def repl(prompt="spill> "):
"The spill read-eval-print loop"
try:
while True:
inp = input(prompt)
if inp:
evaluate(parse(tokenize(inp)))
except EOFError:
print()
def main(args):
env_info = get_sys_env()
place = 'gpu' if env_info['Paddle compiled with cuda'] and env_info[
'GPUs used'] else 'cpu'
paddle.set_device(place)
if not args.cfg:
raise RuntimeError('No configuration file specified.')
cfg = Config(args.cfg)
# Only support for the DeepLabv3+ model
if args.data_format == 'NHWC':
if cfg.dic['model']['type'] != 'DeepLabV3P':
raise ValueError(
'The "NHWC" data format only support the DeepLabV3P model!')
cfg.dic['model']['data_format'] = args.data_format
cfg.dic['model']['backbone']['data_format'] = args.data_format
loss_len = len(cfg.dic['loss']['types'])
for i in range(loss_len):
cfg.dic['loss']['types'][i]['data_format'] = args.data_format
val_dataset = cfg.val_dataset
if val_dataset is None:
raise RuntimeError(
'The verification dataset is not specified in the configuration file.'
)
elif len(val_dataset) == 0:
raise ValueError(
'The length of val_dataset is 0. Please check if your dataset is valid'
)
msg = '\n---------------Config Information---------------\n'
msg += str(cfg)
msg += '------------------------------------------------'
logger.info(msg)
model = cfg.model
if args.model_path:
utils.load_entire_model(model, args.model_path)
logger.info('Loaded trained params of model successfully')
test_config = get_test_config(cfg, args)
config_check(cfg, val_dataset=val_dataset)
evaluate(model,
val_dataset,
num_workers=args.num_workers,
is_view=args.is_view,
save_dir=args.save_dir,
**test_config)
def main(args):
env_info = get_sys_env()
place = 'gpu' if env_info['Paddle compiled with cuda'] and env_info[
'GPUs used'] else 'cpu'
paddle.set_device(place)
if not args.cfg:
raise RuntimeError('No configuration file specified.')
cfg = Config(args.cfg)
val_dataset = cfg.val_dataset
if val_dataset is None:
raise RuntimeError(
'The verification dataset is not specified in the configuration file.'
)
elif len(val_dataset) == 0:
raise ValueError(
'The length of val_dataset is 0. Please check if your dataset is valid'
)
msg = '\n---------------Config Information---------------\n'
msg += str(cfg)
msg += '------------------------------------------------'
logger.info(msg)
model = cfg.model
if args.model_path:
paddleseg.utils.utils.load_entire_model(model, args.model_path)
logger.info('Loaded trained params of model successfully')
config_check(cfg, val_dataset=val_dataset)
evaluate(
model,
val_dataset,
threshold=args.threshold,
nms_kernel=args.nms_kernel,
top_k=args.top_k,
num_workers=args.num_workers,
)
def eval(s):
tokens = tokenization.tokenize(s)
if DEBUG:
print("tokens:", tokens)
tokens, tree = parse.parse(tokens)
if DEBUG:
print("tree:", tree)
tree = parse.quote(tree)
if DEBUG:
print("post quote:", tree)
return [core.evaluate({}, x) for x in tree]
def interpret(filename):
with open(filename, "rt") as file:
inp = file.read()
if inp:
evaluate(parse(tokenize(inp)))
histfile = os.path.join(os.path.expanduser("~"), ".pyclojurehist")
try:
readline.read_history_file(histfile)
except IOError:
# Pass here as there isn't any history file, so one will be
# written by atexit
pass
import atexit
atexit.register(readline.write_history_file, histfile)
parse = lispparser()
lexer = lisplexer()
if __name__ == "__main__":
global_scope = Scope()
scopechain = [global_scope]
while True:
try:
txt = raw_input("pylisp> ")
if re.search('^\s*$', txt):
continue
else:
print(tostring(evaluate(parse(txt), scopechain)))
except EOFError:
break
except KeyboardInterrupt:
print # Give user a newline after Cntrl-C for readability
break
except Exception, e:
print e
def evalparse(x):
return evaluate(parse(x), scopechain)
print(F)
print(">>> F_1 model <<<")
print(F_1)
print(">>> F_2 model <<<")
print(F_2)
print(">>> F_t model <<<")
print(F_t)
# pre-train on source dataset
print("=== Pre-train networks ===")
if cfg.model_trained["pretrain"]:
print("pass")
else:
pre_train(F, F_1, F_2, F_t, source_data_loader)
print(">>> evaluate F+F_1")
evaluate(F, F_1, source_data_loader_test)
print(">>> evaluate F+F_2")
evaluate(F, F_2, source_data_loader_test)
print(">>> evaluate F+F_t")
evaluate(F, F_t, source_data_loader_test)
print("=== Adapt F_t ===")
if cfg.model_trained["domain_adapt"]:
print("pass")
else:
# generate pseudo labels on target dataset
print("--- Generate Pseudo Label ---")
excerpt, pseudo_labels = \
genarate_labels(F, F_1, F_2, target_dataset, cfg.num_target_init)
print(">>> Genrate pseudo labels {}".format(len(pseudo_labels)))
# In[5]:
encoder, decoder = train(series[:-20], series[-20:],
n_steps=200, attn_model="dot",
hidden_size=16, n_layers=1, dropout=0, batch_size=128,
elr=0.001, dlr=0.005, clip=50.0, print_every=20,
teacher_forcing_ratio=lambda x: 1 if x < 20 else 0.5)
# In[6]:
TARGET_IDX = series.shape[1] - 9
preds, attentions = evaluate(series[:-20, TARGET_IDX:(TARGET_IDX+1)], 20, encoder, decoder)
print(np.mean(np.square((preds.numpy() - series[-20:, TARGET_IDX:(TARGET_IDX+1)]))))
plt.plot(np.arange(40), series[:-20, TARGET_IDX])
plt.plot(np.arange(40, 60), preds[:, 0].numpy(), "g-")
plt.plot(np.arange(40, 60), series[-20:, TARGET_IDX], "ro")
plt.title("Pure Sine Wave Prediction")
# In[7]:
show_attention(attentions.numpy()[:, 0, :])
# ## Noisy Data
def evaluate_wrapper(*args, **kargs):
return round(core.evaluate(*args, **kargs), 17)
