diff --git a/miniai/learner.py b/miniai/learner.py
index 2f67522..f2b5665 100644
--- a/miniai/learner.py
+++ b/miniai/learner.py
@@ -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)
+
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()]
@@ -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):
@@ -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()
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):
@@ -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)
@@ -180,7 +193,7 @@ 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])
@@ -188,7 +201,7 @@ 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
@@ -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()
@@ -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))
diff --git a/nbs/09_learner.ipynb b/nbs/09_learner.ipynb
index 2ec0851..756cc74 100644
--- a/nbs/09_learner.ipynb
+++ b/nbs/09_learner.ipynb
@@ -90,7 +90,7 @@
{
"data": {
"application/vnd.jupyter.widget-view+json": {
- "model_id": "6ca050462ee540518c7028378f76b22d",
+ "model_id": "771a91cb084446fb9799692c04c6e7fc",
"version_major": 2,
"version_minor": 0
},
@@ -219,8 +219,8 @@
"name": "stdout",
"output_type": "stream",
"text": [
- "0 True 1.1753045572916667 0.5986833333333333\n",
- "0 False 1.12032890625 0.6135285714285714\n"
+ "0 True 1.17530625 0.5986166666666667\n",
+ "0 False 1.1203782366071429 0.6133857142857143\n"
]
}
],
@@ -364,7 +364,7 @@
"outputs": [],
"source": [
"m,nh = 28*28,50\n",
- "def get_model(): return nn.Sequential(nn.Linear(m,nh), nn.ReLU(), nn.Linear(nh,10))"
+ "def get_model(nh=nh): return nn.Sequential(nn.Linear(m,nh), nn.ReLU(), nn.Linear(nh,10))"
]
},
{
@@ -596,13 +596,16 @@
"outputs": [],
"source": [
"#|export\n",
+ "from torcheval.metrics import Metric, Mean\n",
+ "\n",
"class MetricsCB(Callback):\n",
- " def __init__(self, *ms, **metrics):\n",
+ " def __init__(self, *ms, device=def_device, **metrics):\n",
" for o in ms: metrics[type(o).__name__] = o\n",
" self.metrics = metrics\n",
+ " for m in self.metrics.values(): m.to(device)\n",
" self.all_metrics = copy(metrics)\n",
- " self.all_metrics['loss'] = self.loss = Mean()\n",
- "\n",
+ " self.all_metrics['loss'] = self.loss = Mean(device='cpu' if 'mps' in device else device)\n",
+ " \n",
" def _log(self, d): print(d)\n",
" def before_fit(self, learn): learn.metrics = self\n",
" def before_epoch(self, learn): [o.reset() for o in self.all_metrics.values()]\n",
@@ -614,9 +617,9 @@
" self._log(log)\n",
"\n",
" def after_batch(self, learn):\n",
- " x,y,*_ = to_cpu(learn.batch)\n",
- " for m in self.metrics.values(): m.update(to_cpu(learn.preds), y)\n",
- " self.loss.update(to_cpu(learn.loss), weight=len(x))"
+ " x,y,*_ = learn.batch\n",
+ " for m in self.metrics.values(): m.update(learn.preds.to(m.device), y)\n",
+ " self.loss.update(learn.loss.to(self.loss.device), weight=len(x))"
]
},
{
@@ -758,41 +761,54 @@
"#|export\n",
"class ProgressCB(Callback):\n",
" order = MetricsCB.order+1\n",
- " def __init__(self, plot=False): self.plot = plot\n",
+ " def __init__(self, plot=False, lag=10): fc.store_attr()\n",
" def before_fit(self, learn):\n",
" learn.epochs = self.mbar = master_bar(learn.epochs)\n",
" self.first = True\n",
" if hasattr(learn, 'metrics'): learn.metrics._log = self._log\n",
" self.losses = []\n",
- "\n",
+ " self.gpu_losses = [] \n",
+ " \n",
" def _log(self, d):\n",
" if self.first:\n",
" self.mbar.write(list(d), table=True)\n",
" self.first = False\n",
" self.mbar.write(list(d.values()), table=True)\n",
"\n",
+ " def _plot(self, lag=0): \n",
+ " n = max(0,len(self.gpu_losses)-lag)\n",
+ " if n == 0: return\n",
+ " self.losses, self.gpu_losses = self.losses + [l.item() for l in self.gpu_losses[:n]], self.gpu_losses[n:]\n",
+ " self.mbar.update_graph([[fc.L.range(self.losses), self.losses]])\n",
+ " \n",
" def before_epoch(self, learn): learn.dl = progress_bar(learn.dl, leave=False, parent=self.mbar)\n",
" def after_batch(self, learn):\n",
- " learn.dl.comment = f'{learn.loss:.3f}'\n",
" if self.plot and hasattr(learn, 'metrics') and learn.training:\n",
- " self.losses.append(learn.loss.item())\n",
- " self.mbar.update_graph([[fc.L.range(self.losses), self.losses]])"
+ " self.gpu_losses.append(learn.loss.detach())\n",
+ " if len(self.gpu_losses) > 2* self.lag: self._plot(self.lag)\n",
+ " \n",
+ " def after_epoch(self, learn):\n",
+ " learn.dl.comment = f'{learn.loss:.3f}'\n",
+ " if learn.training: self._plot()"
]
},
{
"cell_type": "code",
"execution_count": null,
- "id": "b82dcb40",
+ "id": "92c0716f-4f85-4468-9a8c-8488ec089587",
"metadata": {},
"outputs": [],
"source": [
- "model = get_model()"
+ "# let's test the lag property and device metrics processing speed ignoring dataset loading time\n",
+ "dlsc = DataLoaders.from_dd(tds, batch_size=bs)\n",
+ "dlsc.train = list(dlsc.train)\n",
+ "dlsc.valid = list(dlsc.valid)"
]
},
{
"cell_type": "code",
"execution_count": null,
- "id": "3b77daf3",
+ "id": "996321ea-1294-490a-bf3c-b6d0fdc8451a",
"metadata": {},
"outputs": [
{
@@ -807,6 +823,9 @@
" /* Needs to be in here for Safari polyfill so background images work as expected. */\n",
" background-size: auto;\n",
" }\n",
+ " progress:not([value]), progress:not([value])::-webkit-progress-bar {\n",
+ " background: repeating-linear-gradient(45deg, #7e7e7e, #7e7e7e 10px, #5c5c5c 10px, #5c5c5c 20px);\n",
+ " }\n",
" .progress-bar-interrupted, .progress-bar-interrupted::-webkit-progress-bar {\n",
" background: #F44336;\n",
" }\n",
@@ -833,14 +852,14 @@
" \n",
"
\n",
" \n",
- " 0.596 | \n",
- " 1.167 | \n",
+ " 0.602 | \n",
+ " 1.180 | \n",
" 0 | \n",
" train | \n",
"
\n",
" \n",
- " 0.729 | \n",
- " 0.794 | \n",
+ " 0.725 | \n",
+ " 0.780 | \n",
" 0 | \n",
" eval | \n",
"
\n",
@@ -856,19 +875,123 @@
},
{
"data": {
- "image/png": "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\n",
+ "image/png": "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\n",
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
+ },
+ {
+ "name": "stdout",
+ "output_type": "stream",
+ "text": [
+ "4.39 s ± 0 ns per loop (mean ± std. dev. of 1 run, 1 loop each)\n"
+ ]
}
],
"source": [
+ "%%timeit -n1 -r1\n",
+ "model = get_model()\n",
"metrics = MetricsCB(accuracy=MulticlassAccuracy())\n",
- "cbs = [TrainCB(), DeviceCB(), metrics, ProgressCB(plot=True)]\n",
- "learn = Learner(model, dls, F.cross_entropy, lr=0.2, cbs=cbs)\n",
+ "cbs = [TrainCB(), DeviceCB(), metrics, ProgressCB(plot=True, lag=0)]\n",
+ "learn = Learner(model, dlsc, F.cross_entropy, lr=0.2, cbs=cbs)\n",
+ "learn.fit(1)"
+ ]
+ },
+ {
+ "cell_type": "code",
+ "execution_count": null,
+ "id": "03ee4172-78bb-469f-a6a5-d853c6f3b5b1",
+ "metadata": {},
+ "outputs": [
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ "\n"
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "text/html": [
+ "\n",
+ " \n",
+ " \n",
+ " accuracy | \n",
+ " loss | \n",
+ " epoch | \n",
+ " train | \n",
+ "
\n",
+ " \n",
+ " \n",
+ " \n",
+ " 0.594 | \n",
+ " 1.169 | \n",
+ " 0 | \n",
+ " train | \n",
+ "
\n",
+ " \n",
+ " 0.708 | \n",
+ " 0.796 | \n",
+ " 0 | \n",
+ " eval | \n",
+ "
\n",
+ " \n",
+ "
"
+ ],
+ "text/plain": [
+ ""
+ ]
+ },
+ "metadata": {},
+ "output_type": "display_data"
+ },
+ {
+ "data": {
+ "image/png": "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\n",
+ "text/plain": [
+ "