Not detecting large spikes in data

[zdalvabaird]

I am using the cosine model on my timeseries which has many different forms of data but it isn't detecting any changes on some of the large spikes. Below here is the model and penalty I am using, and the segment of the graph that I am concerned about. Is there a mistake that I am making which causes it to not detect this as a change point?

for model, p in zip(models=[cosine], pens=[8]):
    print(model)
    try:
        algo = rpt.Pelt(model=models, min_size=0, jump=1)
        algo.fit(signal)
        result = algo.predict(pen=p)