Adjust demo scripts to be Keras 3 compatible (#6761)

This adjusts *_demo.py files to work with Keras 3.

The hparams_demo is fully backward compatible with Keras 2 and forward
compatible with Keras 3.

Unfortunately the graphs_demo is not backward compatible with Keras 2.
Users attempting to run it with Keras 2 will get the following error:

```
  File "/usr/local/google/home/bdubois/.cache/bazel/_bazel_bdubois/079646a57be11faea0b2bfefccb2a81a/execroot/org_tensorflow_tensorboard/bazel-out/k8-fastbuild/bin/tensorboard/plugins/graph/graphs_demo.runfiles/org_tensorflow_tensorboard/tensorboard/plugins/graph/graphs_demo.py", line 128, in profile
    tf.summary.trace_on(profiler=True, profiler_outdir=logdir)
TypeError: trace_on() got an unexpected keyword argument 'profiler_outdir'
```

Amazingly, though, the graph that the demo generates with Keras 3 can be
successfully loaded in the Graph dashboard. This makes me optimistic to
get the Graph plugin fully Keras 3 compatible after addressing the
user-reported error in #6686.

Old Keras 2 Graph: 

![image](https://github.com/tensorflow/tensorboard/assets/17152369/b8745739-ac06-4171-a7bc-c97135b2dec7)
New Keras 3 Graph:

![image](https://github.com/tensorflow/tensorboard/assets/17152369/04dcaf14-3464-47dc-b4a5-373cccd8370b)

tensorboard/plugins/graph/graphs_demo.py

@@ -90,7 +90,7 @@ def keras():
     model = tf.keras.models.Sequential(layers)
     model.compile(
         loss=tf.keras.losses.mean_squared_error,
-        optimizer=tf.keras.optimizers.SGD(lr=0.2),
+        optimizer=tf.keras.optimizers.SGD(learning_rate=0.2),
     )
     model.fit(
         x_train,
@@ -125,11 +125,9 @@ def g(i):
         for step in range(3):
             # Suppress the profiler deprecation warnings from tf.summary.trace_*.
             with _silence_deprecation_warnings():
-                tf.summary.trace_on(profiler=True)
+                tf.summary.trace_on(profiler=True, profiler_outdir=logdir)
                 print(f(tf.constant(step)).numpy())
-                tf.summary.trace_export(
-                    "prof_f", step=step, profiler_outdir=logdir
-                )
+                tf.summary.trace_export("prof_f", step=step)
 
                 tf.summary.trace_on(profiler=False)
                 print(g(tf.constant(step)).numpy())

tensorboard/plugins/hparams/hparams_demo.py

@@ -137,7 +137,11 @@ def model_fn(hparams, seed):
         conv_filters *= 2
 
     model.add(tf.keras.layers.Flatten())
-    model.add(tf.keras.layers.Dropout(hparams[HP_DROPOUT], seed=rng.random()))
+    model.add(
+        tf.keras.layers.Dropout(
+            hparams[HP_DROPOUT], seed=rng.randrange(1 << 32)
+        )
+    )
 
     # Add fully connected layers.
     dense_neurons = 32