def _load_model(self, exp_dir):
config = load_config(os.path.join(exp_dir, 'config.json'))
self.agent = AutoregressiveRNN(config)
self.agent.load_checkpoint('checkpoint.pth.tar')
self.model = self.agent.model
self.model.eval()
self.config = config
def load_exp_data(all_exp_dir):
params = []
losses = []
for exp_name in os.listdir(all_exp_dir):
try:
exp_dir = os.path.join(all_exp_dir, exp_name)
config = load_config(os.path.join(exp_dir, 'config.json'))
vec = config_to_vec(config)
vec['exp_name'] = exp_name
summaries = load_json(
os.path.join(exp_dir, 'summaries', 'all_scalars.json'))
k_loss = get_key_for_metric(summaries.keys(),
'validation/loss/loss')
if k_loss is None:
print('Metric not foud... skipping')
continue
loss = np.average([x[2] for x in summaries[k_loss][-5:]])
params.append(vec)
losses.append(loss)
except FileNotFoundError:
print('File not found... skipping')
continue
return params, losses
def _load_model(self, exp_dir):
config = load_config(os.path.join(exp_dir, 'config.json'))
# config['cuda'] = False
config['gpu_device'] = 9
self.agent = FeedforwardNN(config)
self.agent.load_checkpoint('checkpoint.pth.tar')
self.model = self.agent.model
self.model.eval()
self.config = config
def __init__(self, exp_dir):
self.config = load_config(os.path.join(exp_dir, 'config.json'))
self._load_model(exp_dir)
# self.test_dataset = PyramidImages(500, input_size=self.config.encoder_kwargs.input_size, split='train')
self.test_dataset = PyramidImages(None, input_size=self.config.encoder_kwargs.input_size, split='test',
knowledge_states=self.config.knowledge_states)
self.group_lookup = self._group_lookup_array(GROUPING)
self.dataloader = DataLoader(self.test_dataset, batch_size=64, shuffle=True)
self.n_labels = 5 if self.config.knowledge_states else 13
def __init__(self, problem, exp_dir, strategy='map'):
self.config = load_config(os.path.join(exp_dir, 'config.json'))
self._load_model(exp_dir)
self.strategy = strategy
self.student_data = CitizenshipLabels(
13, split='test', vocab=self.agent.train_dataset.vocab)
self.dataloader = DataLoader(self.student_data,
batch_size=1,
shuffle=False)