Skip to content

Releases: chartes/deucalion-model-af

1.0.0

28 Apr 10:06
3b4b395
Compare
Choose a tag to compare

Trained and tested on https://github.com/chartes/OF3C/releases/tag/1.1

CAS

accuracy precision recall f1 balanced-accuracy support
all 0.9395 0.8571 0.8943 0.8723 0.8943 34136
known-tokens 0.9469 0.8628 0.8996 0.8779 0.8996 31234
unknown-tokens 0.8604 0.8641 0.8532 0.8568 0.8532 2902
ambiguous-tokens 0.9135 0.8542 0.8923 0.8699 0.8923 16788

DEGRE

accuracy precision recall f1 balanced-accuracy support
all 0.9766 0.7993 0.859 0.7884 0.859 34136
known-tokens 0.9806 0.8068 0.8703 0.7979 0.8703 31234
unknown-tokens 0.9328 0.8119 0.6827 0.7365 0.6827 2902
ambiguous-tokens 0.934 0.8018 0.8581 0.7918 0.8581 7423

GENRE

accuracy precision recall f1 balanced-accuracy support
all 0.9492 0.9086 0.8803 0.8933 0.8803 34136
known-tokens 0.9581 0.9177 0.8884 0.9019 0.8884 31234
unknown-tokens 0.8542 0.7856 0.7174 0.7383 0.7174 2902
ambiguous-tokens 0.9243 0.8964 0.8666 0.8803 0.8666 14363

lemma

accuracy precision recall f1 balanced-accuracy support
all 0.9365 0.5843 0.579 0.5749 0.579 34136
known-tokens 0.9624 0.8166 0.8123 0.8091 0.8123 31735
unknown-tokens 0.5939 0.3662 0.3656 0.3611 0.3656 2401
ambiguous-tokens 0.9506 0.6919 0.6919 0.6827 0.6919 19398
unknown-targets 0.0599 0.0289 0.0277 0.0279 0.0277 551

MODE

accuracy precision recall f1 balanced-accuracy support
all 0.9881 0.9323 0.8804 0.902 0.8804 34136
known-tokens 0.9924 0.9484 0.9107 0.9265 0.9107 31234
unknown-tokens 0.9418 0.886 0.7897 0.8264 0.7897 2902
ambiguous-tokens 0.9775 0.9285 0.9131 0.9175 0.9131 6402

NOMB

accuracy precision recall f1 balanced-accuracy support
all 0.9643 0.9551 0.9491 0.952 0.9491 34136
known-tokens 0.9703 0.9617 0.9565 0.959 0.9565 31234
unknown-tokens 0.8997 0.8628 0.8371 0.8494 0.8371 2902
ambiguous-tokens 0.9495 0.9423 0.9383 0.9402 0.9383 15703

PERS

accuracy precision recall f1 balanced-accuracy support
all 0.9847 0.7771 0.7621 0.7694 0.7621 34136
known-tokens 0.988 0.7808 0.7707 0.7756 0.7707 31234
unknown-tokens 0.9493 0.9275 0.8536 0.8873 0.8536 2902
ambiguous-tokens 0.9763 0.7792 0.768 0.7735 0.768 11471

POS

accuracy precision recall f1 balanced-accuracy support
all 0.9482 0.7319 0.732 0.7265 0.732 34136
known-tokens 0.9547 0.7536 0.7616 0.7507 0.7616 31735
unknown-tokens 0.8621 0.5939 0.5594 0.5536 0.5594 2401
ambiguous-tokens 0.9404 0.7428 0.7261 0.7234 0.7261 21497

TEMPS

accuracy precision recall f1 balanced-accuracy support
all 0.9844 0.944 0.9284 0.9355 0.9284 34136
known-tokens 0.9891 0.9572 0.9484 0.9525 0.9484 31234
unknown-tokens 0.9338 0.8696 0.8358 0.8489 0.8358 2902
ambiguous-tokens 0.9631 0.9301 0.9196 0.924 0.9196 6530

0.4.1

05 Dec 16:50
3b4b395
Compare
Choose a tag to compare

Models for lemma and POS are updated.

lemma

accuracy precision recall f1 balanced-accuracy support
all 0.9901 0.9106 0.9119 0.9091 0.9441 87157
known-tokens 0.9919 0.9493 0.9501 0.9482 0.967 86694
unknown-tokens 0.6544 0.4584 0.4599 0.4565 0.622 463
ambiguous-tokens 0.9879 0.8853 0.8866 0.882 0.9195 52598
unknown-targets 0.0612 0.0309 0.0309 0.0309 0.0595 98

POS

accuracy precision recall f1 balanced-accuracy support
all 0.9854 0.8603 0.8617 0.8597 0.8617 87157
known-tokens 0.9862 0.8612 0.8625 0.8606 0.8625 86694
unknown-tokens 0.8402 0.6225 0.6539 0.6284 0.6924 463
ambiguous-tokens 0.9805 0.8648 0.8641 0.8628 0.8641 58082

0.4.0Alpha

04 Jun 08:15
3b4b395
Compare
Choose a tag to compare
0.4.0Alpha Pre-release
Pre-release
Task All - acc pre rec Known tokens - acc pre rec Unknown tokens - acc pre rec Ambiguous tokens - acc pre rec
CAS 95.25 91.75 92.26 95.74 92.18 92.74 84.63 62.6 62.55 94.01 91.77 92.33
DEGRE 98.47 86.45 82.77 98.74 87.75 84.98 92.55 63.81 44.48 95.46 86.91 85.8
GENRE 96.21 93.22 89.03 96.76 93.67 89.66 84.47 88.01 69.87 94.59 92.11 87.9
MODE 99.05 91.59 89 99.36 93.56 92.06 92.39 79.46 74.89 97.61 88.89 92.14
NOMB 97.05 96.5 96.2 97.47 96.92 96.65 88.2 87.85 82.2 96.06 95.67 95.36
PERS 98.79 91.25 85.28 98.95 91.73 85.73 95.34 90.17 90.12 97.72 91.31 85.12
POS 97.55 83 79.9 97.88 84.81 82.11 84.56 54.19 55.92 97.37 83.7 81.2
lemma 97.66 76.08 75.86 98.38 91.72 91.71 69.01 48.86 48.72 97.79 78.24 78.11

0.3.0

14 Dec 09:25
Compare
Choose a tag to compare
Release 0.3.0

0.2.0

03 Jun 10:02
Compare
Choose a tag to compare
v0.2.0

Using pie-flask:v0.0.2 for Docker faster builds

v0.1.0

14 Jan 09:30
Compare
Choose a tag to compare

Cite using DOI

Initial release