readme fix

README.md

@@ -57,11 +57,11 @@ In this project we implement different alternatives to it.
 ## 🗃 Algorithms to implement (from simplest to hardest)
 - [x] [Relaxed Bernoulli](http://proceedings.mlr.press/v119/yamada20a/yamada20a.pdf) 
 - [x] [Correlated relaxed Bernoulli](https://openreview.net/pdf?id=oDFvtxzPOx)
-- [ ] [Gumbel-softmax TOP-K](https://arxiv.org/pdf/1903.06059) 
+- [x] [Gumbel-softmax TOP-K](https://arxiv.org/pdf/1903.06059) 
 - [x] [Straight-Through Bernoulli, distribution (don't mix with Relaxed distribution from pyro)](https://citeseerx.ist.psu.edu/document?repid=rep1&type=pdf&doi=62c76ca0b2790c34e85ba1cce09d47be317c7235) 
 - [x] [Invertible Gaussian reparametrization](https://arxiv.org/abs/1912.09588) with KL implemented
 - [x] [Hard concrete](https://arxiv.org/pdf/1712.01312) 
-- [ ] [REINFORCE](http://www.cs.toronto.edu/~tingwuwang/REINFORCE.pdf)  (not a distribution actually, think how to integrate it with other distributions)
+- [x] [REINFORCE](http://www.cs.toronto.edu/~tingwuwang/REINFORCE.pdf)  (not a distribution actually, think how to integrate it with other distributions)
 - [x] [Logit-normal distribution](https://en.wikipedia.org/wiki/Logit-normal_distribution) and [Laplace-form approximation of Dirichlet](https://stats.stackexchange.com/questions/535560/approximating-the-logit-normal-by-dirichlet) 
 
 ## 📚 Recommended stack