Two fixes:

1. Overwrite previous CloudTuner Runs to avoid erroring out due to failed runs.
2. Remove version number  for kerastuner and tfc in example to use latest.

PiperOrigin-RevId: 361048004

Author: SinaChavoshi

examples/hp_tuning_cifar10_using_google_cloud.ipynb

@@ -286,7 +286,7 @@
         "    max_trials=100,\n",
         "    directory=GCS_BASE_PATH,\n",
         "    study_id=STUDY_ID,\n",
-        "    overwrite=False,\n",
+        "    overwrite=True,\n",
         "    distribution_strategy=distribution_strategy)"
       ]
     },
@@ -339,8 +339,8 @@
         "with open('requirements.txt','w') as f:\n",
         "    f.write('pandas==1.1.5\\n')\n",
         "    f.write('numpy==1.18.5\\n')\n",
-        "    f.write('tensorflow-cloud==0.1.11\\n')\n",
-        "    f.write('keras-tuner==1.0.2\\n')\n",
+        "    f.write('tensorflow-cloud\\n')\n",
+        "    f.write('keras-tuner\\n')\n",
         "\n",
         "# Optional: Some recommended base images. If you provide none the system will choose one for you.\n",
         "TF_GPU_IMAGE= \"tensorflow/tensorflow:latest-gpu\"\n",

examples/hp_tuning_wide_and_deep_model.ipynb

@@ -429,7 +429,7 @@
         "    max_trials=100,\n",
         "    directory=GCS_BASE_PATH,\n",
         "    study_id=STUDY_ID,\n",
-        "    overwrite=False,\n",
+        "    overwrite=True,\n",
         "    distribution_strategy=distribution_strategy)"
       ]
     },