Update README.md

Author: herilalaina

README.md

@@ -2,6 +2,8 @@
 
 # `ifBO`: In-context Freeze-Thaw Bayesian Optimization for Hyperparameter Optimization
 
+[![Open in Spaces](https://huggingface.co/datasets/huggingface/badges/resolve/main/open-in-hf-spaces-md.svg)](https://huggingface.co/spaces/herilalaina/ifbo)
+
 This repository contains the official code for our [ICML 2024 paper](https://openreview.net/forum?id=VyoY3Wh9Wd). `ifBO` is an efficient Bayesian Optimization algorithm that dynamically selects and incrementally evaluates candidates during the optimization process. It uses a model called the `Freeze-Thaw surrogate (FT-PFN)` to predict the performance of candidate configurations as more resources are allocated. The `main` branch includes the necessary API to use `FT-PFN`. Refer to the following sections:
 - [Surrogate API](#surrogate-api): to learn how to initialize and use the surrogate model.
 - [Bayesian Optimization with ifBO](#bayesian-optimization-with-ifbo): to understand how to use `ifBO` for Hyperparameter Optimization.
@@ -101,8 +103,6 @@ Following the PFN approach, the FT-PFN model outputs the Predictive Posterior Di
 
 ## Bayesian Optimization with ifBO
 
-[![Open in Spaces](https://huggingface.co/datasets/huggingface/badges/resolve/main/open-in-hf-spaces-md.svg)](https://huggingface.co/spaces/herilalaina/ifbo)
-
 To use the `ifBO` algorithm in practice, refer to [NePS](https://automl.github.io/neps/latest/), a package for hyperparameter optimization that includes the latest and improved version of `ifBO`. Below is a template example of how to use `ifBO` with NePS. For a complete Python script, see the [full example](https://github.com/automl/neps/blob/master/neps_examples/efficiency/freeze_thaw.py).
 
 ```python