Task 2 Level 3

[MarieJuhlJ]

DCCRN finetuned for 30 epochs on task_2_level_3 on aligned data had CER of 0.657

DCCRN finetuned for 30 epochs on all of task 2 data had a mean CER score on task_2_level_3 of 0.657

[MarieJuhlJ]

DCCRN finetuned for task_2_level_1 with IR input but evaluated without IR gave a CER of 0.555 (just below the recordings' CER score of 0.557

[MarieJuhlJ]

Training on all of the data of task 2 with IR and extra synthesized data per level for 50 epochs gives a CER score of 0.571 on task 2 level 3.

Training on all of the data of task 2 with IR for 50 epochs gives a CER score of 0.580 on task 2 level 3.

[MarieJuhlJ]

Training on IR filtered data from task 2 level 3 gives a CER score of 0.914 when evaluating with linear filter and 0.869 without linear filter. It seems to overfit the training data quite a lot.

[deding2000]

with regularized linear filter we get CER of 0.108