steffen/self-replicating-neural-networks
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

StackPlot

Browse Source
This commit is contained in:
Steffen Illium
2022-02-04 21:07:19 +01:00
parent 21c3e75177
commit 3746c7d26e
3 changed files with 80 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
README.md
View File

@@ -37,7 +37,7 @@
- Es entstehen mehr SRNN
- Der Dropout Effekt wird stärker (diff_ohne_SRNN = _0.0_)
![image info](./figures/lesser_neurons_dropout_stacked_barplot.png)
- [ ] Weiter Trainieren -> 500 Epochs?
- [X] Weiter Trainieren -> 500 Epochs?
- [ ] Loss Gewichtung anpassen
- [x] Training ohne Residual Skip Connection
- Ist kacke

70
experiments/meta_task_exp.py
View File

@@ -37,11 +37,11 @@ else:
pass
from network import MetaNet
from functionalities_test import test_for_fixpoints
from functionalities_test import test_for_fixpoints, FixTypes
WORKER = 10 if not debug else 2
BATCHSIZE = 500 if not debug else 50
EPOCH = 100 if not debug else 3
EPOCH = 400 if not debug else 3
VALIDATION_FRQ = 5 if not debug else 1
SELF_TRAIN_FRQ = 1 if not debug else 1
DEVICE = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
@@ -131,7 +131,30 @@ def checkpoint_and_validate(model, out_path, epoch_n, final_model=False):
return result
def plot_training_particle_types(path_to_dataframe):
plt.clf()
# load from Drive
df = pd.read_csv(path_to_dataframe, index_col=False)
# Set up figure
fig, ax = plt.subplots() # initializes figure and plots
data = df[df['Metric'].isin(FixTypes.all_types())]
fix_types = data['Metric'].unique()
data = data.pivot(index='Epoch', columns='Metric', values='Score').reset_index().fillna(0)
_ = plt.stackplot(data['Epoch'], *[data[fixtype] for fixtype in fix_types], labels=fix_types.tolist())
ax.set(ylabel='Particle Count', xlabel='Epoch')
ax.set_title('Particle Type Count')
fig.legend(loc="center right", title='Particle Type', bbox_to_anchor=(0.85, 0.5))
plt.tight_layout()
if debug:
plt.show()
else:
plt.savefig(Path(path_to_dataframe.parent / 'training_particle_type_lp.png'), dpi=300)
def plot_training_result(path_to_dataframe):
plt.clf()
# load from Drive
df = pd.read_csv(path_to_dataframe, index_col=False)
@@ -163,6 +186,7 @@ def plot_training_result(path_to_dataframe):
else:
plt.savefig(Path(path_to_dataframe.parent / 'training_lineplot.png'), dpi=300)
def flat_for_store(parameters):
return (x.item() for y in parameters for x in y.detach().flatten())
@@ -170,7 +194,7 @@ def flat_for_store(parameters):
if __name__ == '__main__':
self_train = True
training = True
training = False
plotting = True
particle_analysis = True
as_sparse_network_test = True
@@ -265,6 +289,13 @@ if __name__ == '__main__':
weight_store = new_storage_df('weights', meta_weight_count)
metanet.eval()
if particle_analysis:
counter_dict = defaultdict(lambda: 0)
# This returns ID-functions
_ = test_for_fixpoints(counter_dict, list(metanet.particles))
for key, value in dict(counter_dict).items():
step_log = dict(Epoch=int(EPOCH), Batch=BATCHSIZE, Metric=key, Score=value)
train_store.loc[train_store.shape[0]] = step_log
accuracy = checkpoint_and_validate(metanet, run_path, EPOCH, final_model=True)
validation_log = dict(Epoch=EPOCH, Batch=BATCHSIZE,
Metric='Test Accuracy', Score=accuracy.item())
@@ -278,33 +309,35 @@ if __name__ == '__main__':
if plotting:
plot_training_result(df_store_path)
if particle_analysis:
plot_training_particle_types(df_store_path)
exit()
if particle_analysis:
model_path = next(run_path.glob(f'*e100.tp'))
model_path = next(run_path.glob(f'*e{EPOCH}.tp'))
latest_model = torch.load(model_path, map_location=DEVICE).eval()
counter_dict = defaultdict(lambda: 0)
_ = test_for_fixpoints(counter_dict, list(latest_model.particles))
tqdm.write(str(dict(counter_dict)))
zero_ident = torch.load(model_path, map_location=DEVICE).eval().replace_with_zero('identity_func')
zero_other = torch.load(model_path, map_location=DEVICE).eval().replace_with_zero('other_func')
if as_sparse_network_test:
acc_pre = validate(model_path, ratio=0.01).item()
ident_ckpt = set_checkpoint(zero_ident, model_path.parent, -1, final_model=True)
ident_acc_post = validate(ident_ckpt, ratio=0.01).item()
tqdm.write(f'Zero_ident diff = {abs(ident_acc_post-acc_pre)}')
other_ckpt = set_checkpoint(zero_other, model_path.parent, -2, final_model=True)
other_acc_post = validate(other_ckpt, ratio=0.01).item()
tqdm.write(f'Zero_other diff = {abs(other_acc_post - acc_pre)}')
acc_pre = validate(model_path, ratio=1).item()
diff_table = pd.DataFrame(columns=['Particle Type', 'Accuracy', 'Diff'])
for fixpoint_type in FixTypes.all_types():
new_model = torch.load(model_path, map_location=DEVICE).eval().replace_with_zero(fixpoint_type)
new_ckpt = set_checkpoint(new_model, model_path.parent, fixpoint_type, final_model=True)
acc_post = validate(new_ckpt, ratio=1).item()
acc_diff = abs(acc_post-acc_pre)
tqdm.write(f'Zero_ident diff = {acc_diff}')
diff_table.iloc[diff_table.shape[0]] = (fixpoint_type, acc_post, acc_diff)
if plotting:
plt.clf()
fig, ax = plt.subplots(ncols=2)
data = [acc_pre, ident_acc_post, other_acc_post]
labels = ['Full Network', 'Sparse, No Identity', 'Sparse, No Other']
for idx, (score, name) in enumerate(zip(data, labels)):
l = sns.barplot(y=[score], x=['Networks'], color=sns.color_palette()[idx], label=name, ax=ax[0])
barplot = sns.barplot(data=diff_table, y='Accurady', x=['Particle Type'],
color=sns.color_palette()[:diff_table.shape[0]], ax=ax[0])
# noinspection PyUnboundLocalVariable
for idx, patch in enumerate(l.patches):
for idx, patch in enumerate(barplot.patches):
if idx != 0:
# we recenter the bar
patch.set_x(patch.get_x() + idx * 0.035)
@@ -313,7 +346,6 @@ if __name__ == '__main__':
ax[0].set_xlabel('Accuracy')
# ax[0].legend()
counter_dict['full_network'] = sum(counter_dict.values())
ax[1].pie(counter_dict.values(), labels=counter_dict.keys(), colors=sns.color_palette()[:3], )
ax[1].set_title('Particle Count for ')
# ax[1].set_xlabel('')

44
functionalities_test.py
View File

@@ -6,6 +6,19 @@ from tqdm import tqdm
from network import Net
class FixTypes:
divergent = 'divergent'
fix_zero = 'fix_zero'
identity_func = 'identity_func'
fix_sec = 'fix_sec'
other_func = 'other_func'
@classmethod
def all_types(cls):
return [val for key, val in cls.__dict__.items() if isinstance(val, str) and not key.startswith('_')]
def is_divergent(network: Net) -> bool:
return network.input_weight_matrix().isinf().any().item() or network.input_weight_matrix().isnan().any().item()
@@ -16,7 +29,6 @@ def is_identity_function(network: Net, epsilon=pow(10, -5)) -> bool:
target_data = network.create_target_weights(input_data)
predicted_values = network(input_data)
return torch.allclose(target_data.detach(), predicted_values.detach(),
rtol=0, atol=epsilon)
@@ -57,21 +69,21 @@ def test_for_fixpoints(fixpoint_counter: Dict, nets: List, id_functions=None):
for net in tqdm(nets, desc='Fixpoint Tester', total=len(nets)):
if is_divergent(net):
fixpoint_counter["divergent"] += 1
net.is_fixpoint = "divergent"
fixpoint_counter[FixTypes.divergent] += 1
net.is_fixpoint = FixTypes.divergent
elif is_identity_function(net): # is default value
fixpoint_counter["identity_func"] += 1
net.is_fixpoint = "identity_func"
fixpoint_counter[FixTypes.identity_func] += 1
net.is_fixpoint = FixTypes.identity_func
id_functions.append(net)
elif is_zero_fixpoint(net):
fixpoint_counter["fix_zero"] += 1
net.is_fixpoint = "fix_zero"
fixpoint_counter[FixTypes.fix_zero] += 1
net.is_fixpoint = FixTypes.fix_zero
elif is_secondary_fixpoint(net):
fixpoint_counter["fix_sec"] += 1
net.is_fixpoint = "fix_sec"
fixpoint_counter[FixTypes.fix_sec] += 1
net.is_fixpoint = FixTypes.fix_sec
else:
fixpoint_counter["other_func"] += 1
net.is_fixpoint = "other_func"
fixpoint_counter[FixTypes.other_func] += 1
net.is_fixpoint = FixTypes.other_func
return id_functions
@@ -82,14 +94,14 @@ def changing_rate(x_new, x_old):
def test_status(net: Net) -> Net:
if is_divergent(net):
net.is_fixpoint = "divergent"
net.is_fixpoint = FixTypes.divergent
elif is_identity_function(net): # is default value
net.is_fixpoint = "identity_func"
net.is_fixpoint = FixTypes.identity_func
elif is_zero_fixpoint(net):
net.is_fixpoint = "fix_zero"
net.is_fixpoint = FixTypes.fix_zero
elif is_secondary_fixpoint(net):
net.is_fixpoint = "fix_sec"
net.is_fixpoint = FixTypes.fix_sec
else:
net.is_fixpoint = "other_func"
net.is_fixpoint = FixTypes.other_func
return net