Ideally you'd still clearer labels for features, though having the table of features in the beginning of the report helps.
You've implemented my feedback for Fig 2 well. Your peers are still finding this crowded due to number of graphs - can you trim it down?
For Fig 3, make sure your caption and description use the same terms as your legend (you may want to fix the legend?)
Table 1, you can definitely give the features more readable labels.
Additional thought on Table 2: this one might make sense to re-shape so that you have columns "training accuracy" and "testing accuracy" for rows "no feature selection" and "selected features".
Peer feedback: "Include ROC curve, f1 score etc." either address this, mention in issue why you don't think it makes sense to address this, or mention this as something to do in your future directions and then leave the issue open.