Skip to content
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

Jump to bottom

Sped up MetricsCB and ProgressCB #18

Open
wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Sped up MetricsCB and ProgressCB #18

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
49 changes: 31 additions & 18 deletions miniai/learner.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -50,13 +50,16 @@ def to_cpu(x):
return x.detach().cpu()

# %% ../nbs/09_learner.ipynb 35
from torcheval.metrics import Metric, Mean

class MetricsCB(Callback):
def __init__(self, *ms, **metrics):
def __init__(self, *ms, device=def_device, **metrics):
for o in ms: metrics[type(o).__name__] = o
self.metrics = metrics
for m in self.metrics.values(): m.to(device)
self.all_metrics = copy(metrics)
self.all_metrics['loss'] = self.loss = Mean()

self.all_metrics['loss'] = self.loss = Mean(device='cpu' if 'mps' in device else device)
Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

MPS does not support doubles, and Mean fails when placed on device as it weights are double.

def _log(self, d): print(d)
def before_fit(self, learn): learn.metrics = self
def before_epoch(self, learn): [o.reset() for o in self.all_metrics.values()]
Expand All @@ -68,9 +71,9 @@ def after_epoch(self, learn):
self._log(log)

def after_batch(self, learn):
x,y,*_ = to_cpu(learn.batch)
for m in self.metrics.values(): m.update(to_cpu(learn.preds), y)
self.loss.update(to_cpu(learn.loss), weight=len(x))
x,y,*_ = learn.batch
for m in self.metrics.values(): m.update(learn.preds.to(m.device), y)
self.loss.update(learn.loss.to(self.loss.device), weight=len(x))

# %% ../nbs/09_learner.ipynb 36
class DeviceCB(Callback):
Expand All @@ -91,27 +94,37 @@ def zero_grad(self, learn): learn.opt.zero_grad()
# %% ../nbs/09_learner.ipynb 42
class ProgressCB(Callback):
order = MetricsCB.order+1
def __init__(self, plot=False): self.plot = plot
def __init__(self, plot=False, lag=10): fc.store_attr()
jph00 marked this conversation as resolved.
Show resolved Hide resolved
def before_fit(self, learn):
learn.epochs = self.mbar = master_bar(learn.epochs)
self.first = True
if hasattr(learn, 'metrics'): learn.metrics._log = self._log
self.losses = []

self.gpu_losses = []

def _log(self, d):
if self.first:
self.mbar.write(list(d), table=True)
self.first = False
self.mbar.write(list(d.values()), table=True)

def _plot(self, lag=0):
n = max(0,len(self.gpu_losses)-lag)
if n == 0: return
self.losses, self.gpu_losses = self.losses + [l.item() for l in self.gpu_losses[:n]], self.gpu_losses[n:]
self.mbar.update_graph([[fc.L.range(self.losses), self.losses]])

def before_epoch(self, learn): learn.dl = progress_bar(learn.dl, leave=False, parent=self.mbar)
def after_batch(self, learn):
learn.dl.comment = f'{learn.loss:.3f}'
if self.plot and hasattr(learn, 'metrics') and learn.training:
self.losses.append(learn.loss.item())
self.mbar.update_graph([[fc.L.range(self.losses), self.losses]])
self.gpu_losses.append(learn.loss.detach())
if len(self.gpu_losses) > 2* self.lag: self._plot(self.lag)

def after_epoch(self, learn):
learn.dl.comment = f'{learn.loss:.3f}'
if learn.training: self._plot()

# %% ../nbs/09_learner.ipynb 47
# %% ../nbs/09_learner.ipynb 48
class with_cbs:
def __init__(self, nm): self.nm = nm
def __call__(self, f):
Expand All @@ -124,7 +137,7 @@ def _f(o, *args, **kwargs):
finally: o.callback(f'cleanup_{self.nm}')
return _f

# %% ../nbs/09_learner.ipynb 48
# %% ../nbs/09_learner.ipynb 49
class Learner():
def __init__(self, model, dls=(0,), loss_func=F.mse_loss, lr=0.1, cbs=None, opt_func=optim.SGD):
cbs = fc.L(cbs)
Expand Down Expand Up @@ -180,15 +193,15 @@ def callback(self, method_nm): run_cbs(self.cbs, method_nm, self)
@property
def training(self): return self.model.training

# %% ../nbs/09_learner.ipynb 51
# %% ../nbs/09_learner.ipynb 52
class TrainLearner(Learner):
def predict(self): self.preds = self.model(self.batch[0])
def get_loss(self): self.loss = self.loss_func(self.preds, self.batch[1])
def backward(self): self.loss.backward()
def step(self): self.opt.step()
def zero_grad(self): self.opt.zero_grad()

# %% ../nbs/09_learner.ipynb 52
# %% ../nbs/09_learner.ipynb 53
class MomentumLearner(TrainLearner):
def __init__(self, model, dls, loss_func, lr=None, cbs=None, opt_func=optim.SGD, mom=0.85):
self.mom = mom
Expand All @@ -198,10 +211,10 @@ def zero_grad(self):
with torch.no_grad():
for p in self.model.parameters(): p.grad *= self.mom

# %% ../nbs/09_learner.ipynb 57
# %% ../nbs/09_learner.ipynb 58
from torch.optim.lr_scheduler import ExponentialLR

# %% ../nbs/09_learner.ipynb 59
# %% ../nbs/09_learner.ipynb 60
class LRFinderCB(Callback):
def __init__(self, gamma=1.3, max_mult=3): fc.store_attr()

Expand All @@ -224,7 +237,7 @@ def cleanup_fit(self, learn):
plt.plot(self.lrs, self.losses)
plt.xscale('log')

# %% ../nbs/09_learner.ipynb 61
# %% ../nbs/09_learner.ipynb 62
@fc.patch
def lr_find(self:Learner, gamma=1.3, max_mult=3, start_lr=1e-5, max_epochs=10):
self.fit(max_epochs, lr=start_lr, cbs=LRFinderCB(gamma=gamma, max_mult=max_mult))
Loading