Clarify docstrings that mention assert_consumed()

In response to discussions in tensorflow/tensorflow#52346

*   Rename delay / delayed restorations to defer / deferred restorations to be consistent
*   Clarify that "no more assignments" means no more deferred restorations.
*   Remove .assert_consumed() from unrelated examples. Because of deferred restorations, calling .assert_consumed() immediately after restore() is likely to fail. There's also no need to recommend calling assert_consumed() after every restore(), as assert_consumed() is meant to guard against unexpected overwrites when fine-tuning a model, which isn't common.

PiperOrigin-RevId: 417461429

Author: tensorflower-gardener

site/en/guide/checkpoint.ipynb

@@ -423,7 +423,7 @@
         "\n",
         "The optimizer is in red, regular variables are in blue, and the optimizer slot variables are in orange. The other nodes—for example, representing the `tf.train.Checkpoint`—are in black.\n",
         "\n",
-        "Slot variables are part of the optimizer's state, but are created for a specific variable. For example the `'m'` edges above correspond to momentum, which the Adam optimizer tracks for each variable. Slot variables are only saved in a checkpoint if the variable and the optimizer would both be saved, thus the dashed edges."
+        "Slot variables are part of the optimizer's state, but are created for a specific variable. For example, the `'m'` edges above correspond to momentum, which the Adam optimizer tracks for each variable. Slot variables are only saved in a checkpoint if the variable and the optimizer would both be saved, thus the dashed edges."
       ]
     },
     {
@@ -491,11 +491,11 @@
         "id": "KCcmJ-2j9RUP"
       },
       "source": [
-        "### Delayed restorations\n",
+        "### Deferred restorations\n",
         "\n",
-        "`Layer` objects in TensorFlow may delay the creation of variables to their first call, when input shapes are available. For example the shape of a `Dense` layer's kernel depends on both the layer's input and output shapes, and so the output shape required as a constructor argument is not enough information to create the variable on its own. Since calling a `Layer` also reads the variable's value, a restore must happen between the variable's creation and its first use.\n",
+        "`Layer` objects in TensorFlow may defer the creation of variables to their first call, when input shapes are available. For example, the shape of a `Dense` layer's kernel depends on both the layer's input and output shapes, and so the output shape required as a constructor argument is not enough information to create the variable on its own. Since calling a `Layer` also reads the variable's value, a restore must happen between the variable's creation and its first use.\n",
         "\n",
-        "To support this idiom, `tf.train.Checkpoint` queues restores which don't yet have a matching variable."
+        "To support this idiom, `tf.train.Checkpoint` defers restores which don't yet have a matching variable."
       ]
     },
     {
@@ -506,10 +506,10 @@
       },
       "outputs": [],
       "source": [
-        "delayed_restore = tf.Variable(tf.zeros([1, 5]))\n",
-        "print(delayed_restore.numpy())  # Not restored; still zeros\n",
-        "fake_layer.kernel = delayed_restore\n",
-        "print(delayed_restore.numpy())  # Restored"
+        "deferred_restore = tf.Variable(tf.zeros([1, 5]))\n",
+        "print(deferred_restore.numpy())  # Not restored; still zeros\n",
+        "fake_layer.kernel = deferred_restore\n",
+        "print(deferred_restore.numpy())  # Restored"
       ]
     },
     {

site/en/guide/migrate/migrating_checkpoints.ipynb

@@ -834,7 +834,7 @@
         "\n",
         "**TF2 checkpoints work with Keras's `build()` step**\n",
         "\n",
-        "`tf.train.Checkpoint.restore` has a mechanism called *delayed restoration* which\n",
+        "`tf.train.Checkpoint.restore` has a mechanism called *deferred restoration* which\n",
         "allows `tf.Module` and Keras objects to store variable values if the variable has not yet been created. This allows *initialized* models to load weights and  *build* after.\n",
         "\n",
         "```\n",