.

optim_nn_core.py

@@ -723,7 +723,7 @@ class Ritual: # i'm just making up names at this point
             batch_outputs = outputs[bi:bi+batch_size]
 
             if not test_only and self.learner.per_batch:
-                    self.learner.batch(b / batch_count)
+                self.learner.batch(b / batch_count)
 
             predicted = self.learn(batch_inputs, batch_outputs)
             if not test_only:

optim_nn_mnist.py

@@ -1,6 +1,7 @@
 #!/usr/bin/env python3
 
 from optim_nn import *
+from optim_nn_core import _f
 
 #np.random.seed(42069)
 
@@ -13,8 +14,12 @@ from optim_nn import *
 lr = 0.01
 epochs = 24
 starts = 2
+restart_decay = 0.5
 bs = 100
 
+log_fn = 'mnist_losses.npz'
+measure_every_epoch = True
+
 mnist_dim = 28
 mnist_classes = 10
 def get_mnist(fn='mnist.npz'):
@@ -60,7 +65,7 @@ model = Model(x, y, unsafe=True)
 
 optim = Adam()
 learner = SGDR(optim, epochs=epochs//starts, rate=lr,
-               restarts=starts - 1, restart_decay=0.5,
+               restarts=starts - 1, restart_decay=restart_decay,
                expando=lambda i:0)
 
 loss = CategoricalCrossentropy()
@@ -72,15 +77,24 @@ log('parameters', model.param_count)
 
 ritual.prepare(model)
 
-def measure_error():
+batch_losses, batch_mlosses = [], []
+train_losses, train_mlosses = [], []
+valid_losses, valid_mlosses = [], []
+
+def measure_error(quiet=False):
     def print_error(name, inputs, outputs, comparison=None):
         loss, mloss, _, _ = ritual.test_batched(inputs, outputs, bs, return_losses='both')
-        log(name + " loss", "{:12.6e}".format(loss))
-        log(name + " accuracy", "{:6.2f}%".format(mloss * 100))
+        if not quiet:
+            log(name + " loss", "{:12.6e}".format(loss))
+            log(name + " accuracy", "{:6.2f}%".format(mloss * 100))
         return loss, mloss
 
-    print_error("train", inputs, outputs)
-    print_error("valid", valid_inputs, valid_outputs)
+    loss, mloss = print_error("train", inputs, outputs)
+    train_losses.append(loss)
+    train_mlosses.append(mloss)
+    loss, mloss = print_error("valid", valid_inputs, valid_outputs)
+    valid_losses.append(loss)
+    valid_mlosses.append(mloss)
 
 measure_error()
 
@@ -90,7 +104,7 @@ while learner.next():
     shuffled_inputs = inputs[indices]
     shuffled_outputs = outputs[indices]
 
-    avg_loss, avg_mloss, _, _ = ritual.train_batched(
+    avg_loss, avg_mloss, losses, mlosses = ritual.train_batched(
         shuffled_inputs, shuffled_outputs,
         batch_size=bs,
         return_losses='both')
@@ -98,4 +112,22 @@ while learner.next():
     log("epoch {}".format(learner.epoch + 1),
         fmt.format(learner.rate, avg_loss, avg_mloss * 100))
 
-measure_error()
+    batch_losses += losses
+    batch_mlosses += mlosses
+
+    if measure_every_epoch:
+        quiet = learner.epoch + 1 != learner.epochs
+        measure_error(quiet=quiet)
+
+if not measure_every_epoch:
+    measure_error()
+
+if log_fn:
+    log('saving losses', log_fn)
+    np.savez_compressed(log_fn,
+                        batch_losses =np.array(batch_losses,  dtype=_f),
+                        batch_mlosses=np.array(batch_mlosses, dtype=_f),
+                        train_losses =np.array(train_losses,  dtype=_f),
+                        train_mlosses=np.array(train_mlosses, dtype=_f),
+                        valid_losses =np.array(valid_losses,  dtype=_f),
+                        valid_mlosses=np.array(valid_mlosses, dtype=_f))