我们从Python开源项目中，提取了以下50个代码示例，用于说明如何使用joblib.delayed()。
def update_row_param(self, phi_csr, mu0, c, v, r_prev, u_prev, phi_r, phi_u, num_process): nrow = self.y_csr.shape[0] # Update 'c' and 'v' block-wise in parallel. if num_process == 1: r, u = self.update_row_param_blockwise(self.y_csr, phi_csr, mu0, c, v, r_prev, u_prev, phi_r, phi_u) else: n_block = num_process block_ind = np.linspace(0, nrow, 1 + n_block, dtype=int) ru = joblib.Parallel(n_jobs=num_process)( joblib.delayed(self.update_row_param_blockwise)( self.y_csr[block_ind[m]:block_ind[m + 1], :], phi_csr[block_ind[m]:block_ind[m + 1], :], mu0, c, v, r_prev[block_ind[m]:block_ind[m + 1]], u_prev[block_ind[m]:block_ind[m + 1]], phi_r[block_ind[m]:block_ind[m + 1]], phi_u) for m in range(n_block)) r = np.concatenate([ru_i[0] for ru_i in ru]) u = np.vstack([ru_i[1] for ru_i in ru]) return r, u
def evaluate(input_path, n_jobs): aud, ann = zip(*crema.utils.get_ann_audio(input_path)) test_idx = set(pd.read_json('index_test.json')['id']) # drop anything not in the test set ann = [ann_i for ann_i in ann if crema.utils.base(ann_i) in test_idx] aud = [aud_i for aud_i in aud if crema.utils.base(aud_i) in test_idx] stream = tqdm(zip(ann, aud), desc='Evaluating test set', total=len(ann)) results = Parallel(n_jobs=n_jobs)(delayed(track_eval)(ann_i, aud_i) for ann_i, aud_i in stream) df = pd.DataFrame.from_dict(dict(results), orient='index') print('Results') print('-------') print(df.describe()) df.to_json(os.path.join(OUTPUT_PATH, 'test_scores.json'))
def update_col_param(self, phi_csc, mu0, r, u, c_prev, v_prev, phi_c, phi_v, num_process): ncol = self.y_csc.shape[1] if num_process == 1: c, v = self.update_col_param_blockwise(self.y_csc, phi_csc, mu0, r, u, c_prev, v_prev, phi_c, phi_v) else: # Update 'c' and 'v' block-wise in parallel. n_block = num_process block_ind = np.linspace(0, ncol, 1 + n_block, dtype=int) cv = joblib.Parallel(n_jobs=num_process)( joblib.delayed(self.update_col_param_blockwise)( self.y_csc[:, block_ind[m]:block_ind[m + 1]], phi_csc[:, block_ind[m]:block_ind[m + 1]], mu0, r, u, c_prev[block_ind[m]:block_ind[m + 1]], v_prev[block_ind[m]:block_ind[m + 1]], phi_c[block_ind[m]:block_ind[m + 1]], phi_v) for m in range(n_block)) c = np.concatenate([cv_j[0] for cv_j in cv]) v = np.vstack([cv_j[1] for cv_j in cv]) return c, v
def _fit_single_job(self, job_list, X, y): cv_results_ = {} # for i, (train_index, test_index) in job_list: # LOG.info("Training fold %d", i + 1) # # slave_result_ = self._worker( # i, X, y, train_index, test_index) # # _build_cv_results(cv_results_, **slave_result_) slave_results = jl.Parallel(n_jobs=self.n_jobs) \ (jl.delayed(_worker)( self, i, X, y, train_index, test_index) for i, ( train_index, test_index) in job_list) for slave_result_ in slave_results: _build_cv_results(cv_results_, **slave_result_) self.cv_results_ = cv_results_
def degreetocart(data_f1): global df2 df2 = data_f1.copy() print "phase 1" df2['X'] = np.nan df2['Y'] = np.nan df2['Z'] = np.nan df2 = df2.astype(float) print "phase 2" num_cores = multiprocessing.cpu_count() results_x = Parallel(n_jobs=num_cores)(delayed(xloop)(i) for i in xrange(0,len(df2))) print "phase 3" #print results_x #print results_x #print " this is " #print results_x[0] results_y = Parallel(n_jobs=num_cores)(delayed(yloop)(i) for i in xrange(0,len(df2))) print "phase 4" results_z = Parallel(n_jobs=num_cores)(delayed(zloop)(i) for i in xrange(0,len(df2))) print "phase 5" #print results_y #Parallel(n_jobs=num_cores)(delayed(adjloop)(i) for i in xrange(0,len(df2))) for i in xrange(0,len(df2)): print i df2['X'][i] = results_x[i] df2['Y'][i] = results_y[i] df2['Z'][i] = results_z[i]
def shutdown(self): """Shutdown all mturk agents in parallel, otherwise if one mturk agent is disconnected then it could prevent other mturk agents from completing. """ global shutdown_agent def shutdown_agent(agent): try: agent.shutdown(timeout=None) except Exception: agent.shutdown() # not MTurkAgent Parallel( n_jobs=len(self.agents), backend='threading' )(delayed(shutdown_agent)(agent) for agent in self.agents)
def get_distilled_labels(filenames): result_labels = [] print("Creating labels") result_labels = Parallel(n_jobs=num_cores)(delayed(make_label)(long_filename) for long_filename in tqdm(filenames)) return result_labels # This function recives paths to images and lines from file with labels # and returns only path to images that have corresponding label
def _get_data(path, preprocess): data = sio.loadmat(path)['gestures'] data = [np.transpose(np.delete(segment.astype(np.float32), np.s_[7:192:8], 0)) for segment in data.flat] if preprocess: data = list(Context.parallel(jb.delayed(preprocess)(segment, **PREPROCESS_KARGS) for segment in data)) return data # @cached # def _get_data(path, bandstop, cut, downsample): # data = sio.loadmat(path)['gestures'] # data = [np.transpose(np.delete(segment.astype(np.float32), np.s_[7:192:8], 0)) # for segment in data.flat] # if bandstop: # data = list(Context.parallel(jb.delayed(get_bandstop)(segment) for segment in data)) # if cut is not None: # data = list(Context.parallel(jb.delayed(cut)(segment, framerate=FRAMERATE) for segment in data)) # if downsample > 1: # data = [segment[::downsample].copy() for segment in data] # return data
def get_information(ws, x, label, num_of_bins, interval_information_display, model, layerSize, calc_parallel=True, py_hats=0): """Calculate the information for the network for all the epochs and all the layers""" print('Start calculating the information...') bins = np.linspace(-1, 1, num_of_bins) label = np.array(label).astype(np.float) pys, pys1, p_y_given_x, b1, b, unique_a, unique_inverse_x, unique_inverse_y, pxs = extract_probs(label, x) if calc_parallel: params = np.array(Parallel(n_jobs=NUM_CORES )(delayed(calc_information_for_epoch) (i, interval_information_display, ws[i], bins, unique_inverse_x, unique_inverse_y, label, b, b1, len(unique_a), pys, pxs, p_y_given_x, pys1, model.save_file, x.shape[1], layerSize) for i in range(len(ws)))) else: params = np.array([calc_information_for_epoch (i, interval_information_display, ws[i], bins, unique_inverse_x, unique_inverse_y, label, b, b1, len(unique_a), pys, pxs, p_y_given_x, pys1, model.save_file, x.shape[1], layerSize) for i in range(len(ws))]) return params
def _erpac(xp, xa, n_perm, n_jobs): """Sub erpac function [xp] = [xa] = (npts, ntrials) """ npts, ntrials = xp.shape # Compute ERPAC xerpac = np.zeros((npts,)) for t in range(npts): xerpac[t] = circ_corrcc(xp[t, :], xa[t, :])[0] # Compute surrogates: data = Parallel(n_jobs=n_jobs)(delayed(_erpacSuro)( xp, xa, npts, ntrials) for pe in range(n_perm)) suro = np.array(data) # Normalize erpac: xerpac = (xerpac - suro.mean(0))/suro.std(0) # Get p-value: pvalue = norm.cdf(-np.abs(xerpac))*2 return xerpac, pvalue
def _fit(x, y, clf, cv, mf, grp, center, n_jobs): """Sub function for fitting """ # Check the inputs size : x, y = checkXY(x, y, mf, grp, center) rep, nfeat = len(cv), len(x) # Tricks : construct a list of tuple containing the index of # (repetitions,features) & loop on it. Optimal for parallel computing : claIdx, listRep, listFeat = list2index(rep, nfeat) # Run the classification : cvs = Parallel(n_jobs=n_jobs)(delayed(_cvscore)( x[k[1]], y, clone(clf), cv[k[0]]) for k in claIdx) da, y_true, y_pred = zip(*cvs) # Reconstruct elements : da = np.array(groupInList(da, listFeat)) y_true = groupInList(y_true, listFeat) y_pred = groupInList(y_pred, listFeat) return da, x, y, y_true, y_pred
def jobmap(func, INPUT_ITR, FLAG_PARALLEL=False, batch_size=None, *args, **kwargs): n_jobs = -1 if FLAG_PARALLEL else 1 dfunc = joblib.delayed(func) with joblib.Parallel(n_jobs=n_jobs) as MP: # Yield the whole thing if there isn't a batch_size if batch_size is None: for z in MP(dfunc(x, *args, **kwargs) for x in INPUT_ITR): yield z raise StopIteration ITR = iter(INPUT_ITR) progress_bar = tqdm() for block in grouper(ITR, batch_size): MPITR = MP(dfunc(x, *args, **kwargs) for x in block) for k,z in enumerate(MPITR): yield z progress_bar.update(k+1)
def process_and_evaluate(model, X, Y, k, n_jobs=1): """ Arguments: X : query_id, query pairs Y : dict of dicts (harvestable) k : int how many to retrieve """ print("Starting query time with %d jobs" % n_jobs) # TODO can we unzip Y and only pass the fucking chunk of y which # it needs to harvest?? qids_rs = Parallel(n_jobs=n_jobs)(delayed(process_query)(model, x, Y, k) for x in X) print("Evaluating the results:") scores = evaluate_results(qids_rs, Y, k) return scores
def smooth(s,lengthscale,parallel=True): """smoothes s vertically""" if len(s.shape) == 1: s=s[...,None] nChans = s.shape[1] lengthscale=2*round(float(lengthscale)/2) W = np.hamming(min(lengthscale,s.shape[0])) W/= np.sum(W) if s.shape[1]>1: if parallel: njobs=JOBLIB_NCORES else: njobs=1 slidingMean = (Parallel(n_jobs=njobs,backend=JOBLIB_BACKEND,temp_folder=JOBLIB_TEMPFOLDER) (delayed(smoothLine)(s[:,chan],W) for chan in range(nChans))) return np.array(slidingMean).T else: return smoothLine(s[:,0],W)[...,None]
def run(self, model, x, y=None, scoring=None, max_threads=1): # get scorers if scoring is not None: if isinstance(scoring, list): scorers_fn = dict([(self.get_scorer_name(k), get_scorer(k)) for k in scoring]) else: scorers_fn = dict([(self.get_scorer_name(scoring), get_scorer(scoring))]) else: # By default uses the model loss function as scoring function scorers_fn = dict([(model.get_loss_func(), get_scorer(model.get_loss_func()))]) model_cfg = model.to_json() if y is None: args = [(model_cfg['model'], train, test, x, scorers_fn) for train, test in self.cv.split(x, y)] cv_fn = self._do_unsupervised_cv else: args = [(model_cfg['model'], train, test, x, y, scorers_fn) for train, test in self.cv.split(x, y)] cv_fn = self._do_supervised_cv with Parallel(n_jobs=min(max_threads, len(args))) as parallel: cv_results = parallel(delayed(function=cv_fn, check_pickle=False)(*a) for a in args) return self._consolidate_cv_scores(cv_results)
def testParallel(parallel = True): inputs = range(0, 1000, 1) param = 1000 if parallel == True: # parallel stuff # This is reference code for parallel implementation inputs = range(10) num_cores = multiprocessing.cpu_count() results = joblib.Parallel(n_jobs=num_cores)(joblib.delayed(childFunc)(i, param) for i in inputs) else: for i in inputs: childFunc(i) print(results)
def create_log2fc_bigwigs(matrix, outdir, args): # type: (pd.DataFrame, str, Namespace) -> None """Create bigwigs from matrix.""" call("mkdir -p {}".format(outdir), shell=True) genome_size_dict = args.chromosome_sizes input_columns = matrix[args.control] input_rpkm_sum = (1e6 * input_columns / input_columns.sum()).sum(axis=1) / len(args.control) input_rpkm_sum[input_rpkm_sum == 0] = 1 outpaths, data = [], [] for bed_file in matrix[args.treatment]: outpath = join(outdir, splitext(basename(bed_file))[0] + "_log2fc.bw") outpaths.append(outpath) bed_column = matrix[bed_file] bed_column = 1e6 * bed_column / bed_column.sum() divided = bed_column / input_rpkm_sum divided.loc[divided == 0] = 0.01 log2_fc_column = np.log2(divided) data.append(log2_fc_column) Parallel(n_jobs=args.number_cores)(delayed(_create_bigwig)(bed_column, outpath, genome_size_dict) for outpath, bed_column in zip(outpaths, data))
def create_bigwigs(matrix, outdir, args): # type: (pd.DataFrame, str, Namespace) -> None """Create bigwigs from matrix.""" call("mkdir -p {}".format(outdir), shell=True) genome_size_dict = args.chromosome_sizes outpaths, data = [], [] for bed_file in matrix: outpath = join(outdir, splitext(basename(bed_file))[0] + ".bw") outpaths.append(outpath) bed_column = matrix[bed_file] bed_column = 1e6 * bed_column / bed_column.sum() data.append(bed_column) Parallel(n_jobs=args.number_cores)(delayed(_create_bigwig)(bed_column, outpath, genome_size_dict) for outpath, bed_column in zip(outpaths, data))
def count_reads_in_windows(bed_file, args): # type: (str, Namespace) -> List[pd.DataFrame] chromosome_size_dict = args.chromosome_sizes chromosomes = natsorted(list(chromosome_size_dict.keys())) parallel_count_reads = partial(_count_reads_in_windows, bed_file, args) info("Binning chromosomes {}".format(", ".join([c.replace("chr", "") for c in chromosomes]))) chromosome_dfs = Parallel(n_jobs=args.number_cores)( delayed(parallel_count_reads)(chromosome_size_dict[chromosome], chromosome, strand) for chromosome, strand in product(chromosomes, ["+", "-"])) info("Merging the bins on both strands per chromosome.") both_chromosome_strand_dfs = [df_pair for df_pair in _pairwise(chromosome_dfs)] merged_chromosome_dfs = Parallel( n_jobs=args.number_cores)(delayed(merge_chromosome_dfs)(df_pair) for df_pair in both_chromosome_strand_dfs) return merged_chromosome_dfs
def merge_same_files(sample1_dfs, sample2_dfs, nb_cpu): # type: (List[pd.DataFrame], List[pd.DataFrame], int) -> List[pd.DataFrame] # if one list is missing a chromosome, we might pair up the wrong dataframes # therefore creating dicts beforehand to ensure they are paired up properly d1, d2 = ensure_same_chromosomes_in_list(sample1_dfs, sample2_dfs) assert len(d1) == len(d2) logging.info("Merging same class data.") merged_chromosome_dfs = Parallel(n_jobs=nb_cpu)(delayed(_merge_same_files)( d1[chromosome], d2[chromosome]) for chromosome in d1.keys()) return merged_chromosome_dfs
def create_matrixes(chip, input, df, args): # type: (Iterable[pd.DataFrame], Iterable[pd.DataFrame], pd.DataFrame, Namespace) -> List[pd.DataFrame] "Creates matrixes which can be written to file as is (matrix) or as bedGraph." genome = args.chromosome_sizes chip = put_dfs_in_chromosome_dict(chip) input = put_dfs_in_chromosome_dict(input) all_chromosomes = natsorted(set(list(chip.keys()) + list(input.keys()))) # print("df1\n", df, file=sys.stderr) islands = enriched_bins(df, args) # print("islands1\n", islands, file=sys.stderr) logging.info("Creating matrixes from count data.") dfms = Parallel(n_jobs=args.number_cores)(delayed(_create_matrixes)( chromosome, chip, input, islands, genome[chromosome], args.window_size) for chromosome in all_chromosomes) return dfms
def fit(self, X, y): assert isinstance(X, list) #TODO: this should not be an assert assert len(y) > 0 assert len(X) == len(y) # TODO: add support for fitting again after having already performed a fit self.n_labels_ = y.shape[1] self.models_ = [] # Train one model per label. If no data is available for a given label, the model is set to None. models, data = [], [] for idx in range(self.n_labels_): d = [X[i] for i in np.where(y[:, idx] == 1)[0]] if len(d) == 0: model = None else: model = clone(self.model) data.append(d) models.append(model) assert len(models) == len(data) n_jobs = self.n_jobs if self.model.supports_parallel() else 1 self.models_ = Parallel(n_jobs=n_jobs)(delayed(_perform_fit)(models[i], data[i]) for i in range(len(models))) assert len(self.models_) == self.n_labels_
def preprocess_images_multiprocess(path2dataset_orig, prefix_orig, path2dataset_dest, prefix_dest, img_rows, img_cols, img_crop_rows, img_crop_cols): # Origin path = prefix + path -> /mnt/img/img393.JPEG # Destiny path = prefix2 + path -> /mnt/h5/img393.h5 with open(path2dataset_orig, 'rb') as fin: paths = fin.readlines() num_total_paths = len(paths) num_cores = multiprocessing.cpu_count() processed_paths = Parallel(n_jobs=num_cores)(delayed(preprocess_images_worker) \ (line, prefix_orig, prefix_dest, img_rows, img_cols, img_crop_rows, img_crop_cols) for line in paths) processed_paths = [elem for elem in processed_paths if elem] with open(path2dataset_dest, "wb") as fout: fout.writelines(processed_paths) print("Total images pre-processed: %d (remember that corrupted or not present images were discarded)" % len(processed_paths))
def get_graph_stats(graph_obj_handle, prop='degrees'): # if prop == 'degrees': num_cores = multiprocessing.cpu_count() inputs = [int(i*len(graph_obj_handle)/num_cores) for i in range(num_cores)] + [len(graph_obj_handle)] res = Parallel(n_jobs=num_cores)(delayed(get_values)(graph_obj_handle, inputs[i], inputs[i+1], prop) for i in range(num_cores)) stat_dict = {} if 'degrees' in prop: stat_dict['degrees'] = list(set([d for core_res in res for file_res in core_res for d in file_res['degrees']])) if 'edge_labels' in prop: stat_dict['edge_labels'] = list(set([d for core_res in res for file_res in core_res for d in file_res['edge_labels']])) if 'target_mean' in prop or 'target_std' in prop: param = np.array([file_res['params'] for core_res in res for file_res in core_res]) if 'target_mean' in prop: stat_dict['target_mean'] = np.mean(param, axis=0) if 'target_std' in prop: stat_dict['target_std'] = np.std(param, axis=0) return stat_dict
def runStat(dataI, minPts, cut, cpu, fout, hichip=0): """ Calling p-values of interactions for all chromosomes. """ logger.info("Starting estimate significance for interactions") ds = Parallel(n_jobs=cpu)(delayed(getIntSig)( dataI[key]["f"], dataI[key]["records"], minPts, cut) for key in dataI.keys()) ds = [d for d in ds if d is not None] if len(ds) == 0: logger.error("Something wrong, no loops found, sorry, bye.") return 1 ds = pd.concat(ds) try: if hichip: ds = markIntSigHic(ds) else: ds = markIntSig(ds) ds.to_csv(fout + ".loop", sep="\t", index_label="loopId") except: logger.warning( "Something wrong happend to significance estimation, only output called loops" ) ds.to_csv(fout + "_raw.loop", sep="\t", index_label="loopId") return 0
def h0_opt_test_stresstest(IP=None, stress_count=2000, stress_threads=50, **kwargs): ''' Throw {stress_count} domains at the pihole via {stress_threads} threads ''' from joblib import Parallel, delayed top_array = open('topsites.txt').read().splitlines() random.shuffle(top_array) results = Parallel(n_jobs=stress_threads, backend='threading' )(delayed(dns_stress)(IP, site) for site in top_array[:stress_count]) good = sum( 1 for (a,b) in results if a == 'good' ) numbers = [ b for (a,b) in results if a == 'good' ] bad = sum( 1 for (a,b) in results if a == 'bad' ) vmin = min(numbers)*1000 vmax = max(numbers)*1000 vavg = sum(numbers)*1000//len(numbers) vstd = (sum(((n*1000) - vavg) ** 2 for n in numbers) / len(numbers)) ** .5 return not bad or (good/bad)>0.05, "{good}/{bad} min {vmin:.2f}ms avg {vavg:.2f}ms max {vmax:.2f}ms std {vstd:.2f}ms".format(**locals())
def get_X_y(): start = time.time() X = [] y = [] for j in range(10): print('Load folder c{}'.format(j)) path = os.path.join('imgs/train', 'c' + str(j), '*.jpg') files = glob.glob(path) X.extend(Parallel(n_jobs=-1)(delayed(process_image)(im_file) for im_file in files)) y.extend([j]*len(files)) end = time.time() - start print("Time: %.2f seconds" % end) return np.array(X), np.array(y)
def get_X_y(): start = time.time() X = [] y = [] for j in range(10): print('Load folder c{}'.format(j)) path = os.path.join('imgs/train', 'c' + str(j), '*.jpg') files = glob.glob(path) X.extend(Parallel(n_jobs=-1)(delayed(process_image)(im_file) for im_file in files)) y.extend([j]*len(files)) end = time.time() - start print("Time: %.2f seconds" % end) return X, y
def get_other(): start = time.time() X = [] y = [] train_id = [] for j in range(10): one, two = [], [] print('Load folder c{}'.format(j)) path = os.path.join('imgs2/', str(j), '*.jpg') files = glob.glob(path) results = Parallel(n_jobs=-1)(delayed(process_image)(im_file) for im_file in files) one, two = zip(*results) X.extend(one) train_id.extend(two) y.extend([j]*len(files)) end = time.time() - start print("Time: %.2f seconds" % end) return np.array(X), np.array(y), np.array(train_id)
def _resample_labels_joblib(self,num_procs): from joblib import Parallel, delayed import parallel if len(self.labels_list) > 0: parallel.model = self raw = Parallel(n_jobs=num_procs,backend='multiprocessing')\ (delayed(parallel._get_sampled_labels)(idx) for idx in range(len(self.labels_list))) for l, (z,normalizer) in zip(self.labels_list,raw): l.z, l._normalizer = z, normalizer ### Mean Field
def _joblib_resample_states(self,states_list,num_procs): from joblib import Parallel, delayed import parallel # warn('joblib is segfaulting on OS X only, not sure why') if len(states_list) > 0: joblib_args = list_split( [self._get_joblib_pair(s) for s in states_list], num_procs) parallel.model = self parallel.args = joblib_args raw_stateseqs = Parallel(n_jobs=num_procs,backend='multiprocessing')\ (delayed(parallel._get_sampled_stateseq)(idx) for idx in range(len(joblib_args))) for s, (stateseq, log_likelihood) in zip( [s for grp in list_split(states_list,num_procs) for s in grp], [seq for grp in raw_stateseqs for seq in grp]): s.stateseq, s._normalizer = stateseq, log_likelihood
def _joblib_meanfield_update_states(self,states_list,num_procs): if len(states_list) > 0: from joblib import Parallel, delayed import parallel joblib_args = list_split( [self._get_joblib_pair(s) for s in states_list], num_procs) parallel.model = self parallel.args = joblib_args allstats = Parallel(n_jobs=num_procs,backend='multiprocessing')\ (delayed(parallel._get_stats)(idx) for idx in range(len(joblib_args))) for s, stats in zip( [s for grp in list_split(states_list) for s in grp], [s for grp in allstats for s in grp]): s.all_expected_stats = stats
def RelationReport(Tensor,RelationTensor,matrixA,dictIndextoVP,dictVPtoIndex,lambdaFolderName,runDir,num_cores_for_fit_computation): numVP = len(RelationTensor) # dummy code to help parallelize RelIndexFitReport = [] # List of index to fit, indices to be sorted based on fit [(verbPhrase, relNum, tensorSliceNorm, normResidueSlice, Fit), tuples] # for relIndex in range(0,numVP): # verbPhrase,relNum,tensorSliceNorm, normResidueSlice, Fit = computeSliceFit(Tensor[relIndex],RelationTensor[relIndex],matrixA,dictIndextoVP,relIndex) # RelIndexFitReport.append((verbPhrase,relNum,tensorSliceNorm, normResidueSlice, Fit)) RelIndexFitReport = Parallel(n_jobs=num_cores_for_rescal, verbose=1)(delayed(cheaplyComputeSliceFit)(Tensor[relIndex],RelationTensor[relIndex],matrixA,dictIndextoVP,dictIndextoNP,relIndex) for relIndex in range(0,numVP)) RelIndexFitReport.sort(key = lambda x:x[4],reverse=True) # sort based on fit of relations # print(RelIndexFitReport) # check whether sorted. # print('Printing Path') # print(os.path.join(lambdaFolderName,runDir,'RelationReport.txt')) # Writing old relation Report to a file RelationReportHandle = open(os.path.join(os.getcwd(),lambdaFolderName,runDir,'RelationReport.txt'),'w') for lineInfo in RelIndexFitReport: line = 'Relation: '+ str(lineInfo[0])+'\t' +' Relation Number: '+ str(lineInfo[1])+'\t' +' sliceNorm: '+str(lineInfo[2])+'\t' +'errorNorm: '+str(lineInfo[3])+'\t'+' SlicewiseFit: '+str(lineInfo[4])+'\n' print(line) RelationReportHandle.write(line) RelationReportHandle.close() return RelIndexFitReport
def rais(self, data, step = 1000, M = 100, parallel = False, seed = None): num_data = data.shape[0] result = 0 if not parallel: p = [] for i in range(M): logw = self.mcmc_r(data, step, num_data) p.append(logw) p = np.array(p) logmeanp = logmeanexp(p, axis = 0) else: num_cores = multiprocessing.cpu_count() p = Parallel(n_jobs=num_cores)(delayed(self.mcmc_r)(v = data, step = step, num_data = num_data, seed = seed) for i in range(M)) p = np.array(p) logmeanp = logmeanexp(p, axis = 0) result = logmeanp.mean() return result
def ais(rbm, step = 100, M = 100, parallel = False, seed = None): W = rbm.W.data.numpy().T v_bias = rbm.v_bias.data.numpy() h_bias = rbm.h_bias.data.numpy() logZ0 = np.log((1+np.exp(v_bias))).sum() + np.log(1+np.exp(h_bias)).sum() ratio = [] if parallel: num_cores = multiprocessing.cpu_count() results = Parallel(n_jobs=num_cores)(delayed(mcmc)(step = step, seed = seed, W = W, h_bias = h_bias, v_bias = v_bias) for i in range(M)) results = np.array(results).reshape(len(results), 1) logZ = logZ0 + logmeanexp(results, axis = 0) else: for i in range(M): ratio.append(mcmc(step, seed = seed, W = W, h_bias = h_bias, v_bias = v_bias)) ratio = np.array(ratio).reshape(len(ratio),1) logZ = logZ0 + logmeanexp(ratio, axis = 0) return logZ
def ulogprob(v_input, dbn, M = 1000, parallel = False): logw = np.zeros([M, len(v_input)]) # samples = v_input if not parallel: for i in range(M): # samples = v_input # for l in range(dbn.n_layers-1): # logw[i,:] += -dbn.rbm_layers[l].free_energy(samples,dbn.rbm_layers[l].W)[0] # samples = dbn.rbm_layers[l].sample_h_given_v(samples,dbn.rbm_layers[l].W,dbn.rbm_layers[l].h_bias)[0] # logw[i,:] -= -dbn.rbm_layers[l].free_energy_hidden(samples,dbn.rbm_layers[l].W)[0] # logw[i,:] += -dbn.rbm_layers[-1].free_energy(samples,dbn.rbm_layers[-1].W)[0] logw[i,:] += important_sampling(v_input, dbn) else: num_cores = multiprocessing.cpu_count() results = Parallel(n_jobs=num_cores)(delayed(important_sampling)(v_input = v_input, dbn = dbn) for i in range(M)) logw += np.asarray(results) return logmeanexp(logw,0)
def main(): """ Main function. 1. Setup logging 2. Get arguments 3. Get index 4. Process files 5. Write output """ setup_logging() logger = logging.getLogger("stats." + __name__) args = get_args() index = get_index(args) logger.warning("Positions not in annotation will be ignored.") logger.info("Found " + str(len(args.inputs)) + " input file(s):") for input_file in sorted(args.inputs): logger.debug(input_file) if args.is_parallel: stats = Parallel(n_jobs=args.parallel, verbose=100, batch_size=1)(delayed(process_file)(input_file, args.type, index, args.is_parallel) for input_file in args.inputs) else: stats = [] for input_file in args.inputs: output_table = process_file(input_file, args.type, index, args.is_parallel) stats.append(output_table) write_stats(args.out, stats)
def _process_with_joblib(self, X: Union[pd.DataFrame, np.ndarray], n_refs: int, cluster_array: np.ndarray): """ Process calling of .calculate_gap() method using the joblib backend """ with Parallel(n_jobs=self.n_jobs) as parallel: for gap_value, n_clusters in parallel(delayed(self._calculate_gap)(X, n_refs, n_clusters) for n_clusters in cluster_array): yield (gap_value, n_clusters)
def forward_parallel(forward_process,X,n,n_jobs,extra_par): n_row=X.shape[0] parallelizer = Parallel(n_jobs=n_jobs) tasks_iterator = ( delayed(forward_process)(X_row,n,extra_par) for X_row in np.split(X,n_row)) result = parallelizer( tasks_iterator ) # Merging the output of the jobs return np.vstack(result)
def transform(self, Xb, yb): X_n, y_n = super(AugmentBatchIterator, self).transform(Xb, yb) ret = Parallel(n_jobs=-1)(delayed(load_augment_im)(self, name, bb) for name, bb in zip(X_n, y_n)) Xb = np.asarray(map(lambda v: v[0], ret)) yb = np.asarray(map(lambda v: v[1], ret)) return Xb, yb
def transform(self, Xb, yb): X_n, yb = super(LazyBatchIterator, self).transform(Xb, yb) Xb = Parallel(n_jobs=-1)(delayed(load_im_f)(name) for name in X_n) Xb = np.asarray(Xb) return Xb, yb
def process_batch(image_db, label_db, fnames_b, y_b): print "Reading the images and labels" with Parallel(n_jobs=-1) as parallel: Xb = parallel(delayed(load_im_tuple) (fname, i) for i, fname in fnames_b) yb = parallel(delayed(load_y_tuple)(y, i) for i, y in y_b) print "Writing image data" _write_batch_lmdb(image_db, Xb) print "Writing label data" _write_batch_lmdb(label_db, yb)
def get_file_list(folder): names = os.listdir(folder) fnames = [] bboxes = [] bbox_names = map(lambda name: os.path.join( folder, name, '_bboxes.txt'), names) with Parallel(n_jobs=-1) as parallel: dfs = parallel(delayed(_extract_names_bboxes)(bname) for bname in bbox_names) df = pd.concat(dfs, ignore_index=True) df['Flag'] = df['Name'].map(lambda x: True if os.path.exists(x) else False) print "Initial number of images:", df['Name'].count() df = df[df['Flag'] == True] print "Total number of existing images:", df['Name'].count() return df['Name'].values, df['BBox'].values
def pickle_rel_data(): #Parallel(n_jobs=9, backend="multiprocessing")(delayed(pickle_rel)(rel_feature) for rel_feature in # ['rel_degree','rel_in_degree','rel_out_degree','rel_page_rank','rel_local_clust','rel_eigen_centr', # 'rel_hits_hub','rel_hits_authority','rel_kcore']) Parallel(n_jobs=3, backend="multiprocessing")(delayed(pickle_rel)(rel_feature) for rel_feature in ['rel_degree','rel_in_degree','rel_out_degree']) #Parallel(n_jobs=3, backend="multiprocessing")(delayed(pickle_rel)(rel_feature) for rel_feature in # ['rel_hits_hub','rel_hits_authority','rel_kcore']) #Parallel(n_jobs=3, backend="multiprocessing")(delayed(pickle_rel)(rel_feature) for rel_feature in # ['rel_page_rank','rel_local_clust','rel_eigen_centr'])
def pickle_vis_data(): pickle_vis_data_pandas() Parallel(n_jobs=5, backend="multiprocessing")(delayed(pickle_viz)(rel_feature) for rel_feature in ['infobox','lead','left-body','navbox', 'body']) #Parallel(n_jobs=3, backend="multiprocessing")(delayed(pickle_viz_positions)(rel_feature) for rel_feature in # ['links_postions_text','links_postions_x','links_postions_y'])
def batch_train(opt, round_index, round_train_data, round_valid_data, round_valid_weights=None, save_all=True, file_indices=None, return_acc_len=False, seq2seq=False): i = 0 perfs = [] M = len(round_train_data) while i < M: j = min(i + opt['num_machines'], M) cur_perfs = Parallel(n_jobs=j - i, backend='threading') \ (delayed(train)(opt, round_index, train_index, file_indices[train_index] if file_indices else train_index, round_train_data[train_index], round_valid_data[train_index], valid_weights=round_valid_weights[train_index] if round_valid_weights else None, save_all=save_all, return_acc_len=return_acc_len, seq2seq=seq2seq) \ for train_index in range(i, j)) perfs.extend(cur_perfs) i = j error_indices, valid_indices = [], [] for i, perf in enumerate(perfs): if perf == 0.0 or type(perf) == tuple and perf[0] == 0.0: error_indices.append(i) elif i < opt['num_machines']: valid_indices.append(i) M = len(error_indices) TMP_NUM_MACHINES = len(valid_indices) if M > 0 and TMP_NUM_MACHINES > 0: i = 0 error_perfs = [] while i < M: j = min(i + TMP_NUM_MACHINES, M) cur_perfs = Parallel(n_jobs=j - i, backend='threading') \ (delayed(train)(opt, round_index, valid_indices[train_index], file_indices[error_indices[train_index]] if file_indices else error_indices[train_index], round_train_data[error_indices[train_index]], round_valid_data[error_indices[train_index]], valid_weights=round_valid_weights[error_indices[train_index]] if round_valid_weights else None, save_all=save_all, return_acc_len=return_acc_len, seq2seq=seq2seq) \ for train_index in range(i, j)) error_perfs.extend(cur_perfs) i = j for i in range(M): perfs[error_indices[i]] = error_perfs[i] return perfs
