This is an automated email from the ASF dual-hosted git repository. fmcquillan pushed a commit to branch asf-site in repository https://gitbox.apache.org/repos/asf/madlib-site.git
commit 3b665af751366e349420d72fcf1505744fb129cc Author: Domino Valdano <dominoplu...@gmail.com> AuthorDate: Mon Aug 26 13:04:12 2019 -0700 Update demo notebook --- .../Deep-learning/Load-images-v1.ipynb | 20 +++++++------------- 1 file changed, 7 insertions(+), 13 deletions(-) diff --git a/community-artifacts/Deep-learning/Load-images-v1.ipynb b/community-artifacts/Deep-learning/Load-images-v1.ipynb index 15aa948..3209aaf 100644 --- a/community-artifacts/Deep-learning/Load-images-v1.ipynb +++ b/community-artifacts/Deep-learning/Load-images-v1.ipynb @@ -193,7 +193,7 @@ "<a id=\"fetch_numpy\"></a>\n", "# 2. Fetch images then load NumPy array into table\n", "\n", - "<em>iloader.load_dataset_from_np(data_x, data_y, table_name, append=False, no_temp_files=False)</em>\n", + "<em>iloader.load_dataset_from_np(data_x, data_y, table_name, append=False)</em>\n", "\n", "- <em>data_x</em> contains image data in np.array format\n", "\n", @@ -204,13 +204,7 @@ "- If the user passes a <em>table_name</em> while creating ImageLoader object, it will be used for all further calls to load_dataset_from_np. It can be changed by passing it as a parameter during the actual call to load_dataset_from_np, and if so future calls will load to that table name instead. This avoids needing to pass the table_name again every time, but also allows it to be changed at any time.\n", "\n", " \n", - "- <em>append=False</em> attempts to create a new table, while <em>append=True</em> appends more images to an existing table.\n", - "\n", - "\n", - "- EXPERIMENTAL: If <em>no_temp_files=True</em>, the operation will happen without\n", - " writing out the tables to temporary files before loading them.\n", - " Instead, an in-memory filelike buffer (StringIO) will be used\n", - " to build the tables before loading." + "- <em>append=False</em> attempts to create a new table, while <em>append=True</em> appends more images to an existing table." ] }, { @@ -420,8 +414,8 @@ "%sql DROP TABLE IF EXISTS cifar_10_train_data, cifar_10_test_data;\n", "\n", "# Save images to temporary directories and load into database\n", - "iloader.load_dataset_from_np(x_train, y_train, 'cifar_10_train_data', append=False, no_temp_files=False)\n", - "iloader.load_dataset_from_np(x_test, y_test, 'cifar_10_test_data', append=False, no_temp_files=False)" + "iloader.load_dataset_from_np(x_train, y_train, 'cifar_10_train_data', append=False)\n", + "iloader.load_dataset_from_np(x_test, y_test, 'cifar_10_test_data', append=False)" ] }, { @@ -434,12 +428,12 @@ "Uses the Python Imaging Library so supports multiple formats\n", "http://www.pythonware.com/products/pil/\n", "\n", - "<em>load_dataset_from_disk(root_dir, table_name, num_labels='all', append=False, no_temp_files=False)</em>\n", + "<em>load_dataset_from_disk(root_dir, table_name, num_labels='all', append=False)</em>\n", "\n", "- Calling this function will look in <em>root_dir</em> on the local disk of wherever this is being run. It will skip over any files in that directory, but will load images contained in each of its subdirectories. The images should be organized by category/class, where the name of each subdirectory is the label for the images contained within it.\n", "\n", "\n", - "- The <em>table_name</em>, <em>append</em>, and <em>no_temp_files</em> parameters are the same as above The parameter <em>num_labels</em> is an optional parameter which can be used to restrict the number of labels (image classes) loaded, even if more are found in <em>root_dir</em>. For example, for a large dataset you may have hundreds of labels, but only wish to use a subset of that containing a few dozen.\n", + "- The <em>table_name</em> and <em>append</em> parameters are the same as above The parameter <em>num_labels</em> is an optional parameter which can be used to restrict the number of labels (image classes) loaded, even if more are found in <em>root_dir</em>. For example, for a large dataset you may have hundreds of labels, but only wish to use a subset of that containing a few dozen.\n", "\n", "For example, if we put the CIFAR-10 training data is in 10 subdirectories under directory <em>cifar10</em>, with one subdirectory for each class:" ] @@ -596,7 +590,7 @@ "source": [ "%sql drop table if exists cifar_10_train_data_filesystem;\n", "# Load images from file system\n", - "iloader.load_dataset_from_disk('/Users/fmcquillan/tmp/cifar10', 'cifar_10_train_data_filesystem', num_labels='all', append=False, no_temp_files=False)" + "iloader.load_dataset_from_disk('/Users/fmcquillan/tmp/cifar10', 'cifar_10_train_data_filesystem', num_labels='all', append=False)" ] }, {