scikit TSNE template for multiple perplexities / iterations

tsne_template.py

fig, ax = plt.subplots(nrows=2, ncols=3)

perplexity = [[2, 10, 20],
              [30, 50, 100]]

iterations = [[1000, 1000, 1000],
            [1000, 1000, 1000]]

for ri, row in enumerate(ax):
    for ci, col in enumerate(row):
        perp = perplexity[ri][ci]
        tsne_model = TSNE(n_components=2, perplexity=perp, n_iter=1000, random_state=12345)
        tsne_output = tsne_model.fit_transform(uv_converted) 
        col.scatter(tsne_output[:, 0], tsne_output[:, 1])
        col.set_title("perplexity {}".format(perp))

plt.show()