project_name
stringlengths
6
104
file_name
stringlengths
4
89
full_name
stringlengths
1
102
func_name
stringlengths
1
85
docstring
stringlengths
13
836
docstring_tokens
listlengths
4
122
code
stringlengths
23
39.7k
code_tokens
stringlengths
29
44.6k
url
int64
3
986k
huawei-noah/xingtian
learner.py
Learner.submit_algorithm
submit_algorithm
Submit an algorithm, to update algorithm instance description.
[ "Submit", "an", "algorithm,", "to", "update", "algorithm", "instance", "description." ]
def submit_algorithm(self, alg_instance, trainer_obj, shared_buff): self.alg = alg_instance self.trainer = trainer_obj self.shared_buff = shared_buff
['def', 'submit_algorithm(self,', 'alg_instance,', 'trainer_obj,', 'shared_buff):', 'self.alg', '=', 'alg_instance', 'self.trainer', '=', 'trainer_obj', 'self.shared_buff', '=', 'shared_buff']
962,212
huawei-noah/xingtian
remoter.py
remote_run
remote_run
Run command in remote node.
[ "Run", "command", "in", "remote", "node." ]
def remote_run(server_ip, host, passwd, cmd, remote_env): print('remote_env:', remote_env) _env_export = 'export PATH={}/bin:$PATH'.format(remote_env['conda']) if 'env' in remote_env.keys(): for (_key, _val) in remote_env['env'].items(): _env_export += '&& export {}={}'.format(_key, _val...
['def', 'remote_run(server_ip,', 'host,', 'passwd,', 'cmd,', 'remote_env):', "print('remote_env:',", 'remote_env)', '_env_export', '=', "'export", "PATH={}/bin:$PATH'.format(remote_env['conda'])", 'if', "'env'", 'in', 'remote_env.keys():', 'for', '(_key,', '_val)', 'in', "remote_env['env'].items():", '_env_export', '+=...
962,215
huawei-noah/xingtian
trainer.py
build_alg_with_trainer
build_alg_with_trainer
Build an algorithm instance with multi-process trainer.
[ "Build", "an", "algorithm", "instance", "with", "multi-process", "trainer." ]
def build_alg_with_trainer(alg_para, model_q, model_path, process_num): alg_para = deepcopy(alg_para) if process_num >= 2: shared_list_for_train = Manager().list() (alg, subprocess_instance) = start_multi_processes(alg_para, model_q, model_path, process_num, shared_list_for_train) else: ...
['def', 'build_alg_with_trainer(alg_para,', 'model_q,', 'model_path,', 'process_num):', 'alg_para', '=', 'deepcopy(alg_para)', 'if', 'process_num', '>=', '2:', 'shared_list_for_train', '=', 'Manager().list()', '(alg,', 'subprocess_instance)', '=', 'start_multi_processes(alg_para,', 'model_q,', 'model_path,', 'process_n...
962,218
huawei-noah/xingtian
trainer.py
start_multi_processes
start_multi_processes
Start multi processes to train.
[ "Start", "multi", "processes", "to", "train." ]
def start_multi_processes(alg_para, model_q, model_path, process_num, train_list): array_list = init_memory(process_num) event_dict = {} grad_q = Queue() for i in range(process_num): event_dict[i] = Event() weight_list = init_memory(1) grad_process = [Process(target=grad_communicate, arg...
['def', 'start_multi_processes(alg_para,', 'model_q,', 'model_path,', 'process_num,', 'train_list):', 'array_list', '=', 'init_memory(process_num)', 'event_dict', '=', '{}', 'grad_q', '=', 'Queue()', 'for', 'i', 'in', 'range(process_num):', 'event_dict[i]', '=', 'Event()', 'weight_list', '=', 'init_memory(1)', 'grad_pr...
962,219
huawei-noah/xingtian
model.py
check_keep_model
check_keep_model
Check model saved count under path.
[ "Check", "model", "saved", "count", "under", "path." ]
def check_keep_model(model_path, keep_num): target_file = glob.glob(os.path.join(model_path, 'actor*'.format(model_path))) if len(target_file) > keep_num: to_rm_model = sorted(target_file, reverse=True)[keep_num:] for item in to_rm_model: os.remove(item)
['def', 'check_keep_model(model_path,', 'keep_num):', 'target_file', '=', 'glob.glob(os.path.join(model_path,', "'actor*'.format(model_path)))", 'if', 'len(target_file)', '>', 'keep_num:', 'to_rm_model', '=', 'sorted(target_file,', 'reverse=True)[keep_num:]', 'for', 'item', 'in', 'to_rm_model:', 'os.remove(item)']
962,221
huawei-noah/xingtian
model.py
XTModel.set_weights
set_weights
Set weight with memory tensor.
[ "Set", "weight", "with", "memory", "tensor." ]
def set_weights(self, weights): with self.graph.as_default(): self.actor_var.set_weights(weights)
['def', 'set_weights(self,', 'weights):', 'with', 'self.graph.as_default():', 'self.actor_var.set_weights(weights)']
962,224
huawei-noah/xingtian
model_utils.py
get_mlp_default_settings
get_mlp_default_settings
Get default setting for mlp model.
[ "Get", "default", "setting", "for", "mlp", "model." ]
def get_mlp_default_settings(kind): if kind == 'hidden_sizes': return [64, 64] elif kind == 'activation': return 'tanh' else: raise KeyError('unknown type: {}'.format(kind))
['def', 'get_mlp_default_settings(kind):', 'if', 'kind', '==', "'hidden_sizes':", 'return', '[64,', '64]', 'elif', 'kind', '==', "'activation':", 'return', "'tanh'", 'else:', 'raise', "KeyError('unknown", 'type:', "{}'.format(kind))"]
962,226
huawei-noah/xingtian
model_utils.py
custom_norm_initializer
custom_norm_initializer
Perform Customize norm initializer for op.
[ "Perform", "Customize", "norm", "initializer", "for", "op." ]
def custom_norm_initializer(std=0.5): def _initializer(shape, dtype=None, partition_info=None): out = np.random.randn(*shape).astype(np.float32) out *= std / np.sqrt(np.square(out).sum(axis=0, keepdims=True)) return tf.constant(out) return _initializer
['def', 'custom_norm_initializer(std=0.5):', 'def', '_initializer(shape,', 'dtype=None,', 'partition_info=None):', 'out', '=', 'np.random.randn(*shape).astype(np.float32)', 'out', '*=', 'std', '/', 'np.sqrt(np.square(out).sum(axis=0,', 'keepdims=True))', 'return', 'tf.constant(out)', 'return', '_initializer']
962,229
huawei-noah/xingtian
model_zeus.py
XTModelZeus.predict
predict
Do predict use the latest model.
[ "Do", "predict", "use", "the", "latest", "model." ]
def predict(self, state): return self.model.predict(state)
['def', 'predict(self,', 'state):', 'return', 'self.model.predict(state)']
962,231
huawei-noah/xingtian
tf_compat.py
import_tf_compact
import_tf_compact
Import tensorflow with compact behavior.
[ "Import", "tensorflow", "with", "compact", "behavior." ]
def import_tf_compact(): if 'tensorflow' not in sys.modules: try: import tensorflow.compat.v1 as tf tf.disable_v2_behavior() except ImportError: import tensorflow as tf tf.logging.set_verbosity(tf.logging.ERROR) return tf else: return s...
['def', 'import_tf_compact():', 'if', "'tensorflow'", 'not', 'in', 'sys.modules:', 'try:', 'import', 'tensorflow.compat.v1', 'as', 'tf', 'tf.disable_v2_behavior()', 'except', 'ImportError:', 'import', 'tensorflow', 'as', 'tf', 'tf.logging.set_verbosity(tf.logging.ERROR)', 'return', 'tf', 'else:', 'return', "sys.modules...
962,234
huawei-noah/xingtian
tf_compat.py
get_tf_major
get_tf_major
Get major of tensorflow version.
[ "Get", "major", "of", "tensorflow", "version." ]
def get_tf_major(): return int(tf.__version__.split('.')[0])
['def', 'get_tf_major():', 'return', "int(tf.__version__.split('.')[0])"]
962,237
huawei-noah/xingtian
tf_dist.py
ActionDist.sample
sample
Sample action from this distribution.
[ "Sample", "action", "from", "this", "distribution." ]
def sample(self, repeat): raise NotImplementedError
['def', 'sample(self,', 'repeat):', 'raise', 'NotImplementedError']
962,238
huawei-noah/xingtian
tf_utils.py
norm_initializer
norm_initializer
Build customized norm initializer.
[ "Build", "customized", "norm", "initializer." ]
def norm_initializer(std=0.5): def _initializer(shape, dtype=None, partition_info=None): out = np.random.randn(*shape).astype(np.float32) out *= std / np.sqrt(np.square(out).sum(axis=0, keepdims=True)) return tf.constant(out) return _initializer
['def', 'norm_initializer(std=0.5):', 'def', '_initializer(shape,', 'dtype=None,', 'partition_info=None):', 'out', '=', 'np.random.randn(*shape).astype(np.float32)', 'out', '*=', 'std', '/', 'np.sqrt(np.square(out).sum(axis=0,', 'keepdims=True))', 'return', 'tf.constant(out)', 'return', '_initializer']
962,240
huawei-noah/xingtian
tf_utils.py
TFVariables.get_weights
get_weights
Get weights with dict type.
[ "Get", "weights", "with", "dict", "type." ]
def get_weights(self): _weights = self.session.run(self.node_hub_with_order) return _weights
['def', 'get_weights(self):', '_weights', '=', 'self.session.run(self.node_hub_with_order)', 'return', '_weights']
962,242
huawei-noah/xingtian
tf_utils.py
TFVariables.set_weights
set_weights
Set weights with dict type.
[ "Set", "weights", "with", "dict", "type." ]
def set_weights(self, to_weights): nodes_to_assign = [self._to_assign_node_dict[node_name] for node_name in to_weights.keys() if node_name in self._to_assign_node_dict] if not nodes_to_assign: print('to_weights: ', to_weights) raise KeyError("NO node's weights could assign in self.graph {} vs {}...
['def', 'set_weights(self,', 'to_weights):', 'nodes_to_assign', '=', '[self._to_assign_node_dict[node_name]', 'for', 'node_name', 'in', 'to_weights.keys()', 'if', 'node_name', 'in', 'self._to_assign_node_dict]', 'if', 'not', 'nodes_to_assign:', "print('to_weights:", "',", 'to_weights)', 'raise', 'KeyError("NO', "node's...
962,243
huawei-noah/xingtian
tf_utils.py
TFVariables.set_weights_with_npz
set_weights_with_npz
Set weight with numpy file.
[ "Set", "weight", "with", "numpy", "file." ]
def set_weights_with_npz(self, npz_file: str): weights = self.read_weights(npz_file) self.set_weights(weights)
['def', 'set_weights_with_npz(self,', 'npz_file:', 'str):', 'weights', '=', 'self.read_weights(npz_file)', 'self.set_weights(weights)']
962,246
huawei-noah/xingtian
dqn_cnn.py
DqnCnn.create_model
create_model
Create Deep-Q CNN network.
[ "Create", "Deep-Q", "CNN", "network." ]
def create_model(self, model_info): state = Input(shape=self.state_dim, dtype='uint8') state1 = Lambda(lambda x: K.cast(x, dtype='float32') / 255.0)(state) convlayer = Conv2D(32, (8, 8), strides=(4, 4), activation='relu', padding='valid')(state1) convlayer = Conv2D(64, (4, 4), strides=(2, 2), activation...
['def', 'create_model(self,', 'model_info):', 'state', '=', 'Input(shape=self.state_dim,', "dtype='uint8')", 'state1', '=', 'Lambda(lambda', 'x:', 'K.cast(x,', "dtype='float32')", '/', '255.0)(state)', 'convlayer', '=', 'Conv2D(32,', '(8,', '8),', 'strides=(4,', '4),', "activation='relu',", "padding='valid')(state1)", ...
962,248
huawei-noah/xingtian
dqn_mlp.py
layer_add
layer_add
Compute Q given Advantage and V.
[ "Compute", "Q", "given", "Advantage", "and", "V." ]
def layer_add(x): return x[0] + x[1]
['def', 'layer_add(x):', 'return', 'x[0]', '+', 'x[1]']
962,251
huawei-noah/xingtian
impala_cnn.py
impala_loss
impala_loss
Compute loss for impala.
[ "Compute", "loss", "for", "impala." ]
def impala_loss(advantage): def loss(y_true, y_pred): policy = y_pred log_policy = K.log(policy + 1e-10) entropy = -policy * K.log(policy + 1e-10) cross_entropy = -y_true * log_policy return K.mean(advantage * cross_entropy - ENTROPY_LOSS * entropy, 1) return loss
['def', 'impala_loss(advantage):', 'def', 'loss(y_true,', 'y_pred):', 'policy', '=', 'y_pred', 'log_policy', '=', 'K.log(policy', '+', '1e-10)', 'entropy', '=', '-policy', '*', 'K.log(policy', '+', '1e-10)', 'cross_entropy', '=', '-y_true', '*', 'log_policy', 'return', 'K.mean(advantage', '*', 'cross_entropy', '-', 'EN...
962,255
huawei-noah/xingtian
impala_mlp.py
impala_loss
impala_loss
Compute loss for IMPALA.
[ "Compute", "loss", "for", "IMPALA." ]
def impala_loss(advantage): def loss(y_true, y_pred): policy = y_pred log_policy = K.log(policy + 1e-10) entropy = -policy * log_policy cross_entropy = -y_true * log_policy return K.mean(advantage * cross_entropy - ENTROPY_LOSS * entropy) return loss
['def', 'impala_loss(advantage):', 'def', 'loss(y_true,', 'y_pred):', 'policy', '=', 'y_pred', 'log_policy', '=', 'K.log(policy', '+', '1e-10)', 'entropy', '=', '-policy', '*', 'log_policy', 'cross_entropy', '=', '-y_true', '*', 'log_policy', 'return', 'K.mean(advantage', '*', 'cross_entropy', '-', 'ENTROPY_LOSS', '*',...
962,264
huawei-noah/xingtian
muzero_utils.py
scale_gradient
scale_gradient
Scales the gradient for the backward pass.
[ "Scales", "the", "gradient", "for", "the", "backward", "pass." ]
def scale_gradient(tensor, scale): return tensor * scale + tf.stop_gradient(tensor) * (1 - scale)
['def', 'scale_gradient(tensor,', 'scale):', 'return', 'tensor', '*', 'scale', '+', 'tf.stop_gradient(tensor)', '*', '(1', '-', 'scale)']
962,275
huawei-noah/xingtian
ppo_mlp_zeus.py
value_loss
value_loss
Compute value loss for PPO.
[ "Compute", "value", "loss", "for", "PPO." ]
def value_loss(target_v, out_v, old_v): vpredclipped = old_v + tf.clip_by_value(out_v - old_v, -VF_CLIP, VF_CLIP) vf_losses1 = tf.square(out_v - target_v) vf_losses2 = tf.square(vpredclipped - target_v) vf_loss = 0.5 * tf.reduce_mean(tf.maximum(vf_losses1, vf_losses2)) return vf_loss
['def', 'value_loss(target_v,', 'out_v,', 'old_v):', 'vpredclipped', '=', 'old_v', '+', 'tf.clip_by_value(out_v', '-', 'old_v,', '-VF_CLIP,', 'VF_CLIP)', 'vf_losses1', '=', 'tf.square(out_v', '-', 'target_v)', 'vf_losses2', '=', 'tf.square(vpredclipped', '-', 'target_v)', 'vf_loss', '=', '0.5', '*', 'tf.reduce_mean(tf....
962,276
huawei-noah/xingtian
qmix_tf.py
QMixModel.build_actor_graph
build_actor_graph
Build explorer graph with minimum principle.
[ "Build", "explorer", "graph", "with", "minimum", "principle." ]
def build_actor_graph(self): with self.graph.as_default(): with tf.variable_scope('explore_agent'): (self.agent_outs, self.hidden_outs) = self.build_agent_net(inputs_obs=self.ph_obs, seq_max=1, obs_lengths=[1 for _ in range(self.n_agents)], hidden_state_in=self.ph_hidden_states_in) self....
['def', 'build_actor_graph(self):', 'with', 'self.graph.as_default():', 'with', "tf.variable_scope('explore_agent'):", '(self.agent_outs,', 'self.hidden_outs)', '=', 'self.build_agent_net(inputs_obs=self.ph_obs,', 'seq_max=1,', 'obs_lengths=[1', 'for', '_', 'in', 'range(self.n_agents)],', 'hidden_state_in=self.ph_hidde...
962,280
huawei-noah/xingtian
qmix_tf.py
QMixModel.reset_hidden_state
reset_hidden_state
Reset hidden state with value assign.
[ "Reset", "hidden", "state", "with", "value", "assign." ]
def reset_hidden_state(self): self.hi_out_val = self.hi_out_val_default
['def', 'reset_hidden_state(self):', 'self.hi_out_val', '=', 'self.hi_out_val_default']
962,282
huawei-noah/xingtian
__init__.py
register_zeus
register_zeus
Import and register zeus modules automatically.
[ "Import", "and", "register", "zeus", "modules", "automatically." ]
def register_zeus(backend): from zeus.datasets import register_datasets from zeus.modules import register_modules from zeus.networks import register_networks from zeus.evaluator import register_evaluator from zeus.trainer import register_trainer, trainer_api from zeus.metrics import register_met...
['def', 'register_zeus(backend):', 'from', 'zeus.datasets', 'import', 'register_datasets', 'from', 'zeus.modules', 'import', 'register_modules', 'from', 'zeus.networks', 'import', 'register_networks', 'from', 'zeus.evaluator', 'import', 'register_evaluator', 'from', 'zeus.trainer', 'import', 'register_trainer,', 'train...
962,302
huawei-noah/xingtian
__init__.py
is_torch_backend
is_torch_backend
Return whether is pytorch backend or not.
[ "Return", "whether", "is", "pytorch", "backend", "or", "not." ]
def is_torch_backend(): return os.environ.get('BACKEND_TYPE', None) == 'PYTORCH'
['def', 'is_torch_backend():', 'return', "os.environ.get('BACKEND_TYPE',", 'None)', '==', "'PYTORCH'"]
962,306
huawei-noah/xingtian
__init__.py
is_tf_backend
is_tf_backend
Return whether is tensorflow backend or not.
[ "Return", "whether", "is", "tensorflow", "backend", "or", "not." ]
def is_tf_backend(): return os.environ.get('BACKEND_TYPE', None) == 'TENSORFLOW'
['def', 'is_tf_backend():', 'return', "os.environ.get('BACKEND_TYPE',", 'None)', '==', "'TENSORFLOW'"]
962,307
huawei-noah/xingtian
config.py
build_tree
build_tree
Convert plaint dictionary to a tree dictionary.
[ "Convert", "plaint", "dictionary", "to", "a", "tree", "dictionary." ]
def build_tree(data): result = {} for (key, value) in data.items(): if '.' in key: _keys = key.split('.') _tree = {} _tree[_keys[-1]] = value _keys.reverse() for sub_key in _keys[1:]: _tree = {sub_key: _tree} branch ...
['def', 'build_tree(data):', 'result', '=', '{}', 'for', '(key,', 'value)', 'in', 'data.items():', 'if', "'.'", 'in', 'key:', '_keys', '=', "key.split('.')", '_tree', '=', '{}', '_tree[_keys[-1]]', '=', 'value', '_keys.reverse()', 'for', 'sub_key', 'in', '_keys[1:]:', '_tree', '=', '{sub_key:', '_tree}', 'branch', '=',...
962,314
huawei-noah/xingtian
config_serializable.py
ConfigSerializable.rules
rules
Return rules for checking.
[ "Return", "rules", "for", "checking." ]
def rules(cls): return {}
['def', 'rules(cls):', 'return', '{}']
962,317
huawei-noah/xingtian
config_serializable.py
ConfigSerializable.backup_original_value
backup_original_value
Backup class original data.
[ "Backup", "class", "original", "data." ]
def backup_original_value(cls, force=False): if not cls.__original__value__ or force: cls.__original__value__ = cls().to_json() return cls.__original__value__
['def', 'backup_original_value(cls,', 'force=False):', 'if', 'not', 'cls.__original__value__', 'or', 'force:', 'cls.__original__value__', '=', 'cls().to_json()', 'return', 'cls.__original__value__']
962,318
huawei-noah/xingtian
task_ops.py
TaskOps.model_zoo_path
model_zoo_path
Return model zoo path.
[ "Return", "model", "zoo", "path." ]
def model_zoo_path(self): return General.model_zoo.model_zoo_path
['def', 'model_zoo_path(self):', 'return', 'General.model_zoo.model_zoo_path']
962,341
huawei-noah/xingtian
user_config.py
UserConfig.merge_reference
merge_reference
Merge config with reference the specified config with ref item.
[ "Merge", "config", "with", "reference", "the", "specified", "config", "with", "ref", "item." ]
def merge_reference(child): if not isinstance(child, dict): return ref = child.get('ref') if not ref: return ref_dict = deepcopy(UserConfig().data) for key in ref.split('.'): ref_dict = ref_dict.get(key) not_merge_keys = ['callbacks', 'lazy_built'] for key in not_merg...
['def', 'merge_reference(child):', 'if', 'not', 'isinstance(child,', 'dict):', 'return', 'ref', '=', "child.get('ref')", 'if', 'not', 'ref:', 'return', 'ref_dict', '=', 'deepcopy(UserConfig().data)', 'for', 'key', 'in', "ref.split('.'):", 'ref_dict', '=', 'ref_dict.get(key)', 'not_merge_keys', '=', "['callbacks',", "'l...
962,347
huawei-noah/xingtian
utils.py
copy_search_file
copy_search_file
Copy files from srcDir to desDir.
[ "Copy", "files", "from", "srcDir", "to", "desDir." ]
def copy_search_file(srcDir, desDir): ls = os.listdir(srcDir) for line in ls: filePath = os.path.join(srcDir, line) if os.path.isfile(filePath): shutil.copy(filePath, desDir)
['def', 'copy_search_file(srcDir,', 'desDir):', 'ls', '=', 'os.listdir(srcDir)', 'for', 'line', 'in', 'ls:', 'filePath', '=', 'os.path.join(srcDir,', 'line)', 'if', 'os.path.isfile(filePath):', 'shutil.copy(filePath,', 'desDir)']
962,354
huawei-noah/xingtian
message.py
get_msg_info
get_msg_info
Get message ctr info.
[ "Get", "message", "ctr", "info." ]
def get_msg_info(msg, key): return msg['ctr_info'].get(key)
['def', 'get_msg_info(msg,', 'key):', 'return', "msg['ctr_info'].get(key)"]
962,356
huawei-noah/xingtian
message.py
set_msg_info
set_msg_info
Set message ctr info.
[ "Set", "message", "ctr", "info." ]
def set_msg_info(msg, **kwargs): msg['ctr_info'].update(**kwargs)
['def', 'set_msg_info(msg,', '**kwargs):', "msg['ctr_info'].update(**kwargs)"]
962,357
huawei-noah/xingtian
share_buffer.py
test_buf_get_live
test_buf_get_live
Test share buf live count.
[ "Test", "share", "buf", "live", "count." ]
def test_buf_get_live(): live_count = 10 logging.set_verbosity(logging.DEBUG) share_buf = ShareBuf(live=live_count, size=20000000, start=True) data = {'d{}'.format(i): np.array(np.arange(i)) for i in range(5, 8)} ds = serialize(data).to_buffer() b_id = share_buf.put(data_buffer=ds) for _ in ...
['def', 'test_buf_get_live():', 'live_count', '=', '10', 'logging.set_verbosity(logging.DEBUG)', 'share_buf', '=', 'ShareBuf(live=live_count,', 'size=20000000,', 'start=True)', 'data', '=', "{'d{}'.format(i):", 'np.array(np.arange(i))', 'for', 'i', 'in', 'range(5,', '8)}', 'ds', '=', 'serialize(data).to_buffer()', 'b_i...
962,358
huawei-noah/xingtian
share_buffer.py
test_share_buf_io
test_share_buf_io
Test share buf io-out.
[ "Test", "share", "buf", "io-out." ]
def test_share_buf_io(): logging.set_verbosity(logging.DEBUG) share_buf = ShareBuf(live=10, size=20000000, start=True) data = {'d{}'.format(i): np.array(np.arange(i)) for i in range(5, 8)} print(data) ds = serialize(data).to_buffer() b_id = share_buf.put(data_buffer=ds) print('b_id', b_id) ...
['def', 'test_share_buf_io():', 'logging.set_verbosity(logging.DEBUG)', 'share_buf', '=', 'ShareBuf(live=10,', 'size=20000000,', 'start=True)', 'data', '=', "{'d{}'.format(i):", 'np.array(np.arange(i))', 'for', 'i', 'in', 'range(5,', '8)}', 'print(data)', 'ds', '=', 'serialize(data).to_buffer()', 'b_id', '=', 'share_bu...
962,359
huawei-noah/xingtian
share_buffer.py
ShareBuf.plus_one_live
plus_one_live
Add one live value.
[ "Add", "one", "live", "value." ]
def plus_one_live(self): self.live_threshold += 1 self._update_vanish_attr(self.live_threshold)
['def', 'plus_one_live(self):', 'self.live_threshold', '+=', '1', 'self._update_vanish_attr(self.live_threshold)']
962,361
huawei-noah/xingtian
share_buffer.py
ShareBuf.reduce_once
reduce_once
Reduce one times of this object.
[ "Reduce", "one", "times", "of", "this", "object." ]
def reduce_once(self, object_id): if object_id not in self.live_info: logging.debug('obj_id: {} is deleted yet'.format(object_id)) else: self.live_info[object_id] -= 1
['def', 'reduce_once(self,', 'object_id):', 'if', 'object_id', 'not', 'in', 'self.live_info:', "logging.debug('obj_id:", '{}', 'is', 'deleted', "yet'.format(object_id))", 'else:', 'self.live_info[object_id]', '-=', '1']
962,363
huawei-noah/xingtian
share_buffer.py
ShareBuf.put
put
Put data buffer for share.
[ "Put", "data", "buffer", "for", "share." ]
def put(self, data_buffer, special_live=None): client = self.connect() object_id = client.put_raw_buffer(data_buffer) self._init_obj(object_id.binary(), special_live) ready_vanish_ids = self._get_vanish_obj() if ready_vanish_ids: client.delete(ready_vanish_ids) return object_id.binary()
['def', 'put(self,', 'data_buffer,', 'special_live=None):', 'client', '=', 'self.connect()', 'object_id', '=', 'client.put_raw_buffer(data_buffer)', 'self._init_obj(object_id.binary(),', 'special_live)', 'ready_vanish_ids', '=', 'self._get_vanish_obj()', 'if', 'ready_vanish_ids:', 'client.delete(ready_vanish_ids)', 're...
962,364
huawei-noah/xingtian
share_buffer.py
ShareBuf.get_with_live_consume
get_with_live_consume
Get a object data from plasma server with id, and reduce live count.
[ "Get", "a", "object", "data", "from", "plasma", "server", "with", "id,", "and", "reduce", "live", "count." ]
def get_with_live_consume(self, object_id_byte): data = self._get_buf(object_id_byte) self.reduce_once(object_id_byte) return data
['def', 'get_with_live_consume(self,', 'object_id_byte):', 'data', '=', 'self._get_buf(object_id_byte)', 'self.reduce_once(object_id_byte)', 'return', 'data']
962,366
huawei-noah/xingtian
share_by_plasma.py
ShareByPlasma.send
send
Send data to plasma server.
[ "Send", "data", "to", "plasma", "server." ]
def send(self, data, name=None, block=True): data_buffer = serialize(data['data']).to_buffer() compress_type = data['ctr_info'].get('compress_type', 'auto') if compress_type in ['auto', 'compress']: if sys.getsizeof(bytes(data_buffer)) > self.compress_threhold or compress_type == 'compress': ...
['def', 'send(self,', 'data,', 'name=None,', 'block=True):', 'data_buffer', '=', "serialize(data['data']).to_buffer()", 'compress_type', '=', "data['ctr_info'].get('compress_type',", "'auto')", 'if', 'compress_type', 'in', "['auto',", "'compress']:", 'if', 'sys.getsizeof(bytes(data_buffer))', '>', 'self.compress_threho...
962,369
huawei-noah/xingtian
share_by_plasma.py
ShareByPlasma.recv
recv
Receive data from plasma server.
[ "Receive", "data", "from", "plasma", "server." ]
def recv(self, name=None, block=True): if not block and self.control_q.empty(): return None ctr_info = self.control_q.get() object_id = ctr_info['object_id'] compress_flag = ctr_info.get('compress_flag', False) client = self.connect() data = client.get_buffers([object_id])[0] if comp...
['def', 'recv(self,', 'name=None,', 'block=True):', 'if', 'not', 'block', 'and', 'self.control_q.empty():', 'return', 'None', 'ctr_info', '=', 'self.control_q.get()', 'object_id', '=', "ctr_info['object_id']", 'compress_flag', '=', "ctr_info.get('compress_flag',", 'False)', 'client', '=', 'self.connect()', 'data', '=',...
962,370
huawei-noah/xingtian
share_by_raw_array.py
ShareByRawArray.recv
recv
Get data from share memory.
[ "Get", "data", "from", "share", "memory." ]
def recv(self, name=None): (data_id, len_data) = self.control_q.get() data = pyarrow.deserialize(lz4.frame.decompress(memoryview(self.mem)[int(data_id * self.size_mem_agent):int(data_id * self.size_mem_agent + len_data)])) return data
['def', 'recv(self,', 'name=None):', '(data_id,', 'len_data)', '=', 'self.control_q.get()', 'data', '=', 'pyarrow.deserialize(lz4.frame.decompress(memoryview(self.mem)[int(data_id', '*', 'self.size_mem_agent):int(data_id', '*', 'self.size_mem_agent', '+', 'len_data)]))', 'return', 'data']
962,377
huawei-noah/xingtian
share_by_raw_array.py
ShareByRawArray.recv_bytes
recv_bytes
Get data from share memory without deserialize.
[ "Get", "data", "from", "share", "memory", "without", "deserialize." ]
def recv_bytes(self, block): (data_id, len_data) = self.control_q.get() return memoryview(self.mem)[int(data_id * self.size_mem_agent):int(data_id * self.size_mem_agent + len_data)]
['def', 'recv_bytes(self,', 'block):', '(data_id,', 'len_data)', '=', 'self.control_q.get()', 'return', 'memoryview(self.mem)[int(data_id', '*', 'self.size_mem_agent):int(data_id', '*', 'self.size_mem_agent', '+', 'len_data)]']
962,378
huawei-noah/xingtian
share_by_raw_array.py
ShareByRawArray.send_bytes
send_bytes
Put data in share memory without serialize.
[ "Put", "data", "in", "share", "memory", "without", "serialize." ]
def send_bytes(self, data): (data_id, data_buffer) = data memmove(addressof(self.mem) + int(data_id) * self.size_mem_agent, data_buffer, len(data_buffer)) self.control_q.put((data_id, len(data_buffer)))
['def', 'send_bytes(self,', 'data):', '(data_id,', 'data_buffer)', '=', 'data', 'memmove(addressof(self.mem)', '+', 'int(data_id)', '*', 'self.size_mem_agent,', 'data_buffer,', 'len(data_buffer))', 'self.control_q.put((data_id,', 'len(data_buffer)))']
962,379
huawei-noah/xingtian
uni_comm.py
UniComm.send
send
Create common send interface.
[ "Create", "common", "send", "interface." ]
def send(self, data, name=None, block=True, **kwargs): return self.comm.send(data, name, block, **kwargs)
['def', 'send(self,', 'data,', 'name=None,', 'block=True,', '**kwargs):', 'return', 'self.comm.send(data,', 'name,', 'block,', '**kwargs)']
962,385
huawei-noah/xingtian
uni_comm.py
UniComm.send_multipart
send_multipart
Create common send_multipart interface.
[ "Create", "common", "send_multipart", "interface." ]
def send_multipart(self, data): return self.comm.send_multipart(data)
['def', 'send_multipart(self,', 'data):', 'return', 'self.comm.send_multipart(data)']
962,389
huawei-noah/xingtian
uni_comm.py
UniComm.recv_multipart
recv_multipart
Create common recv_multipart interface.
[ "Create", "common", "recv_multipart", "interface." ]
def recv_multipart(self): return self.comm.recv_multipart()
['def', 'recv_multipart(self):', 'return', 'self.comm.recv_multipart()']
962,390
huawei-noah/xingtian
benchmark_data.py
Data.get_version
get_version
Get database version info.
[ "Get", "database", "version", "info." ]
def get_version(self): return self.VERSION
['def', 'get_version(self):', 'return', 'self.VERSION']
962,391
huawei-noah/xingtian
check.py
make_rules
make_rules
Make new rule in dict for attr.
[ "Make", "new", "rule", "in", "dict", "for", "attr." ]
def make_rules(adict, attr_name, if_required, types, scpoe=None): if attr_name not in adict: adict[attr_name] = {} adict[attr_name]['required'] = if_required adict[attr_name]['type'] = types if scpoe: adict[attr_name]['scope'] = scpoe return adict
['def', 'make_rules(adict,', 'attr_name,', 'if_required,', 'types,', 'scpoe=None):', 'if', 'attr_name', 'not', 'in', 'adict:', 'adict[attr_name]', '=', '{}', "adict[attr_name]['required']", '=', 'if_required', "adict[attr_name]['type']", '=', 'types', 'if', 'scpoe:', "adict[attr_name]['scope']", '=', 'scpoe', 'return',...
962,392
huawei-noah/xingtian
check.py
BaseChecking.check_all
check_all
Check rules for attr.
[ "Check", "rules", "for", "attr." ]
def check_all(cls, attr_name, rules, checked_cls_name, config): for subclass in cls.__subclasses__(): subclass.check(attr_name, rules, checked_cls_name, config)
['def', 'check_all(cls,', 'attr_name,', 'rules,', 'checked_cls_name,', 'config):', 'for', 'subclass', 'in', 'cls.__subclasses__():', 'subclass.check(attr_name,', 'rules,', 'checked_cls_name,', 'config)']
962,393
huawei-noah/xingtian
common.py
bytes_to_str
bytes_to_str
Bytes to string, used after data transform by internet.
[ "Bytes", "to", "string,", "used", "after", "data", "transform", "by", "internet." ]
def bytes_to_str(data): if isinstance(data, bytes): return data if sys.version_info.major == 2 else data.decode('ascii') if isinstance(data, dict): return dict(map(bytes_to_str, data.items())) if isinstance(data, tuple): return map(bytes_to_str, data) return data
['def', 'bytes_to_str(data):', 'if', 'isinstance(data,', 'bytes):', 'return', 'data', 'if', 'sys.version_info.major', '==', '2', 'else', "data.decode('ascii')", 'if', 'isinstance(data,', 'dict):', 'return', 'dict(map(bytes_to_str,', 'data.items()))', 'if', 'isinstance(data,', 'tuple):', 'return', 'map(bytes_to_str,', '...
962,400
huawei-noah/xingtian
common.py
get_host_ip
get_host_ip
Get local ip address.
[ "Get", "local", "ip", "address." ]
def get_host_ip(): try: s = socket.socket(socket.AF_INET, socket.SOCK_DGRAM) s.connect(('8.8.8.8', 80)) ip = s.getsockname()[0] finally: s.close() return ip
['def', 'get_host_ip():', 'try:', 's', '=', 'socket.socket(socket.AF_INET,', 'socket.SOCK_DGRAM)', "s.connect(('8.8.8.8',", '80))', 'ip', '=', 's.getsockname()[0]', 'finally:', 's.close()', 'return', 'ip']
962,401
huawei-noah/xingtian
evaluate_xt.py
make_workspace_if_not_exist
make_workspace_if_not_exist
Make workspace if not exist.
[ "Make", "workspace", "if", "not", "exist." ]
def make_workspace_if_not_exist(benchmark_args, subdir='models', task_name=None): (workspace, archive_root, bm_id) = _make_workspace(benchmark_args, task_postfix=task_name) make_dirs_if_not_exist(workspace) if isinstance(subdir, str): make_dirs_if_not_exist(os.path.join(workspace, subdir)) elif ...
['def', 'make_workspace_if_not_exist(benchmark_args,', "subdir='models',", 'task_name=None):', '(workspace,', 'archive_root,', 'bm_id)', '=', '_make_workspace(benchmark_args,', 'task_postfix=task_name)', 'make_dirs_if_not_exist(workspace)', 'if', 'isinstance(subdir,', 'str):', 'make_dirs_if_not_exist(os.path.join(works...
962,410
huawei-noah/xingtian
evaluate_xt.py
read_train_event_id
read_train_event_id
Read train event id.
[ "Read", "train", "event", "id." ]
def read_train_event_id(benchmark_args): (archive_root, bm_id) = _get_archive_bm_basic_info(benchmark_args) return fetch_train_event(archive_root, bm_id, single=True)
['def', 'read_train_event_id(benchmark_args):', '(archive_root,', 'bm_id)', '=', '_get_archive_bm_basic_info(benchmark_args)', 'return', 'fetch_train_event(archive_root,', 'bm_id,', 'single=True)']
962,412
huawei-noah/xingtian
evaluate_xt.py
get_bm_args_from_config
get_bm_args_from_config
Get bm args from config.
[ "Get", "bm", "args", "from", "config." ]
def get_bm_args_from_config(config): alg_para = config['alg_para'] env_para = config['env_para'] agent_para = config['agent_para'] model_info = config['model_para'] alg_para['model_info'] = model_info bm_info = config.get('benchmark', dict()) return parse_benchmark_args(env_para, alg_para, a...
['def', 'get_bm_args_from_config(config):', 'alg_para', '=', "config['alg_para']", 'env_para', '=', "config['env_para']", 'agent_para', '=', "config['agent_para']", 'model_info', '=', "config['model_para']", "alg_para['model_info']", '=', 'model_info', 'bm_info', '=', "config.get('benchmark',", 'dict())', 'return', 'pa...
962,413
huawei-noah/xingtian
evaluate_xt.py
read_train_records_from_config
read_train_records_from_config
Read train records from config.
[ "Read", "train", "records", "from", "config." ]
def read_train_records_from_config(config, use_index='step', stage='both'): bm_args = get_bm_args_from_config(config) return read_train_records(bm_args, use_index, stage)
['def', 'read_train_records_from_config(config,', "use_index='step',", "stage='both'):", 'bm_args', '=', 'get_bm_args_from_config(config)', 'return', 'read_train_records(bm_args,', 'use_index,', 'stage)']
962,414
huawei-noah/xingtian
get_xt_config.py
finditem
finditem
Find key in dict.
[ "Find", "key", "in", "dict." ]
def finditem(obj, key): if not isinstance(obj, dict): return None elif key in obj: return obj[key] for (k, v) in obj.items(): ret_obj = finditem(v, key) if ret_obj is not None: return ret_obj
['def', 'finditem(obj,', 'key):', 'if', 'not', 'isinstance(obj,', 'dict):', 'return', 'None', 'elif', 'key', 'in', 'obj:', 'return', 'obj[key]', 'for', '(k,', 'v)', 'in', 'obj.items():', 'ret_obj', '=', 'finditem(v,', 'key)', 'if', 'ret_obj', 'is', 'not', 'None:', 'return', 'ret_obj']
962,417
huawei-noah/xingtian
get_xt_config.py
parse_xt_multi_case_paras
parse_xt_multi_case_paras
Parse the multi-case config file entrance for benchmark.
[ "Parse", "the", "multi-case", "config", "file", "entrance", "for", "benchmark." ]
def parse_xt_multi_case_paras(config_file, key_fields=('alg_config', 'agent_config')): with open(config_file) as file_hander: yaml_obj = yaml.safe_load(file_hander) (parse_candidate, combination_count) = _get_combination_info(yaml_obj, key_fields) para_prod_val = _get_product_value(parse_candidate) ...
['def', 'parse_xt_multi_case_paras(config_file,', "key_fields=('alg_config',", "'agent_config')):", 'with', 'open(config_file)', 'as', 'file_hander:', 'yaml_obj', '=', 'yaml.safe_load(file_hander)', '(parse_candidate,', 'combination_count)', '=', '_get_combination_info(yaml_obj,', 'key_fields)', 'para_prod_val', '=', '...
962,418
huawei-noah/xingtian
hw_cloud_helper.py
sync_data_from_s3
sync_data_from_s3
Sync data from user's s3 path to local machine, auto-check the local path firstly.
[ "Sync", "data", "from", "user's", "s3", "path", "to", "local", "machine,", "auto-check", "the", "local", "path", "firstly." ]
def sync_data_from_s3(s3_path, destination): local_makedir_if_not_existed(destination) if not mox.file.is_directory(s3_path): mox.file.copy(s3_path, destination) else: mox.file.copy_parallel(s3_path, destination)
['def', 'sync_data_from_s3(s3_path,', 'destination):', 'local_makedir_if_not_existed(destination)', 'if', 'not', 'mox.file.is_directory(s3_path):', 'mox.file.copy(s3_path,', 'destination)', 'else:', 'mox.file.copy_parallel(s3_path,', 'destination)']
962,425
huawei-noah/xingtian
local_data.py
open_file
open_file
Need close by hand.
[ "Need", "close", "by", "hand." ]
def open_file(file_path, open_type): if file_path.startswith('s3://'): import moxing as mox ret_handle = mox.file.File(file_path, open_type) else: ret_handle = open(file_path, open_type) return ret_handle
['def', 'open_file(file_path,', 'open_type):', 'if', "file_path.startswith('s3://'):", 'import', 'moxing', 'as', 'mox', 'ret_handle', '=', 'mox.file.File(file_path,', 'open_type)', 'else:', 'ret_handle', '=', 'open(file_path,', 'open_type)', 'return', 'ret_handle']
962,426
huawei-noah/xingtian
logger.py
time_to_str
time_to_str
Convert seconds to days, hours, minutes and seconds.
[ "Convert", "seconds", "to", "days,", "hours,", "minutes", "and", "seconds." ]
def time_to_str(sec): (days, remainder) = divmod(sec, 60 * 60 * 24) (hours, remainder) = divmod(remainder, 60 * 60) (minutes, seconds) = divmod(remainder, 60) _str = '' if days > 0: _str += '{:d} days, '.format(int(days)) if hours > 0: _str += '{:d} hours, '.format(int(hours)) ...
['def', 'time_to_str(sec):', '(days,', 'remainder)', '=', 'divmod(sec,', '60', '*', '60', '*', '24)', '(hours,', 'remainder)', '=', 'divmod(remainder,', '60', '*', '60)', '(minutes,', 'seconds)', '=', 'divmod(remainder,', '60)', '_str', '=', "''", 'if', 'days', '>', '0:', '_str', '+=', "'{:d}", 'days,', "'.format(int(d...
962,429
huawei-noah/xingtian
logger.py
Logger.elapsed_time
elapsed_time
Elapsed time set as an property.
[ "Elapsed", "time", "set", "as", "an", "property." ]
def elapsed_time(self): return time() - self.abs_start
['def', 'elapsed_time(self):', 'return', 'time()', '-', 'self.abs_start']
962,430
huawei-noah/xingtian
logger.py
Logger.update
update
Update value could been rewrite.
[ "Update", "value", "could", "been", "rewrite." ]
def update(self, **kwargs): self.records.update(kwargs)
['def', 'update(self,', '**kwargs):', 'self.records.update(kwargs)']
962,431
huawei-noah/xingtian
logger.py
Logger.train_reward_avg
train_reward_avg
Train reward average could been property.
[ "Train", "reward", "average", "could", "been", "property." ]
def train_reward_avg(self): if not self.records['train_reward']: return np.nan return np.mean(self.records['train_reward'][-100:])
['def', 'train_reward_avg(self):', 'if', 'not', "self.records['train_reward']:", 'return', 'np.nan', 'return', "np.mean(self.records['train_reward'][-100:])"]
962,434
huawei-noah/xingtian
logger.py
StatsRecorder.could_show_stats
could_show_stats
Check whether show or not.
[ "Check", "whether", "show", "or", "not." ]
def could_show_stats(self): if self._data.get('step', 0) - self._last_show_step >= self.show_interval: self._last_show_step = self._data.get('step', 0) return True return False
['def', 'could_show_stats(self):', 'if', "self._data.get('step',", '0)', '-', 'self._last_show_step', '>=', 'self.show_interval:', 'self._last_show_step', '=', "self._data.get('step',", '0)', 'return', 'True', 'return', 'False']
962,438
huawei-noah/xingtian
logger.py
StatsRecorder.assemble_records
assemble_records
Assemble the data format for tensorboard.
[ "Assemble", "the", "data", "format", "for", "tensorboard." ]
def assemble_records(self): record_list = list() for _key in BOARD_GROUP_MAP.keys(): try: g_key = self.add_board_prefix(_key) if not self._data[_key]: continue if np.nan is self._data[_key]: continue record_list.append((g_ke...
['def', 'assemble_records(self):', 'record_list', '=', 'list()', 'for', '_key', 'in', 'BOARD_GROUP_MAP.keys():', 'try:', 'g_key', '=', 'self.add_board_prefix(_key)', 'if', 'not', 'self._data[_key]:', 'continue', 'if', 'np.nan', 'is', 'self._data[_key]:', 'continue', 'record_list.append((g_key,', 'self._data[_key],', "s...
962,440
huawei-noah/xingtian
logger.py
StatsRecorder.process_stats
process_stats
Process a stats received.
[ "Process", "a", "stats", "received." ]
def process_stats(self, stats): if stats.get('ctr_info'): if stats.get('ctr_info').get('cmd') == 'stats_msg{}'.format(self.name): self.record_explore_status(stats['data']) elif stats.get('is_bm'): self.local_data_writer.insert_records(stats['data']) bm_data2board = list() ...
['def', 'process_stats(self,', 'stats):', 'if', "stats.get('ctr_info'):", 'if', "stats.get('ctr_info').get('cmd')", '==', "'stats_msg{}'.format(self.name):", "self.record_explore_status(stats['data'])", 'elif', "stats.get('is_bm'):", "self.local_data_writer.insert_records(stats['data'])", 'bm_data2board', '=', 'list()'...
962,442
huawei-noah/xingtian
printer.py
print_immediately
print_immediately
Print some string immediately.
[ "Print", "some", "string", "immediately." ]
def print_immediately(to_str): print(to_str) sys.stdout.flush()
['def', 'print_immediately(to_str):', 'print(to_str)', 'sys.stdout.flush()']
962,443
huawei-noah/xingtian
printer.py
debug_within_interval
debug_within_interval
Print with time interval.
[ "Print", "with", "time", "interval." ]
def debug_within_interval(logs=None, interval=10, func=None, human_able=False, **kwargs): global LAST_PRINT if time() - LAST_PRINT > interval: if func and callable(func): func(**kwargs) if logs: logs_human = pprint.pformat(logs, indent=0, width=1) if human_able else logs ...
['def', 'debug_within_interval(logs=None,', 'interval=10,', 'func=None,', 'human_able=False,', '**kwargs):', 'global', 'LAST_PRINT', 'if', 'time()', '-', 'LAST_PRINT', '>', 'interval:', 'if', 'func', 'and', 'callable(func):', 'func(**kwargs)', 'if', 'logs:', 'logs_human', '=', 'pprint.pformat(logs,', 'indent=0,', 'widt...
962,444
huawei-noah/xingtian
profiler.py
do_profile
do_profile
Create dummy for import error.
[ "Create", "dummy", "for", "import", "error." ]
def do_profile(follow=[], profiler=None): def inner(func): def nothing(*args, **kwargs): return func(*args, **kwargs) return nothing return inner
['def', 'do_profile(follow=[],', 'profiler=None):', 'def', 'inner(func):', 'def', 'nothing(*args,', '**kwargs):', 'return', 'func(*args,', '**kwargs)', 'return', 'nothing', 'return', 'inner']
962,446
huawei-noah/xingtian
profiler.py
save_and_dump_stats
save_and_dump_stats
Create utils for save stats into file.
[ "Create", "utils", "for", "save", "stats", "into", "file." ]
def save_and_dump_stats(profiler, stats_file='default_stats.pkl'): if not profiler: print('invalid profiler handler!') return if os.path.exists(stats_file): print('remove {}, and re-write it.'.format(stats_file)) os.remove(stats_file) else: print('write into file: {}'...
['def', 'save_and_dump_stats(profiler,', "stats_file='default_stats.pkl'):", 'if', 'not', 'profiler:', "print('invalid", 'profiler', "handler!')", 'return', 'if', 'os.path.exists(stats_file):', "print('remove", '{},', 'and', 're-write', "it.'.format(stats_file))", 'os.remove(stats_file)', 'else:', "print('write", 'into...
962,447
huawei-noah/xingtian
profiler.py
show_stats_file
show_stats_file
Create utils for display stats.
[ "Create", "utils", "for", "display", "stats." ]
def show_stats_file(stats_file): if not show_text: print("Please use 'pip install line_profiler`, return with nothing do!") return def load_stats(filename): with open(filename, 'rb') as stats_handle: return pickle.load(stats_handle) print(load_stats(stats_file)) tmp_...
['def', 'show_stats_file(stats_file):', 'if', 'not', 'show_text:', 'print("Please', 'use', "'pip", 'install', 'line_profiler`,', 'return', 'with', 'nothing', 'do!")', 'return', 'def', 'load_stats(filename):', 'with', 'open(filename,', "'rb')", 'as', 'stats_handle:', 'return', 'pickle.load(stats_handle)', 'print(load_st...
962,448
huawei-noah/xingtian
profile_stats.py
SingleTracker.average
average
Mean time of `with` interaction.
[ "Mean", "time", "of", "`with`", "interaction." ]
def average(self): if not self.with_time_list: return np.nan return np.nanmean(self.with_time_list) * 1000
['def', 'average(self):', 'if', 'not', 'self.with_time_list:', 'return', 'np.nan', 'return', 'np.nanmean(self.with_time_list)', '*', '1000']
962,450
huawei-noah/xingtian
profile_stats.py
PredictStats.get
get
Get agent status and clear the buffer.
[ "Get", "agent", "status", "and", "clear", "the", "buffer." ]
def get(self): ret = {'mean_predictor_wait_ms': self.obs_wait_time * 1000 / self.iters, 'mean_predictor_infer_ms': self.inference_time * 1000 / self.iters} self.reset() return ret
['def', 'get(self):', 'ret', '=', "{'mean_predictor_wait_ms':", 'self.obs_wait_time', '*', '1000', '/', 'self.iters,', "'mean_predictor_infer_ms':", 'self.inference_time', '*', '1000', '/', 'self.iters}', 'self.reset()', 'return', 'ret']
962,451
huawei-noah/xingtian
profile_stats.py
AgentGroupStats.update_with_agent_stats
update_with_agent_stats
Update agent status to agent group.
[ "Update", "agent", "status", "to", "agent", "group." ]
def update_with_agent_stats(self, agent_stats: list): _steps = [sta['mean_env_step_time_ms'] for sta in agent_stats] _infers = [sta['mean_inference_time_ms'] for sta in agent_stats] _iters = [sta['iters'] for sta in agent_stats] self._stats.update({'mean_env_step_ms': np.nanmean(_steps), 'mean_inference...
['def', 'update_with_agent_stats(self,', 'agent_stats:', 'list):', '_steps', '=', "[sta['mean_env_step_time_ms']", 'for', 'sta', 'in', 'agent_stats]', '_infers', '=', "[sta['mean_inference_time_ms']", 'for', 'sta', 'in', 'agent_stats]', '_iters', '=', "[sta['iters']", 'for', 'sta', 'in', 'agent_stats]', "self._stats.up...
962,453
huawei-noah/xingtian
profile_stats.py
AgentGroupStats.get
get
Get the newest one-explore-status of agent group.
[ "Get", "the", "newest", "one-explore-status", "of", "agent", "group." ]
def get(self): self._stats.update({'explore_ms': self.explore_time_in_epi * 1000, 'wait_model_ms': self.wait_model_time * 1000, 'restore_model_ms': self.restore_model_time * 1000}) if self.iters > 0: self._stats.update({'mean_env_step_ms': self.env_step_time * 1000 / self.iters, 'mean_inference_ms': sel...
['def', 'get(self):', "self._stats.update({'explore_ms':", 'self.explore_time_in_epi', '*', '1000,', "'wait_model_ms':", 'self.wait_model_time', '*', '1000,', "'restore_model_ms':", 'self.restore_model_time', '*', '1000})', 'if', 'self.iters', '>', '0:', "self._stats.update({'mean_env_step_ms':", 'self.env_step_time', ...
962,454
huawei-noah/xingtian
profile_stats.py
TimerRecorder.get_metric
get_metric
Fetch the newest time record.
[ "Fetch", "the", "newest", "time", "record." ]
def get_metric(self, fields): ret = dict() for _task in fields: if not self.track_stub[_task]: continue ret.update({'{}_{}_mean_ms'.format(self.style, _task): 1000 * np.nanmean(self.track_stub[_task]), '{}_{}_max_ms'.format(self.style, _task): 1000 * np.max(self.track_stub[_task]), '...
['def', 'get_metric(self,', 'fields):', 'ret', '=', 'dict()', 'for', '_task', 'in', 'fields:', 'if', 'not', 'self.track_stub[_task]:', 'continue', "ret.update({'{}_{}_mean_ms'.format(self.style,", '_task):', '1000', '*', 'np.nanmean(self.track_stub[_task]),', "'{}_{}_max_ms'.format(self.style,", '_task):', '1000', '*',...
962,455
huawei-noah/xingtian
profile_stats.py
TimerRecorder.report_if_need
report_if_need
Rreport the time metric if need.
[ "Rreport", "the", "time", "metric", "if", "need." ]
def report_if_need(self, field_sets=None, **kwargs): if time() - self.last_report_time >= self.report_interval: to_log = self.get_metric(field_sets or self.fields) if kwargs: to_log.update(kwargs) to_log_format = pprint.pformat(to_log, indent=0, width=1) logging.debug('\n...
['def', 'report_if_need(self,', 'field_sets=None,', '**kwargs):', 'if', 'time()', '-', 'self.last_report_time', '>=', 'self.report_interval:', 'to_log', '=', 'self.get_metric(field_sets', 'or', 'self.fields)', 'if', 'kwargs:', 'to_log.update(kwargs)', 'to_log_format', '=', 'pprint.pformat(to_log,', 'indent=0,', 'width=...
962,456
huawei-noah/xingtian
coco.py
collate_fn
collate_fn
Collate fn for data loader.
[ "Collate", "fn", "for", "data", "loader." ]
def collate_fn(batch): return tuple(zip(*batch))
['def', 'collate_fn(batch):', 'return', 'tuple(zip(*batch))']
962,469
huawei-noah/xingtian
div2k.py
DIV2K.dataset_init
dataset_init
Costruct method, which will load some dateset information.
[ "Costruct", "method,", "which", "will", "load", "some", "dateset", "information." ]
def dataset_init(self): self.args.root_HR = FileOps.download_dataset(self.args.root_HR) self.args.root_LR = FileOps.download_dataset(self.args.root_LR) if self.args.subfile is not None: with open(self.args.subfile) as f: file_names = sorted([line.rstrip('\n') for line in f]) ...
['def', 'dataset_init(self):', 'self.args.root_HR', '=', 'FileOps.download_dataset(self.args.root_HR)', 'self.args.root_LR', '=', 'FileOps.download_dataset(self.args.root_LR)', 'if', 'self.args.subfile', 'is', 'not', 'None:', 'with', 'open(self.args.subfile)', 'as', 'f:', 'file_names', '=', "sorted([line.rstrip('\\n')"...
962,470
huawei-noah/xingtian
auto_lane_pointlane_codec.py
PointLaneCodec.uniform_sample_lane_y_axis
uniform_sample_lane_y_axis
Ensure y from bottom of image.
[ "Ensure", "y", "from", "bottom", "of", "image." ]
def uniform_sample_lane_y_axis(self, x_pt_list, y_pt_list): if len(x_pt_list) < 2 or len(y_pt_list) < 2: return (-1, -1, [], []) max_y = y_pt_list[-1] if max_y < self.input_height - 1: y1 = y_pt_list[-2] y2 = y_pt_list[-1] x1 = x_pt_list[-2] x2 = x_pt_list[-1] ...
['def', 'uniform_sample_lane_y_axis(self,', 'x_pt_list,', 'y_pt_list):', 'if', 'len(x_pt_list)', '<', '2', 'or', 'len(y_pt_list)', '<', '2:', 'return', '(-1,', '-1,', '[],', '[])', 'max_y', '=', 'y_pt_list[-1]', 'if', 'max_y', '<', 'self.input_height', '-', '1:', 'y1', '=', 'y_pt_list[-2]', 'y2', '=', 'y_pt_list[-1]', ...
962,501
huawei-noah/xingtian
auto_lane_pointlane_codec.py
PointLaneCodec.get_one_line_pass_anchors
get_one_line_pass_anchors
Get one line pass all anchors.
[ "Get", "one", "line", "pass", "all", "anchors." ]
def get_one_line_pass_anchors(self, startpos, endpos, xlist, y_list, anchor_count): anchor_list = [] anchor_distance_result = [] Gt_loc_list = [] for i in range(0, endpos - startpos + 1): h = self.feature_height - 1 - int((startpos + i) * self.interval / self.step_h) w = int(xlist[i] / s...
['def', 'get_one_line_pass_anchors(self,', 'startpos,', 'endpos,', 'xlist,', 'y_list,', 'anchor_count):', 'anchor_list', '=', '[]', 'anchor_distance_result', '=', '[]', 'Gt_loc_list', '=', '[]', 'for', 'i', 'in', 'range(0,', 'endpos', '-', 'startpos', '+', '1):', 'h', '=', 'self.feature_height', '-', '1', '-', 'int((st...
962,502
huawei-noah/xingtian
avazu_util.py
BaseDataset.summary
summary
Summarize the data set.
[ "Summarize", "the", "data", "set." ]
def summary(self): logging.info(self.__class__.__name__, 'data set summary:') logging.info('train set: ', self.train_size) logging.info('\tpositive samples: ', self.pos_train_samples) logging.info('\tnegative samples: ', self.neg_train_samples) logging.info('\tpositive ratio: ', self.train_pos_ratio...
['def', 'summary(self):', 'logging.info(self.__class__.__name__,', "'data", 'set', "summary:')", "logging.info('train", 'set:', "',", 'self.train_size)', "logging.info('\\tpositive", 'samples:', "',", 'self.pos_train_samples)', "logging.info('\\tnegative", 'samples:', "',", 'self.neg_train_samples)', "logging.info('\\t...
962,520
huawei-noah/xingtian
dataset.py
Dataset.transforms
transforms
Transform function which can replace transforms.
[ "Transform", "function", "which", "can", "replace", "transforms." ]
def transforms(self): return self._transforms
['def', 'transforms(self):', 'return', 'self._transforms']
962,526
huawei-noah/xingtian
dataset.py
Dataset.transforms
transforms
Set function of transforms.
[ "Set", "function", "of", "transforms." ]
def transforms(self, value): self._transforms = value
['def', 'transforms(self,', 'value):', 'self._transforms', '=', 'value']
962,527
huawei-noah/xingtian
adapter.py
TorchAdapter.sampler
sampler
Set function of sampler.
[ "Set", "function", "of", "sampler." ]
def sampler(self, value): self._sampler = value
['def', 'sampler(self,', 'value):', 'self._sampler', '=', 'value']
962,588
huawei-noah/xingtian
adapter.py
TfAdapter.data_map_func
data_map_func
Apply data map function from raw data.
[ "Apply", "data", "map", "function", "from", "raw", "data." ]
def data_map_func(self, images_index, label_index): if not self.is_detection: (image, label) = tf.numpy_function(self._get_item, [images_index, label_index], [self.image_dtype_tf, self.label_dtype_tf]) if self.fixed_size: image.set_shape(self.image_shape) label.set_shape(self...
['def', 'data_map_func(self,', 'images_index,', 'label_index):', 'if', 'not', 'self.is_detection:', '(image,', 'label)', '=', 'tf.numpy_function(self._get_item,', '[images_index,', 'label_index],', '[self.image_dtype_tf,', 'self.label_dtype_tf])', 'if', 'self.fixed_size:', 'image.set_shape(self.image_shape)', 'label.se...
962,592
huawei-noah/xingtian
imagenet.py
Imagenet.input_fn
input_fn
Define input_fn used by Tensorflow Estimator.
[ "Define", "input_fn", "used", "by", "Tensorflow", "Estimator." ]
def input_fn(self): data_files = os.path.join(self.data_path, 'train/train-*' if self.mode == 'train' else 'val/val-*') dataset = tf.data.Dataset.list_files(data_files, shuffle=False) if self.world_size > 1: dataset = dataset.shard(self.world_size, self.rank) if self.mode == 'train': dat...
['def', 'input_fn(self):', 'data_files', '=', 'os.path.join(self.data_path,', "'train/train-*'", 'if', 'self.mode', '==', "'train'", 'else', "'val/val-*')", 'dataset', '=', 'tf.data.Dataset.list_files(data_files,', 'shuffle=False)', 'if', 'self.world_size', '>', '1:', 'dataset', '=', 'dataset.shard(self.world_size,', '...
962,595
huawei-noah/xingtian
__init__.py
register_transforms
register_transforms
Import and register transforms automatically.
[ "Import", "and", "register", "transforms", "automatically." ]
def register_transforms(backend): import zeus if zeus.is_gpu_device(): from .ImageTransform import ImageTransform from .Invert import Invert from .MaskTransform import MaskTransform from .Posterize import Posterize from .RandomCrop_pair import RandomCrop_pair from...
['def', 'register_transforms(backend):', 'import', 'zeus', 'if', 'zeus.is_gpu_device():', 'from', '.ImageTransform', 'import', 'ImageTransform', 'from', '.Invert', 'import', 'Invert', 'from', '.MaskTransform', 'import', 'MaskTransform', 'from', '.Posterize', 'import', 'Posterize', 'from', '.RandomCrop_pair', 'import', ...
962,600
huawei-noah/xingtian
device_evaluator.py
DeviceEvaluator.train_process
train_process
Validate process for the model validate worker.
[ "Validate", "process", "for", "the", "model", "validate", "worker." ]
def train_process(self): init_log(level=General.logger.level, log_file='device_evaluator_{}.log'.format(self.worker_id), log_path=self.local_log_path) logging.info('start davinci or mobile evaluate process') self.load_model() self.valid_loader = self._init_dataloader(mode='test') performance = self....
['def', 'train_process(self):', 'init_log(level=General.logger.level,', "log_file='device_evaluator_{}.log'.format(self.worker_id),", 'log_path=self.local_log_path)', "logging.info('start", 'davinci', 'or', 'mobile', 'evaluate', "process')", 'self.load_model()', 'self.valid_loader', '=', "self._init_dataloader(mode='te...
962,606
huawei-noah/xingtian
evaluator.py
Evaluator.size
size
Return the size of current evaluator list.
[ "Return", "the", "size", "of", "current", "evaluator", "list." ]
def size(self): return len(self.sub_worker_list)
['def', 'size(self):', 'return', 'len(self.sub_worker_list)']
962,607
huawei-noah/xingtian
flops_and_params.py
add_new_hooks
add_new_hooks
Add new register hooks to custom hooks.
[ "Add", "new", "register", "hooks", "to", "custom", "hooks." ]
def add_new_hooks(custom_hooks): import torch.nn as nn from thop.profile import register_hooks from thop.vision.basic_hooks import count_softmax from zeus.modules.operators import ops add_register_hooks = {nn.PReLU: register_hooks[nn.ReLU], nn.ELU: register_hooks[nn.ReLU], nn.Softmax: count_softmax,...
['def', 'add_new_hooks(custom_hooks):', 'import', 'torch.nn', 'as', 'nn', 'from', 'thop.profile', 'import', 'register_hooks', 'from', 'thop.vision.basic_hooks', 'import', 'count_softmax', 'from', 'zeus.modules.operators', 'import', 'ops', 'add_register_hooks', '=', '{nn.PReLU:', 'register_hooks[nn.ReLU],', 'nn.ELU:', '...
962,618
huawei-noah/xingtian
__init__.py
register_metrics
register_metrics
Import and register metrics automatically.
[ "Import", "and", "register", "metrics", "automatically." ]
def register_metrics(backend): if backend == 'pytorch': from . import pytorch elif backend == 'tensorflow': from . import tensorflow elif backend == 'mindspore': from . import mindspore
['def', 'register_metrics(backend):', 'if', 'backend', '==', "'pytorch':", 'from', '.', 'import', 'pytorch', 'elif', 'backend', '==', "'tensorflow':", 'from', '.', 'import', 'tensorflow', 'elif', 'backend', '==', "'mindspore':", 'from', '.', 'import', 'mindspore']
962,624
huawei-noah/xingtian
metrics.py
Metrics.reset
reset
Reset states for new evaluation after each epoch.
[ "Reset", "states", "for", "new", "evaluation", "after", "each", "epoch." ]
def reset(self): self.metric_results = dict()
['def', 'reset(self):', 'self.metric_results', '=', 'dict()']
962,628
huawei-noah/xingtian
auc_metrics.py
AUC.summary
summary
Summary all cached records, here is the last pfm record.
[ "Summary", "all", "cached", "records,", "here", "is", "the", "last", "pfm", "record." ]
def summary(self): return self.pfm
['def', 'summary(self):', 'return', 'self.pfm']
962,636
huawei-noah/xingtian
lane_metric.py
LaneMetricCore.summary
summary
Summary all record from result cache, and get performance.
[ "Summary", "all", "record", "from", "result", "cache,", "and", "get", "performance." ]
def summary(self): hit_num = sum((result['hit_num'] for result in self.result_record)) pr_num = sum((result['pr_num'] for result in self.result_record)) gt_num = sum((result['gt_num'] for result in self.result_record)) precision = hit_num / (pr_num + sys.float_info.epsilon) recall = hit_num / (gt_nu...
['def', 'summary(self):', 'hit_num', '=', "sum((result['hit_num']", 'for', 'result', 'in', 'self.result_record))', 'pr_num', '=', "sum((result['pr_num']", 'for', 'result', 'in', 'self.result_record))', 'gt_num', '=', "sum((result['gt_num']", 'for', 'result', 'in', 'self.result_record))', 'precision', '=', 'hit_num', '/...
962,653
huawei-noah/xingtian
metrics.py
MetricBase.summary
summary
Summary all cached records, called after valid.
[ "Summary", "all", "cached", "records,", "called", "after", "valid." ]
def summary(self): raise NotImplementedError
['def', 'summary(self):', 'raise', 'NotImplementedError']
962,657
huawei-noah/xingtian
compressed_model_filter.py
CompressedModelFilter.select_satisfied_model
select_satisfied_model
Select satisfied models by standard.
[ "Select", "satisfied", "models", "by", "standard." ]
def select_satisfied_model(self, standard, num): (target, restrict) = self._parse_standard(standard) candidates = self._filtrate(restrict) satisfied_models = self._choose_models(candidates, target, num) return satisfied_models
['def', 'select_satisfied_model(self,', 'standard,', 'num):', '(target,', 'restrict)', '=', 'self._parse_standard(standard)', 'candidates', '=', 'self._filtrate(restrict)', 'satisfied_models', '=', 'self._choose_models(candidates,', 'target,', 'num)', 'return', 'satisfied_models']
962,691