def print_table(table, header_text, row_labels, col_labels, colwidth=10,
latex=True):
"""Pretty-print a 2D array of data, optionally with row/col labels"""
print('')
if latex:
num_cols = len(table[0])
print("\\begin{center}")
print("\\begin{tabular}{l" + "c" * num_cols + "}")
print("\\toprule")
print('\multicolumn{8}{c}{'+header_text+'} \\')
print("\\midrule")
else:
print('--------', header_text)
for row, label in zip(table, row_labels):
row.insert(0, label)
if latex:
col_labels = ["\\textbf{" + str(col_label) + "}"
for col_label in col_labels]
table.insert(0, col_labels)
for r, row in enumerate(table):
misc.print_row(row, colwidth=colwidth, latex=latex)
if latex and r == 0:
print("\\midrule")
if latex:
print("\\bottomrule")
print("\\end{tabular}")
print("\\end{center}")
def print_table(table, header_text, row_labels, col_labels, colwidth=10,
latex=True):
"""Pretty-print a 2D array of data, optionally with row/col labels"""
print("")
if latex:
num_cols = len(table[0])
print("\\begin{center}")
print("\\adjustbox{max width=\\textwidth}{%")
print("\\begin{tabular}{l" + "c" * num_cols + "}")
print("\\toprule")
else:
print("--------", header_text)
for row, label in zip(table, row_labels):
row.insert(0, label)
if latex:
col_labels = ["\\textbf{" + str(col_label).replace("%", "\\%") + "}"
for col_label in col_labels]
table.insert(0, col_labels)
for r, row in enumerate(table):
misc.print_row(row, colwidth=colwidth, latex=latex)
if latex and r == 0:
print("\\midrule")
if latex:
print("\\bottomrule")
print("\\end{tabular}}")
print("\\end{center}")
results = {
'step': step,
'epoch': step / steps_per_epoch,
}
for key, val in checkpoint_vals.items():
results[key] = np.mean(val)
evals = zip(eval_loader_names, eval_loaders, eval_weights)
for name, loader, weights in evals:
acc = misc.accuracy(algorithm, loader, weights, device)
results[name + '_acc'] = acc
results_keys = sorted(results.keys())
if results_keys != last_results_keys:
misc.print_row(results_keys, colwidth=12)
last_results_keys = results_keys
misc.print_row([results[key] for key in results_keys], colwidth=12)
results.update({'hparams': hparams, 'args': vars(args)})
epochs_path = os.path.join(args.output_dir, 'results.jsonl')
with open(epochs_path, 'a') as f:
f.write(json.dumps(results, sort_keys=True) + "\n")
algorithm_dict = algorithm.state_dict()
start_step = step + 1
checkpoint_vals = collections.defaultdict(lambda: [])
records = []
with open(epochs_path, 'r') as f:
if args.algorithm == 'MULDENS':
eval_dict = {}
for name, loader, weights in evals:
eval_dict[name] = [loader, weights]
models_selected = [
1, 1, 1
] #step_vals['models_selected'] # random stuff. we will not use it anyway
correct_models_selected_for_each_domain = np.nan * np.ones(
len(dataset))
train_envs = [
i for i in range(len(dataset)) if i not in args.test_envs
]
for t, m in zip(train_envs, models_selected):
correct_models_selected_for_each_domain[t] = m
acc_flags = {'ensemble_for_obs': True, 'compute_test_beta': False}
results_MULDENS = misc.MULDENS_accuracy(
algorithm, eval_dict, args.test_envs,
correct_models_selected_for_each_domain, device, acc_flags)
results_MULDENS['step'] = step_number
results_keys = sorted(
[k for k in results_MULDENS.keys() if 'acc' in k])
results_keys.append('step')
if i == 0:
misc.print_row(results_keys, colwidth=25)
misc.print_row([results_MULDENS[key] for key in results_keys],
colwidth=25)
