def load_model_decode(model_dir, data, name, gpu, seg=True):
data.HP_gpu = gpu
print "Load Model from file: ", model_dir
model = SeqModel(data)
## load model need consider if the model trained in GPU and load in CPU, or vice versa
# if not gpu:
# model.load_state_dict(torch.load(model_dir), map_location=lambda storage, loc: storage)
# # model = torch.load(model_dir, map_location=lambda storage, loc: storage)
# else:
model.load_state_dict(torch.load(model_dir))
# model = torch.load(model_dir)
print("Decode %s data ..." % (name))
start_time = time.time()
speed, acc, p, r, f, pred_results = evaluate(data, model, name)
end_time = time.time()
time_cost = end_time - start_time
if seg:
print(
"%s: time:%.2fs, speed:%.2fst/s; acc: %.4f, p: %.4f, r: %.4f, f: %.4f"
% (name, time_cost, speed, acc, p, r, f))
else:
print("%s: time:%.2fs, speed:%.2fst/s; acc: %.4f" %
(name, time_cost, speed, acc))
return pred_results
def load_model(model_dir, data, gpu):
data.HP_gpu = gpu
print
"Load Model from file: ", model_dir
model = SeqModel(data)
## load model need consider if the model trained in GPU and load in CPU, or vice versa
if not gpu:
model.load_state_dict(
torch.load(model_dir, map_location=lambda storage, loc: storage))
# model = torch.load(model_dir, map_location=lambda storage, loc: storage)
else:
model.load_state_dict(torch.load(model_dir))
# model = torch.load(model_dir)
return model
class Tagger():
def __init__(self, model_dir, dset_dir, gpu, seg):
self.model_dir = model_dir
self.dset_dir = dset_dir
self.data = load_data_setting(dset_dir)
self.data.HP_gpu = gpu
self.model = SeqModel(self.data)
self.model.load_state_dict(torch.load(self.model_dir))
def change_inlines(self, text):
self.data.inference_single_with_gaz(text)
def load_model_inference(self, seg=True):
#self.model = SeqModel(self.data)
return inference(self.data, self.model)
def load_model_decode(model_dir, data, name, gpu, seg=True):
data.HP_gpu = gpu
print ("Load Model from file: ", model_dir)
model = SeqModel(data)
## load model need consider if the model trained in GPU and load in CPU, or vice versa
# if not gpu:
# model.load_state_dict(torch.load(model_dir), map_location=lambda storage, loc: storage)
# # model = torch.load(model_dir, map_location=lambda storage, loc: storage)
# else:
model.load_state_dict(torch.load(model_dir))
# model = torch.load(model_dir)
print("Decode %s data ..."%(name))
start_time = time.time()
speed, acc, p, r, f, pred_results = evaluate(data, model, name)
end_time = time.time()
time_cost = end_time - start_time
if seg:
print("%s: time:%.2fs, speed:%.2fst/s; acc: %.4f, p: %.4f, r: %.4f, f: %.4f"%(name, time_cost, speed, acc, p, r, f))
else:
print("%s: time:%.2fs, speed:%.2fst/s; acc: %.4f"%(name, time_cost, speed, acc))
return pred_results
def load_model_inference(model_dir, data, name, gpu, seg=True):
data.HP_gpu = gpu
model = SeqModel(data)
model.load_state_dict(torch.load(model_dir))
return inference(data, model)
# -*- coding: utf-8 -*-