From ef1307c25306c404492f1aae1d13a316dd6832b5 Mon Sep 17 00:00:00 2001 From: William L Hamilton Date: Wed, 31 May 2017 14:51:58 +0100 Subject: [PATCH] Update README.md --- README.md | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/README.md b/README.md index c82837c..db46159 100644 --- a/README.md +++ b/README.md @@ -47,9 +47,10 @@ The user must also specify a --model, the variants of which are described in det #### Logging directory Finally, a --base_log_dir should be specified (it defaults to the current directory). The output of the model and log files will be stored in a subdirectory of the base_log_dir. +The path to the logged data will be of the form `-/graphsage-/`. The supervised model will output F1 scores, while the unsupervised model will train embeddings and store them. -The unsupervised embeddings will be stored at val.npy with val.txt specifying the order of embeddings as a per-line list of node ids. -Note that the full log outputs and stored embeddings can be 5-10Gb in size (on the full data). +The unsupervised embeddings will be stored in a numpy formated file named val.npy with val.txt specifying the order of embeddings as a per-line list of node ids. +Note that the full log outputs and stored embeddings can be 5-10Gb in size (on the full data when running with the unsupervised variant). #### Using the output of the unsupervised models