python Precision and F-score are ill-defined

Question

원래 소스코드는 label이 mask, withoutmask 두개였는데 저는 mask, withoutmask, wrongmask 해서 마스크 감지할려고 소스를 수정중입니다. 근데 wrongmask파일을 만들고 실행을 했는데 wrong_mask에서는 precision, recall f1-score가 다 0이뜹니다. UndefinedMetricWarning: Precision and F-score are ill-defined and being set to 0.0 in labels with no predicted samples. Use &#96;zero_division&#96; parameter to control this behavior. _warn_prf(average, modifier, msg_start, len(result)) precision recall f1-score support with_mask 0.95 0.99 0.97 373without_mask 0.99 0.99 0.99 386 wrong_mask 0.00 0.00 0.00 16 accuracy 0.97 775 macro avg 0.65 0.66 0.65 775weighted avg 0.95 0.97 0.96 775# train the head of the network print(&#34;[INFO] training head...&#34;) H &#61; model.fit( aug.flow(trainX, trainY, batch_size&#61;BS), steps_per_epoch&#61;len(trainX) // BS, validation_data&#61;(testX, testY), validation_steps&#61;len(testX) // BS, epochs&#61;EPOCHS) # make predictions on the testing set print(&#34;[INFO] evaluating network...&#34;) predIdxs &#61; model.predict(testX, batch_size&#61;BS) # for each image in the testing set we need to find the index of the # label with corresponding largest predicted probability predIdxs &#61; np.argmax(predIdxs, axis&#61;1) # show a nicely formatted classification report print(classification_report(testY.argmax(axis&#61;1), predIdxs, target_names&#61;labelNames)) # serialize the model to disk print(&#34;[INFO] saving mask detector model...&#34;) model.save(args[&#34;model&#34;], save_format&#61;&#34;h5&#34;) # plot the training loss and accuracy N &#61; EPOCHS plt.style.use(&#34;ggplot&#34;) plt.figure() plt.plot(np.arange(0, N), H.history[&#34;loss&#34;], label&#61;&#34;train_loss&#34;) plt.plot(np.arange(0, N), H.history[&#34;val_loss&#34;], label&#61;&#34;val_loss&#34;) plt.plot(np.arange(0, N), H.history[&#34;accuracy&#34;], label&#61;&#34;train_acc&#34;) plt.plot(np.arange(0, N), H.history[&#34;val_accuracy&#34;], label&#61;&#34;val_acc&#34;) plt.title(&#34;Training Loss and Accuracy&#34;) plt.xlabel(&#34;Epoch #&#34;) plt.ylabel(&#34;Loss/Accuracy&#34;) plt.legend(loc&#61;&#34;lower left&#34;) plt.savefig(args[&#34;plot&#34;])제 생각에는print(classification_report(testY.argmax(axis&#61;1), predIdxs, target_names&#61;labelNames))이 행에서 제대로 안되는 것 같은데,어떻게 해결할 수 있을까요.

python Precision and F-score are ill-defined

0개의 답변

#machine-learning 목록