Fix training hang due to constrained layout of matplotlib

This commit is contained in:
Yin Li 2020-07-27 12:39:41 -07:00
parent bf9b0ba426
commit 8ce13e67f6
2 changed files with 36 additions and 39 deletions

View File

@ -24,61 +24,58 @@ def plt_slices(*fields, size=64, title=None, cmap=None, norm=None):
if title is not None:
assert len(title) == nf
cmap = np.broadcast_to(cmap, (nf,))
norm = np.broadcast_to(norm, (nf,))
im_size = 2
cbar_height = 0.3
cbar_frac = cbar_height / (nc * im_size + cbar_height)
cbar_height = 0.2
fig, axes = plt.subplots(
nc, nf,
nc + 1, nf,
squeeze=False,
figsize=(nf * im_size, nc * im_size + cbar_height),
dpi=100,
constrained_layout=True,
gridspec_kw={'height_ratios': nc * [im_size] + [cbar_height]}
)
def quantize(x):
return 2 ** round(log2(x), ndigits=1)
for f, field in enumerate(fields):
for f, (field, cmap_col, norm_col) in enumerate(zip(fields, cmap, norm)):
all_non_neg = (field >= 0).all()
all_non_pos = (field <= 0).all()
if cmap is None:
if cmap_col is None:
if all_non_neg:
cmap_ = 'viridis'
cmap_col = 'viridis'
elif all_non_pos:
warnings.warn('no implementation for all non-positive values')
cmap_ = None
cmap_col = None
else:
cmap_ = 'RdBu_r'
else:
cmap_ = cmap
cmap_col = 'RdBu_r'
if norm is None:
if norm_col is None:
l2, l1, h1, h2 = np.percentile(field, [2.5, 16, 84, 97.5])
w1, w2 = (h1 - l1) / 2, (h2 - l2) / 2
if all_non_neg:
if h1 > 0.1 * h2:
norm_ = Normalize(vmin=0, vmax=quantize(2 * h2))
norm_col = Normalize(vmin=0, vmax=quantize(2 * h2))
else:
norm_ = LogNorm(vmin=quantize(0.5 * l2), vmax=quantize(2 * h2))
norm_col = LogNorm(vmin=quantize(0.5 * l2), vmax=quantize(2 * h2))
elif all_non_pos:
warnings.warn('no implementation for all non-positive values')
norm_ = None
norm_col = None
else:
if w1 > 0.1 * w2:
vlim = quantize(2.5 * w1)
norm_ = Normalize(vmin=-vlim, vmax=vlim)
norm_col = Normalize(vmin=-vlim, vmax=vlim)
else:
vlim = quantize(w2)
norm_ = SymLogNorm(linthresh=0.1 * w1, vmin=-vlim, vmax=vlim)
else:
norm_ = norm
norm_col = SymLogNorm(linthresh=0.1 * w1, vmin=-vlim, vmax=vlim)
for c in range(field.shape[0]):
s = (c,) + (0,) * (nd - 2) + (slice(64),) * 2
axes[c, f].pcolormesh(field[s], cmap=cmap_, norm=norm_)
axes[c, f].pcolormesh(field[s], cmap=cmap_col, norm=norm_col)
axes[c, f].set_aspect('equal')
@ -92,15 +89,11 @@ def plt_slices(*fields, size=64, title=None, cmap=None, norm=None):
axes[c, f].axis('off')
fig.colorbar(
ScalarMappable(norm=norm_, cmap=cmap_),
ax=axes[:, f],
ScalarMappable(norm=norm_col, cmap=cmap_col),
cax=axes[-1, f],
orientation='horizontal',
fraction=cbar_frac,
pad=0,
shrink=0.9,
aspect=10,
)
fig.set_constrained_layout_pads(w_pad=2/72, h_pad=2/72, wspace=0, hspace=0)
fig.tight_layout()
return fig

View File

@ -330,12 +330,14 @@ def train(epoch, loader, model, lag2eul, criterion,
logger.add_scalar('loss/epoch/train/lxe', epoch_loss.prod(),
global_step=epoch+1)
logger.add_figure('fig/epoch/train', plt_slices(
input[-1], lag_out[-1], lag_tgt[-1], lag_out[-1] - lag_tgt[-1],
eul_out[-1], eul_tgt[-1], eul_out[-1] - eul_tgt[-1],
title=['in', 'lag_out', 'lag_tgt', 'lag_out - lag_tgt',
'eul_out', 'eul_tgt', 'eul_out - eul_tgt'],
), global_step=epoch+1)
fig = plt_slices(
input[-1], lag_out[-1], lag_tgt[-1], lag_out[-1] - lag_tgt[-1],
eul_out[-1], eul_tgt[-1], eul_out[-1] - eul_tgt[-1],
title=['in', 'lag_out', 'lag_tgt', 'lag_out - lag_tgt',
'eul_out', 'eul_tgt', 'eul_out - eul_tgt'],
)
logger.add_figure('fig/epoch/train', fig, global_step=epoch+1)
fig.clf()
return epoch_loss
@ -380,12 +382,14 @@ def validate(epoch, loader, model, lag2eul, criterion, logger, device, args):
logger.add_scalar('loss/epoch/val/lxe', epoch_loss.prod(),
global_step=epoch+1)
logger.add_figure('fig/epoch/val', plt_slices(
input[-1], lag_out[-1], lag_tgt[-1], lag_out[-1] - lag_tgt[-1],
eul_out[-1], eul_tgt[-1], eul_out[-1] - eul_tgt[-1],
title=['in', 'lag_out', 'lag_tgt', 'lag_out - lag_tgt',
'eul_out', 'eul_tgt', 'eul_out - eul_tgt'],
), global_step=epoch+1)
fig = plt_slices(
input[-1], lag_out[-1], lag_tgt[-1], lag_out[-1] - lag_tgt[-1],
eul_out[-1], eul_tgt[-1], eul_out[-1] - eul_tgt[-1],
title=['in', 'lag_out', 'lag_tgt', 'lag_out - lag_tgt',
'eul_out', 'eul_tgt', 'eul_out - eul_tgt'],
)
logger.add_figure('fig/epoch/val', fig, global_step=epoch+1)
fig.clf()
return epoch_loss