn_state: Optional[List[torch.Tensor]]=None, self_attn_mask: Optional[torch.Tensor]=None, self_attn_padding_mask: Optional[torch.Tensor]=None, need_attn: bool=False, need_head_weights: bool=False, past_key: Optional[Tensor]=None, past_value: Optional[Tensor]=None, past_key_padding_mask: Optional[torch.Tensor]=None, past_kv_forward: str='none'): if need_head_weights: need_attn = True residual = x if self.normalize_before: x = self.self_attn_layer_norm(x) _self_attn_input_buffer = self.self_attn._get_input_buffer(incremental_state) y = x if self.plug_in_dec_self_attn: (x, attn) = self.self_attn(query=x, key=y, value=y, key_padding_mask=self_attn_padding_mask, incremental_state=incremental_state, need_weights=False, attn_mask=self_attn_mask, past_key=past_value, past_value=past_value, past_key_padding_mask=past_key_padding_mask, past_kv_forward=past_kv_forward) else: (x, attn) = self.self_attn(query=x, key=y, value=y, key_padding_mask=self_attn_padding_mask, incremental_state=incremental_state, need_weights=False, attn_mask=self_attn_mask) x = self.dropout_module(x) x = (residual + x) if (not self.normalize_before): x = self.self_attn_layer_norm(x) if (self.encoder_attn is not None): residual = x if self.normalize_before: x = self.encoder_attn_layer_norm(x) (x, attn) = self.encoder_attn(query=x, key=encoder_out, value=encoder_out, key_padding_mask=encoder_padding_mask, incremental_state=incremental_state, static_kv=True, need_weights=(need_attn or ((not self.training) and self.need_attn)), need_head_weights=need_head_weights, past_key=past_key, past_value=past_value, past_key_padding_mask=past_key_padding_mask, past_kv_forward=past_kv_forward) x = self.dropout_module(x) x = (residual + x) if (not self.normalize_before): x = self.encoder_attn_layer_norm(x) residual = x if self.normalize_before: x = self.final_layer_norm(x) x = self.activation_fn(self.fc1(x)) x = self.activation_dropout_module(x) x = self.fc2(x) x = self.dropout_module(x) x = (residual + x) if (not self.normalize_before): x = self.final_layer_norm(x) return (x, attn, None) def make_generation_fast_(self, need_attn: bool=False, **kwargs): self.need_attn = need_attn
def show_hsv_equalized(directory, fileName): cap = cv2.VideoCapture((directory + fileName)) frameNumber = 1 frameLuminosityInfo = {} while cap.isOpened(): (ret, frame) = cap.read() if (frame is None): frameLuminosityInfo = pd.DataFrame(frameLuminosityInfo) frameLuminosityInfo.to_csv(((directory + fileName) + '.csv')) return frameLuminosityInfo break blur_frame = cv2.medianBlur(frame, 3) (H, S, V) = cv2.split(cv2.cvtColor(blur_frame, cv2.COLOR_BGR2HSV)) eq_V = cv2.equalizeHist(V) eq_image = cv2.cvtColor(cv2.merge([H, S, eq_V]), cv2.COLOR_HSV2RGB) frameLuminosityInfo[frameNumber] = {'meanValue': cv2.meanStdDev(V)[0][0][0], 'stdValue': cv2.meanStdDev(V)[1][0][0]} cv2.imshow('frame', eq_image) if ((cv2.waitKey(1) & 255) == ord('q')): break frameNumber = (frameNumber + 1) cap.release() cv2.destroyAllWindows()
class SRS_GoogLeNet(nn.Module): def __init__(self): super(SRS_GoogLeNet, self).__init__() self.pre_layers = nn.Sequential(nn.Conv2d(3, 192, kernel_size=3, padding=1), nn.BatchNorm2d(192), SRS()) self.a3 = Inception(192, 64, 96, 128, 16, 32, 32) self.b3 = Inception(256, 128, 128, 192, 32, 96, 64) self.maxpool = nn.MaxPool2d(3, stride=2, padding=1) self.a4 = Inception(480, 192, 96, 208, 16, 48, 64) self.b4 = Inception(512, 160, 112, 224, 24, 64, 64) self.c4 = Inception(512, 128, 128, 256, 24, 64, 64) self.d4 = Inception(512, 112, 144, 288, 32, 64, 64) self.e4 = Inception(528, 256, 160, 320, 32, 128, 128) self.a5 = Inception(832, 256, 160, 320, 32, 128, 128) self.b5 = Inception(832, 384, 192, 384, 48, 128, 128) self.avgpool = nn.AvgPool2d(8, stride=1) self.linear = nn.Linear(1024, 100) def forward(self, x): out = self.pre_layers(x) out = self.a3(out) out = self.b3(out) out = self.maxpool(out) out = self.a4(out) out = self.b4(out) out = self.c4(out) out = self.d4(out) out = self.e4(out) out = self.maxpool(out) out = self.a5(out) out = self.b5(out) out = self.avgpool(out) out = out.view(out.size(0), (- 1)) out = self.linear(out) return out
def _test_typing(code, inject=False): sys.path.insert(0, TOOLS_DIR) try: import jedityper finally: sys.path.remove(TOOLS_DIR) lines = [] with _tempfile(code) as f: types = jedityper.analyse(f.name) if inject: lines = jedityper.inject_types(f.name, types) return (types, lines)
def test_final_low_high_pred() -> None: y_preds = np.array([[4, 3, 2], [3, 3, 3], [2, 3, 4]]) y_pred_low = np.array([4, 3, 2]) y_pred_up = np.array([3, 3, 3]) with pytest.warns(UserWarning, match='WARNING: The predictions of .*'): check_lower_upper_bounds(y_preds, y_pred_low, y_pred_up)
def pytest_addoption(parser): parser.addoption('--runupstream', action='store_true', help='run upstream tests') parser.addoption('--runslow', action='store_true', help='run slow tests') parser.addoption('--runcorpus', action='store_true', help='run tests with corpus path dependency') parser.addoption('--practice', action='store_true', help='for test scripts only for practice and not real test cases.') parser.addoption('--runextra', action='store_true', help='run tests with extra dependencies') parser.addoption('--fairseq', action='store_true', help='run tests with fairseq dependencies') parser.addoption('--upstream_names', action='store')
def test_shuffle_players(): key = jax.random.PRNGKey(0) for i in range(100): (key, subkey) = jax.random.split(key) shuffled_players = _shuffle_players(subkey) assert ((shuffled_players[0] - shuffled_players[2]) % 2) assert ((shuffled_players[1] - shuffled_players[3]) % 2)
def integrity_check(model_tuned, hf_hub_name): model_sum = sum((param.sum() for param in model_tuned.state_dict().values())).item() model_sum_file = hf_hub_download(repo_id=hf_hub_name, filename='model_sum.txt') with open(model_sum_file, 'r') as f: model_sum_hf_hub = float(f.read()) return np.isclose(model_sum_hf_hub, model_sum)
def main(): args = parse_args() random.seed(args.seed) data = utils.load_csv_text(args.csv_filename, True) label_map = {youtube_id: text for (youtube_id, _, text, _) in data} infos = [] filenames = list(args.root_audio.rglob('*.wav')) for filename in filenames: suffix = filename.with_suffix('').relative_to(args.root_audio) frame_dir = (args.root_frame / suffix) n_frames = len(list(frame_dir.rglob('*.jpg'))) if (n_frames > (args.fps * 8)): youtube_id = filename.stem label = label_map[youtube_id].replace(', ', ' ') infos.append(f'{filename},{frame_dir},{n_frames},{label}') print(f'{len(infos)} audio/frames pairs found.') n_train = int((len(infos) * (1 - args.ratio))) random.shuffle(infos) trainset = infos[:n_train] valset = infos[n_train:] for (name, subset) in zip(('train', 'val'), (trainset, valset)): filename = (args.out_dir / f'{name}.csv') with open(filename, 'w') as f: for item in subset: f.write((item + '\n')) print(f'{len(subset)} items saved to {filename}.') print('Done!')
class Trainer(object): def __init__(self, network, optimizer, dataloader, args): self.args = args self.network = network self.optimizer = optimizer self.dataloader = dataloader if (not os.path.exists('weights')): os.makedirs('weights') self.timeformat = '%Y-%m-%d %H:%M:%S' def train(self): lossAcc = 0.0 self.network.train() dataiter = iter(self.dataloader) for _ in range((self.args.resume_iter // self.args.lr_step)): self.adjustLR() self.showLR() for step in range(self.args.resume_iter, self.args.max_step): losses = [] for _ in range(self.args.iter_size): try: (data, target) = next(dataiter) except StopIteration: dataiter = iter(self.dataloader) (data, target) = next(dataiter) (data, target) = (data.cuda(self.args.gpu_id), target.cuda(self.args.gpu_id)) (data, target) = (Variable(data), Variable(target)) loss = self.network(data, target) if np.isnan(float(loss.data[0])): raise ValueError('loss is nan while training') losses.append(loss) lossAcc += loss.data[0] bLoss = torch.mean(torch.cat(losses)) self.optimizer.zero_grad() bLoss.backward() self.optimizer.step() if ((step > 0) and ((step % self.args.lr_step) == 0)): self.adjustLR() self.showLR() if (((step + 1) % self.args.disp_interval) == 0): timestr = time.strftime(self.timeformat, time.localtime()) print('{} iter={} loss={:<8.2f}'.format(timestr, (step + 1), ((lossAcc / self.args.disp_interval) / self.args.iter_size))) lossAcc = 0.0 if (((step + 1) % self.args.save_interval) == 0): torch.save(self.network.state_dict(), './weights/hed_sklarge/{}_{}.pth'.format(self.args.network, (step + 1))) torch.save(self.network.state_dict(), './weights/hed_sklarge/{}.pth'.format(self.args.network)) def adjustLR(self): for param_group in self.optimizer.param_groups: param_group['lr'] *= self.args.lr_gamma def showLR(self): for param_group in self.optimizer.param_groups: print(param_group['lr'], end=' ') print('')
def make_folder(path, version): if (not os.path.exists(os.path.join(path, version))): os.makedirs(os.path.join(path, version))
def get_filenames(data_root, task, sub_task, split=''): if (task == 'concode'): data_dir = '{}/{}'.format(data_root, task) train_fn = '{}/train.json'.format(data_dir) dev_fn = '{}/dev.json'.format(data_dir) test_fn = '{}/test.json'.format(data_dir) elif (task == 'summarize'): data_dir = '{}/{}/{}'.format(data_root, task, sub_task) train_fn = '{}/train.jsonl'.format(data_dir) dev_fn = '{}/valid.jsonl'.format(data_dir) test_fn = '{}/test.jsonl'.format(data_dir) elif (task == 'refine'): data_dir = '{}/{}/{}'.format(data_root, task, sub_task) train_fn = '{}/train.buggy-fixed.buggy,{}/train.buggy-fixed.fixed'.format(data_dir, data_dir) dev_fn = '{}/valid.buggy-fixed.buggy,{}/valid.buggy-fixed.fixed'.format(data_dir, data_dir) test_fn = '{}/test.buggy-fixed.buggy,{}/test.buggy-fixed.fixed'.format(data_dir, data_dir) elif (task == 'translate'): data_dir = '{}/{}'.format(data_root, task) if (sub_task == 'cs-java'): train_fn = '{}/train.java-cs.txt.cs,{}/train.java-cs.txt.java'.format(data_dir, data_dir) dev_fn = '{}/valid.java-cs.txt.cs,{}/valid.java-cs.txt.java'.format(data_dir, data_dir) test_fn = '{}/test.java-cs.txt.cs,{}/test.java-cs.txt.java'.format(data_dir, data_dir) else: train_fn = '{}/train.java-cs.txt.java,{}/train.java-cs.txt.cs'.format(data_dir, data_dir) dev_fn = '{}/valid.java-cs.txt.java,{}/valid.java-cs.txt.cs'.format(data_dir, data_dir) test_fn = '{}/test.java-cs.txt.java,{}/test.java-cs.txt.cs'.format(data_dir, data_dir) elif (task == 'clone'): data_dir = '{}/{}'.format(data_root, task) train_fn = '{}/train.txt'.format(data_dir) dev_fn = '{}/valid.txt'.format(data_dir) test_fn = '{}/test.txt'.format(data_dir) elif (task == 'defect'): data_dir = '{}/{}'.format(data_root, task) train_fn = '{}/train.jsonl'.format(data_dir) dev_fn = '{}/valid.jsonl'.format(data_dir) test_fn = '{}/test.jsonl'.format(data_dir) if (split == 'train'): return train_fn elif (split == 'dev'): return dev_fn elif (split == 'test'): return test_fn else: return (train_fn, dev_fn, test_fn)
class UploadCommand(BaseUserCommand): def walk_dir(self, rel_path): entries: List[os.DirEntry] = list(os.scandir(rel_path)) files = [(os.path.join(os.getcwd(), f.path), f.path) for f in entries if f.is_file()] for f in entries: if f.is_dir(): files += self.walk_dir(f.path) return files def run(self): token = HfFolder.get_token() if (token is None): print('Not logged in') exit(1) local_path = os.path.abspath(self.args.path) if os.path.isdir(local_path): if (self.args.filename is not None): raise ValueError('Cannot specify a filename override when uploading a folder.') rel_path = os.path.basename(local_path) files = self.walk_dir(rel_path) elif os.path.isfile(local_path): filename = (self.args.filename if (self.args.filename is not None) else os.path.basename(local_path)) files = [(local_path, filename)] else: raise ValueError('Not a valid file or directory: {}'.format(local_path)) for (filepath, filename) in files: print('About to upload file {} to S3 under filename {}'.format(ANSI.bold(filepath), ANSI.bold(filename))) choice = input('Proceed? [Y/n] ').lower() if (not ((choice == '') or (choice == 'y') or (choice == 'yes'))): print('Abort') exit() print(ANSI.bold('Uploading... This might take a while if files are large')) for (filepath, filename) in files: access_url = self._api.presign_and_upload(token=token, filename=filename, filepath=filepath) print('Your file now lives at:') print(access_url)
class PositionwiseFeedForwardNetwork(nn.Module): def __init__(self, d_model: int=512, d_ff: int=2048, dropout_p: float=0.3) -> None: super(PositionwiseFeedForwardNetwork, self).__init__() self.feed_forward = nn.Sequential(nn.Linear(d_model, d_ff), nn.Dropout(dropout_p), nn.ReLU(), nn.Linear(d_ff, d_model), nn.Dropout(dropout_p)) def forward(self, inputs: torch.Tensor) -> torch.Tensor: return self.feed_forward(inputs)
def add_overviews(rst_out, tile_size, verbose=False): overview_level = rasterio.rio.overview.get_maximum_overview_level(*rst_out.shape, tile_size) overviews = [(2 ** j) for j in range(1, (overview_level + 1))] if verbose: print(f'Adding pyramid overviews to raster {overviews}') rst_out.build_overviews(overviews, rasterio.warp.Resampling.average) rst_out.update_tags(ns='rio_overview', resampling='nearest') tags = rst_out.tags() tags.update(OVR_RESAMPLING_ALG='NEAREST') rst_out.update_tags(**tags) rst_out._set_all_scales([rst_out.scales[(b - 1)] for b in rst_out.indexes]) rst_out._set_all_offsets([rst_out.offsets[(b - 1)] for b in rst_out.indexes])
def stackedunet128(output_stride='32'): return Stackedunet_imagenet(in_dim=512, start_planes=64, filters_base=128, num_classes=1000, depth=4, ost=output_stride)
def register_Ns3RrpaaWifiManager_methods(root_module, cls): cls.add_constructor([param('ns3::RrpaaWifiManager const &', 'arg0')]) cls.add_constructor([]) cls.add_method('AssignStreams', 'int64_t', [param('int64_t', 'stream')]) cls.add_method('GetTypeId', 'ns3::TypeId', [], is_static=True) cls.add_method('SetHeSupported', 'void', [param('bool', 'enable')], is_virtual=True) cls.add_method('SetHtSupported', 'void', [param('bool', 'enable')], is_virtual=True) cls.add_method('SetVhtSupported', 'void', [param('bool', 'enable')], is_virtual=True) cls.add_method('SetupMac', 'void', [param('ns3::Ptr< ns3::WifiMac > const', 'mac')], is_virtual=True) cls.add_method('SetupPhy', 'void', [param('ns3::Ptr< ns3::WifiPhy > const', 'phy')], is_virtual=True) cls.add_method('DoCreateStation', 'ns3::WifiRemoteStation *', [], is_const=True, visibility='private', is_virtual=True) cls.add_method('DoGetDataTxVector', 'ns3::WifiTxVector', [param('ns3::WifiRemoteStation *', 'station')], visibility='private', is_virtual=True) cls.add_method('DoGetRtsTxVector', 'ns3::WifiTxVector', [param('ns3::WifiRemoteStation *', 'station')], visibility='private', is_virtual=True) cls.add_method('DoNeedRts', 'bool', [param('ns3::WifiRemoteStation *', 'st'), param('ns3::Ptr< ns3::Packet const >', 'packet'), param('bool', 'normally')], visibility='private', is_virtual=True) cls.add_method('DoReportDataFailed', 'void', [param('ns3::WifiRemoteStation *', 'station')], visibility='private', is_virtual=True) cls.add_method('DoReportDataOk', 'void', [param('ns3::WifiRemoteStation *', 'station'), param('double', 'ackSnr'), param('ns3::WifiMode', 'ackMode'), param('double', 'dataSnr')], visibility='private', is_virtual=True) cls.add_method('DoReportFinalDataFailed', 'void', [param('ns3::WifiRemoteStation *', 'station')], visibility='private', is_virtual=True) cls.add_method('DoReportFinalRtsFailed', 'void', [param('ns3::WifiRemoteStation *', 'station')], visibility='private', is_virtual=True) cls.add_method('DoReportRtsFailed', 'void', [param('ns3::WifiRemoteStation *', 'station')], visibility='private', is_virtual=True) cls.add_method('DoReportRtsOk', 'void', [param('ns3::WifiRemoteStation *', 'station'), param('double', 'ctsSnr'), param('ns3::WifiMode', 'ctsMode'), param('double', 'rtsSnr')], visibility='private', is_virtual=True) cls.add_method('DoReportRxOk', 'void', [param('ns3::WifiRemoteStation *', 'station'), param('double', 'rxSnr'), param('ns3::WifiMode', 'txMode')], visibility='private', is_virtual=True) cls.add_method('IsLowLatency', 'bool', [], is_const=True, visibility='private', is_virtual=True) return
class TestSegmentedImageDataset(): def test_init(self): pass def test_len(self): pass def test_get_item(self): pass
def get_wav_files(tgt_split_path): wavs_pattern = os.path.join(tgt_split_path, '**', '*.wav') return list(sorted(glob(wavs_pattern)))
class ResnetCSNNoGC(Resnet3d101): def __init__(self, tw=8, sample_size=112, e_dim=7, decoders=None): decoders = ([Decoder3dNoGC()] if (decoders is None) else decoders) print('Creating decoders {}'.format(decoders)) super(ResnetCSNNoGC, self).__init__(tw, sample_size, e_dim, decoders) self.encoder = Encoder3d_csn_ir(tw, sample_size)
def message(con=None, log=None, queue=None): if (con and (log == '')): log = sb.colors.strip(con) if (con and (not quiet)): print(con, flush=True) if log: if queue: queue.put(log) else: __prolog.append(log)
class ImmutableMultiDictMixin(ImmutableDictMixin): def __reduce_ex__(self, protocol): return (type(self), (list(iteritems(self, multi=True)),)) def _iter_hashitems(self): return iteritems(self, multi=True) def add(self, key, value): is_immutable(self) def popitemlist(self): is_immutable(self) def poplist(self, key): is_immutable(self) def setlist(self, key, new_list): is_immutable(self) def setlistdefault(self, key, default_list=None): is_immutable(self)
def _object_to_tensor(obj): buffer = pickle.dumps(obj) byte_storage = torch.ByteStorage.from_buffer(buffer) byte_tensor = torch.ByteTensor(byte_storage) local_size = torch.LongTensor([byte_tensor.numel()]) return (byte_tensor, local_size)
() ('--checkpoint-file', type=click.Path(exists=True)) ('--data-dir', default='data/record', type=click.Path(exists=True)) ('--doc-stride', default=128) ('--do-eval/--no-eval', default=True) ('--do-train/--no-train', default=True) ('--eval-batch-size', default=32) ('--max-query-length', default=90) ('--max-seq-length', default=512) ('--num-train-epochs', default=2.0) ('--seed', default=4) ('--train-batch-size', default=1) _args _obj def run(common_args, **task_args): task_args.update(common_args) args = Namespace(**task_args) set_seed(args.seed) args.experiment.log_parameters({p.name: getattr(args, p.name) for p in run.params}) args.model_config.vocab_size += 3 word_emb = args.model_weights['embeddings.word_embeddings.weight'] highlight_emb = word_emb[args.tokenizer.convert_tokens_to_ids([''])[0]].unsqueeze(0) placeholder_emb = word_emb[args.tokenizer.convert_tokens_to_ids(['#'])[0]].unsqueeze(0) marker_emb = word_emb[args.tokenizer.convert_tokens_to_ids(['*'])[0]].unsqueeze(0) args.model_weights['embeddings.word_embeddings.weight'] = torch.cat([word_emb, highlight_emb, placeholder_emb, marker_emb]) args.tokenizer.add_special_tokens(dict(additional_special_tokens=[HIGHLIGHT_TOKEN, PLACEHOLDER_TOKEN, ENTITY_MARKER_TOKEN])) args.model_config.entity_vocab_size = 2 entity_emb = args.model_weights['entity_embeddings.entity_embeddings.weight'] mask_emb = entity_emb[args.entity_vocab[MASK_TOKEN]].unsqueeze(0) args.model_weights['entity_embeddings.entity_embeddings.weight'] = torch.cat([entity_emb[:1], mask_emb]) results = {} if args.do_train: model = LukeForEntitySpanQA(args) model.load_state_dict(args.model_weights, strict=False) model.to(args.device) (train_dataloader, _, _, _) = load_examples(args, 'train') num_train_steps_per_epoch = (len(train_dataloader) // args.gradient_accumulation_steps) num_train_steps = int((num_train_steps_per_epoch * args.num_train_epochs)) best_dev_score = [(- 1)] best_weights = [None] def step_callback(model, global_step): if (((global_step % num_train_steps_per_epoch) == 0) and (args.local_rank in (0, (- 1)))): epoch = int(((global_step / num_train_steps_per_epoch) - 1)) dev_results = evaluate(args, model, fold='dev') args.experiment.log_metrics({f'dev_{k}_epoch{epoch}': v for (k, v) in dev_results.items()}, epoch=epoch) results.update({f'dev_{k}_epoch{epoch}': v for (k, v) in dev_results.items()}) tqdm.write(('dev: ' + str(dev_results))) if (dev_results['exact_match'] > best_dev_score[0]): if hasattr(model, 'module'): best_weights[0] = {k: v.to('cpu').clone() for (k, v) in model.module.state_dict().items()} else: best_weights[0] = {k: v.to('cpu').clone() for (k, v) in model.state_dict().items()} best_dev_score[0] = dev_results['exact_match'] results['best_epoch'] = epoch model.train() trainer = Trainer(args, model=model, dataloader=train_dataloader, num_train_steps=num_train_steps, step_callback=step_callback) trainer.train() if (args.do_train and (args.local_rank in (0, (- 1)))): logger.info('Saving the model checkpoint to %s', args.output_dir) torch.save(best_weights[0], os.path.join(args.output_dir, WEIGHTS_NAME)) if (args.local_rank not in (0, (- 1))): return {} model = None torch.cuda.empty_cache() if args.do_eval: model = LukeForEntitySpanQA(args) if args.checkpoint_file: model.load_state_dict(torch.load(args.checkpoint_file, map_location='cpu')) else: model.load_state_dict(torch.load(os.path.join(args.output_dir, WEIGHTS_NAME), map_location='cpu')) model.to(args.device) output_file = os.path.join(args.output_dir, 'predictions.json') results.update({f'dev_{k}': v for (k, v) in evaluate(args, model, fold='dev', output_file=output_file).items()}) logger.info('Results: %s', json.dumps(results, indent=2, sort_keys=True)) args.experiment.log_metrics(results) with open(os.path.join(args.output_dir, 'results.json'), 'w') as f: json.dump(results, f) return results
def test_pcpvt(): (H, W) = (224, 224) temp = torch.randn((1, 3, H, W)) model = PCPVT(embed_dims=[32, 64, 160, 256], num_heads=[1, 2, 5, 8], mlp_ratios=[8, 8, 4, 4], qkv_bias=True, depths=[3, 4, 6, 3], sr_ratios=[8, 4, 2, 1], norm_after_stage=False) model.init_weights() outs = model(temp) assert (outs[0].shape == (1, 32, (H // 4), (W // 4))) assert (outs[1].shape == (1, 64, (H // 8), (W // 8))) assert (outs[2].shape == (1, 160, (H // 16), (W // 16))) assert (outs[3].shape == (1, 256, (H // 32), (W // 32)))