Add learning curve data from models in task_set.

PiperOrigin-RevId: 297986362

task_set/README.md

@@ -2,15 +2,26 @@
 This directory contains a variety of optimization problems for use in evaluating
 and meta-training learned optimizers. It is decribed in
 "Using a thousand optimization tasks to learn hyperparameter search strategies"
-[arxiv](link_to_paper).
+[arxiv](https://arxiv.org/abs/2002.11887).
 
 The problems are implemented as tensorflow 1.x style models mostly using Sonnet.
 
-# Usage to train models
+## Learning curves of trained models
+As part of this repository we are releasing learning curves and corresponding
+hyperparameters for roughly 29 million models. The data is stored in a cloud
+bucket in npz format here: `gs://task_set_data/task_set_data/`.
+
+For ease of analysis, we provide a sample [colab](https://colab.research.google.com/drive/1BYJzTx2MiJWbM4ydFoQvu2yon65banBj).
+
+We hope this data can be used to gain insight into both optimizers, and probe notions of task similarity.
+See our [paper](https://arxiv.org/abs/2002.11887) for examples of what can be done with data.
+
+## Usage to train models
 In addition to model definition, we also provide training scripts.
 
 `python3 -m task_set.train_inner --optimizer_name="adam4p_wide_grid_seed107" --task_name="mlp_family_seed117" --output_directory="/tmp/root_data_dir"`
 
-# Requirements:
+
+## Requirements:
 As of now, we only support tensorflow version 1.0 (e.g. tensorflow-1.15) and
 1.x sonnet. See requirements.txt for full versions required.