Skip to content

Separate code for building sweep df and plot stability #36

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
fabian-sp merged 1 commit into from
Jun 10, 2025
Merged

Separate code for building sweep df and plot stability #36

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions .gitignore
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -1,3 +1,5 @@
*.DS_Store

# Data files
data/
# Checkpoints
Expand Down
29 changes: 5 additions & 24 deletions stepback/plotting.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -44,34 +44,15 @@ def plot_stability(R,
if isinstance(score, str):
score = [score]

base_df = R.base_df.copy()
id_df = R.id_df.copy()

grouped = base_df.groupby(['name', xaxis])
max_epoch = grouped['epoch'].max()
assert len(max_epoch.unique()) == 1, "It seems that different setups ran for different number of epochs."

if cutoff is None:
cutoff_epoch = (max_epoch[0], max_epoch[0])
else:
cutoff_epoch = (cutoff, max_epoch[0])

fig, axs = plt.subplots(len(score),1,figsize=figsize)

for j, s in enumerate(score):
# filter epochs
sub_df = base_df[(base_df.epoch >= cutoff_epoch[0]) & (base_df.epoch <= cutoff_epoch[1])]
# select the columns to group by
grouping_cols = [c for c in id_df.columns if c not in ignore_columns]
# group by all id_cols
df = sub_df.groupby(grouping_cols)[[s, s+'_std']].mean() # use dropna=False if we would have nan values
# move xaxis out of grouping
df = df.reset_index(level=xaxis)
# make xaxis float
df[xaxis] = df[xaxis].astype('float')

# get method and learning rate with best score
# best_ind, best_x = df.index[df[s].argmax()], df[xaxis][df[s].argmax()]
df = R.build_sweep_df(score=s,
xaxis=xaxis,
ignore_columns=ignore_columns,
cutoff=cutoff
)

ax = axs.ravel()[j] if len(score) > 1 else axs
# .unique(level=) might be useful at some point
Expand Down
32 changes: 32 additions & 0 deletions stepback/record.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -189,6 +189,38 @@ def _build_base_df(self, agg='mean'):

return df

def build_sweep_df(self, score='val_score', xaxis='lr', ignore_columns=list(), cutoff=None):

base_df = self.base_df.copy()
id_df = self.id_df.copy()

grouped = base_df.groupby(['name', xaxis])
max_epoch = grouped['epoch'].max()
assert len(max_epoch.unique()) == 1, "It seems that different setups ran for different number of epochs."

if cutoff is None:
cutoff_epoch = (max_epoch[0], max_epoch[0])
else:
cutoff_epoch = (cutoff, max_epoch[0])

# filter epochs
sub_df = base_df[(base_df.epoch >= cutoff_epoch[0])
&
(base_df.epoch <= cutoff_epoch[1])]
# select the columns to group by
grouping_cols = [c for c in id_df.columns if c not in ignore_columns]
# group by all id_cols
df = sub_df.groupby(grouping_cols)[[score, score+'_std']].mean()
# move xaxis out of grouping
df = df.reset_index(level=xaxis)
# make xaxis float
df[xaxis] = df[xaxis].astype('float')

# get method and learning rate with best score
# best_ind, best_x = df.index[df[s].argmax()], df[xaxis][df[s].argmax()]

return df

#============ DATABASE =================================
#=======================================================

Expand Down