Slight update

README.md

@@ -17,7 +17,7 @@ widget:
 
 # Ukrainian flair embeddings (forward)
 
-Trained for 10 epochs on the texts from ubertext2.0 (WIP).
+Trained for 12+ epochs on the texts from ubertext2.0 (WIP).
 The characters dictionary used for training is in `flair_dictionary.pkl` file
 
 For more information on flair embeddings see [the article](https://github.com/flairNLP/flair/blob/master/resources/docs/embeddings/FLAIR_EMBEDDINGS.md) or the paper below:

best-lm.pt

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:d952419535bc66f60333174cad83dac8e95b83da1546a1ea5303823d758c1802
+oid sha256:fdae6a5c00bd208946ba0adf9f2f455701fdc71bb91157fc3372834ed8332e56
 size 22791455

loss.txt

@@ -306,3 +306,78 @@
 | end of split  54 / 28 | epoch  10 | time: 3388.66s | valid loss 1.0212 | valid ppl 2.7766 | learning rate 5.0000
 | end of split  55 / 28 | epoch  10 | time: 3386.75s | valid loss 1.0212 | valid ppl 2.7764 | learning rate 5.0000
 | end of split  56 / 28 | epoch  10 | time: 3386.25s | valid loss 1.0213 | valid ppl 2.7767 | learning rate 5.0000
+| end of split  29 / 28 | epoch  11 | time: 3361.50s | valid loss 1.0212 | valid ppl 2.7765 | learning rate 5.0000
+| end of split  30 / 28 | epoch  11 | time: 3388.02s | valid loss 1.0212 | valid ppl 2.7765 | learning rate 5.0000
+| end of split  31 / 28 | epoch  11 | time: 3389.23s | valid loss 1.0211 | valid ppl 2.7761 | learning rate 5.0000
+| end of split  32 / 28 | epoch  11 | time: 3376.47s | valid loss 1.0210 | valid ppl 2.7760 | learning rate 5.0000
+| end of split  33 / 28 | epoch  11 | time: 3378.54s | valid loss 1.0211 | valid ppl 2.7763 | learning rate 5.0000
+| end of split  34 / 28 | epoch  11 | time: 3371.86s | valid loss 1.0210 | valid ppl 2.7761 | learning rate 5.0000
+| end of split  35 / 28 | epoch  11 | time: 988.54s | valid loss 1.0211 | valid ppl 2.7762 | learning rate 5.0000
+| end of split  36 / 28 | epoch  11 | time: 3369.15s | valid loss 1.0210 | valid ppl 2.7761 | learning rate 5.0000
+| end of split  37 / 28 | epoch  11 | time: 3362.72s | valid loss 1.0209 | valid ppl 2.7758 | learning rate 5.0000
+| end of split  38 / 28 | epoch  11 | time: 3363.26s | valid loss 1.0210 | valid ppl 2.7759 | learning rate 5.0000
+| end of split  39 / 28 | epoch  11 | time: 3359.86s | valid loss 1.0210 | valid ppl 2.7760 | learning rate 5.0000
+| end of split  40 / 28 | epoch  11 | time: 3338.89s | valid loss 1.0209 | valid ppl 2.7758 | learning rate 5.0000
+| end of split  41 / 28 | epoch  11 | time: 3356.02s | valid loss 1.0209 | valid ppl 2.7756 | learning rate 5.0000
+| end of split  42 / 28 | epoch  11 | time: 3351.44s | valid loss 1.0208 | valid ppl 2.7753 | learning rate 5.0000
+| end of split  43 / 28 | epoch  11 | time: 3350.87s | valid loss 1.0207 | valid ppl 2.7751 | learning rate 5.0000
+| end of split  44 / 28 | epoch  11 | time: 3346.91s | valid loss 1.0207 | valid ppl 2.7752 | learning rate 5.0000
+| end of split  45 / 28 | epoch  11 | time: 3348.82s | valid loss 1.0206 | valid ppl 2.7749 | learning rate 5.0000
+| end of split  46 / 28 | epoch  11 | time: 3348.50s | valid loss 1.0207 | valid ppl 2.7750 | learning rate 5.0000
+| end of split  47 / 28 | epoch  11 | time: 3346.52s | valid loss 1.0206 | valid ppl 2.7748 | learning rate 5.0000
+| end of split  48 / 28 | epoch  11 | time: 3341.43s | valid loss 1.0206 | valid ppl 2.7748 | learning rate 5.0000
+| end of split  49 / 28 | epoch  11 | time: 3342.42s | valid loss 1.0205 | valid ppl 2.7747 | learning rate 5.0000
+| end of split  50 / 28 | epoch  11 | time: 3361.90s | valid loss 1.0205 | valid ppl 2.7747 | learning rate 5.0000
+| end of split  51 / 28 | epoch  11 | time: 3373.79s | valid loss 1.0205 | valid ppl 2.7745 | learning rate 5.0000
+| end of split  52 / 28 | epoch  11 | time: 3380.88s | valid loss 1.0205 | valid ppl 2.7746 | learning rate 5.0000
+| end of split  53 / 28 | epoch  11 | time: 3380.44s | valid loss 1.0204 | valid ppl 2.7743 | learning rate 5.0000
+| end of split  54 / 28 | epoch  11 | time: 3379.94s | valid loss 1.0204 | valid ppl 2.7743 | learning rate 5.0000
+| end of split  55 / 28 | epoch  11 | time: 3379.47s | valid loss 1.0204 | valid ppl 2.7742 | learning rate 5.0000
+| end of split  56 / 28 | epoch  11 | time: 3380.66s | valid loss 1.0204 | valid ppl 2.7742 | learning rate 5.0000
+| end of split  29 / 28 | epoch  12 | time: 3378.16s | valid loss 1.0206 | valid ppl 2.7749 | learning rate 5.0000
+| end of split  30 / 28 | epoch  12 | time: 3397.83s | valid loss 1.0205 | valid ppl 2.7746 | learning rate 5.0000
+| end of split  31 / 28 | epoch  12 | time: 3392.19s | valid loss 1.0204 | valid ppl 2.7742 | learning rate 5.0000
+| end of split  32 / 28 | epoch  12 | time: 3379.40s | valid loss 1.0204 | valid ppl 2.7743 | learning rate 5.0000
+| end of split  33 / 28 | epoch  12 | time: 3373.61s | valid loss 1.0203 | valid ppl 2.7740 | learning rate 5.0000
+| end of split  34 / 28 | epoch  12 | time: 3369.09s | valid loss 1.0202 | valid ppl 2.7738 | learning rate 5.0000
+| end of split  35 / 28 | epoch  12 | time: 3370.15s | valid loss 1.0202 | valid ppl 2.7738 | learning rate 5.0000
+| end of split  36 / 28 | epoch  12 | time: 3364.76s | valid loss 1.0202 | valid ppl 2.7736 | learning rate 5.0000
+| end of split  37 / 28 | epoch  12 | time: 3362.81s | valid loss 1.0202 | valid ppl 2.7738 | learning rate 5.0000
+| end of split  38 / 28 | epoch  12 | time: 3361.73s | valid loss 1.0201 | valid ppl 2.7736 | learning rate 5.0000
+| end of split  39 / 28 | epoch  12 | time: 3362.24s | valid loss 1.0201 | valid ppl 2.7734 | learning rate 5.0000
+| end of split  40 / 28 | epoch  12 | time: 3349.23s | valid loss 1.0201 | valid ppl 2.7735 | learning rate 5.0000
+| end of split  41 / 28 | epoch  12 | time: 3349.66s | valid loss 1.0200 | valid ppl 2.7732 | learning rate 5.0000
+| end of split  42 / 28 | epoch  12 | time: 3354.36s | valid loss 1.0200 | valid ppl 2.7733 | learning rate 5.0000
+| end of split  43 / 28 | epoch  12 | time: 3337.30s | valid loss 1.0200 | valid ppl 2.7731 | learning rate 5.0000
+| end of split  44 / 28 | epoch  12 | time: 3354.63s | valid loss 1.0200 | valid ppl 2.7733 | learning rate 5.0000
+| end of split  45 / 28 | epoch  12 | time: 983.22s | valid loss 1.0200 | valid ppl 2.7732 | learning rate 5.0000
+| end of split  46 / 28 | epoch  12 | time: 3353.47s | valid loss 1.0200 | valid ppl 2.7731 | learning rate 5.0000
+| end of split  47 / 28 | epoch  12 | time: 3353.04s | valid loss 1.0199 | valid ppl 2.7730 | learning rate 5.0000
+| end of split  48 / 28 | epoch  12 | time: 3362.69s | valid loss 1.0200 | valid ppl 2.7731 | learning rate 5.0000
+| end of split  49 / 28 | epoch  12 | time: 3392.80s | valid loss 1.0198 | valid ppl 2.7726 | learning rate 5.0000
+| end of split  50 / 28 | epoch  12 | time: 3394.63s | valid loss 1.0198 | valid ppl 2.7727 | learning rate 5.0000
+| end of split  51 / 28 | epoch  12 | time: 3382.77s | valid loss 1.0199 | valid ppl 2.7728 | learning rate 5.0000
+| end of split  52 / 28 | epoch  12 | time: 3385.26s | valid loss 1.0199 | valid ppl 2.7729 | learning rate 5.0000
+| end of split  53 / 28 | epoch  12 | time: 3384.68s | valid loss 1.0198 | valid ppl 2.7725 | learning rate 5.0000
+| end of split  54 / 28 | epoch  12 | time: 3381.93s | valid loss 1.0198 | valid ppl 2.7726 | learning rate 5.0000
+| end of split  55 / 28 | epoch  12 | time: 3398.40s | valid loss 1.0197 | valid ppl 2.7723 | learning rate 5.0000
+| end of split  56 / 28 | epoch  12 | time: 3396.09s | valid loss 1.0198 | valid ppl 2.7726 | learning rate 5.0000
+| end of split  29 / 28 | epoch  13 | time: 3377.87s | valid loss 1.0197 | valid ppl 2.7723 | learning rate 5.0000
+| end of split  30 / 28 | epoch  13 | time: 3374.68s | valid loss 1.0196 | valid ppl 2.7721 | learning rate 5.0000
+| end of split  31 / 28 | epoch  13 | time: 3387.69s | valid loss 1.0196 | valid ppl 2.7722 | learning rate 5.0000
+| end of split  32 / 28 | epoch  13 | time: 990.82s | valid loss 1.0197 | valid ppl 2.7723 | learning rate 5.0000
+| end of split  33 / 28 | epoch  13 | time: 3369.69s | valid loss 1.0195 | valid ppl 2.7719 | learning rate 5.0000
+| end of split  34 / 28 | epoch  13 | time: 3370.78s | valid loss 1.0197 | valid ppl 2.7723 | learning rate 5.0000
+| end of split  35 / 28 | epoch  13 | time: 3370.56s | valid loss 1.0195 | valid ppl 2.7719 | learning rate 5.0000
+| end of split  36 / 28 | epoch  13 | time: 3360.93s | valid loss 1.0195 | valid ppl 2.7719 | learning rate 5.0000
+| end of split  37 / 28 | epoch  13 | time: 3361.03s | valid loss 1.0196 | valid ppl 2.7720 | learning rate 5.0000
+| end of split  38 / 28 | epoch  13 | time: 3361.72s | valid loss 1.0196 | valid ppl 2.7722 | learning rate 5.0000
+| end of split  39 / 28 | epoch  13 | time: 3350.87s | valid loss 1.0195 | valid ppl 2.7718 | learning rate 5.0000
+| end of split  40 / 28 | epoch  13 | time: 3347.90s | valid loss 1.0195 | valid ppl 2.7718 | learning rate 5.0000
+| end of split  41 / 28 | epoch  13 | time: 3345.82s | valid loss 1.0197 | valid ppl 2.7722 | learning rate 5.0000
+| end of split  42 / 28 | epoch  13 | time: 3354.18s | valid loss 1.0194 | valid ppl 2.7716 | learning rate 5.0000
+| end of split  43 / 28 | epoch  13 | time: 3350.06s | valid loss 1.0203 | valid ppl 2.7741 | learning rate 5.0000
+| end of split  44 / 28 | epoch  13 | time: 3348.70s | valid loss 1.0194 | valid ppl 2.7716 | learning rate 5.0000
+| end of split  45 / 28 | epoch  13 | time: 3351.28s | valid loss 1.0194 | valid ppl 2.7714 | learning rate 5.0000
+| end of split  46 / 28 | epoch  13 | time: 3347.01s | valid loss 1.0194 | valid ppl 2.7714 | learning rate 5.0000
+| end of split  47 / 28 | epoch  13 | time: 3338.57s | valid loss 1.0193 | valid ppl 2.7713 | learning rate 5.0000