I have used prodigy to train a model for customized entity, It is pretty good. and the result is so
Loaded model en_core_web_sm
Using 20% of accept/reject examples (1330) for evaluation
Using 100% of remaining examples (5332) for training
Dropout: 0.2 Batch size: 32 Iterations: 50
BEFORE 0.005
Correct 23
Incorrect 4442
Entities 2779
Unknown 0
# LOSS RIGHT WRONG ENTS SKIP ACCURACY
01 30.157 1484 435 1694 0 0.773
02 25.616 1575 306 1747 0 0.837
03 25.160 1608 262 1769 0 0.860
04 24.729 1626 237 1780 0 0.873
05 24.756 1622 220 1755 0 0.881
06 24.248 1632 204 1759 0 0.889
07 24.315 1637 178 1743 0 0.902
08 24.172 1647 162 1747 0 0.910
09 24.171 1645 150 1731 0 0.916
10 24.011 1639 161 1730 0 0.911
11 23.908 1638 156 1723 0 0.913
12 24.130 1648 155 1742 0 0.914
13 23.913 1643 157 1734 0 0.913
14 23.997 1647 163 1748 0 0.910
15 23.901 1646 160 1743 0 0.911
16 23.961 1642 158 1733 0 0.912
17 23.912 1650 154 1745 0 0.915
18 23.871 1650 147 1738 0 0.918
19 23.949 1645 159 1740 0 0.912
20 23.929 1650 153 1744 0 0.915
21 23.891 1643 160 1737 0 0.911
22 23.916 1649 152 1741 0 0.916
23 23.822 1651 157 1750 0 0.913
24 23.806 1654 144 1743 0 0.920
25 23.910 1652 151 1746 0 0.916
26 23.780 1648 151 1738 0 0.916
27 23.841 1653 147 1744 0 0.918
28 23.864 1643 152 1729 0 0.915
29 23.975 1648 154 1741 0 0.915
30 23.691 1646 147 1730 0 0.918
31 23.959 1649 143 1732 0 0.920
32 23.810 1647 143 1728 0 0.920
33 23.745 1648 143 1730 0 0.920
34 24.077 1650 143 1734 0 0.920
35 23.737 1647 150 1735 0 0.917
36 23.815 1648 146 1733 0 0.919
37 23.872 1653 138 1735 0 0.923
38 23.936 1652 146 1741 0 0.919
39 23.890 1650 150 1741 0 0.917
40 23.929 1646 151 1734 0 0.916
41 23.890 1643 149 1726 0 0.917
42 24.018 1645 152 1733 0 0.915
43 23.718 1639 155 1724 0 0.914
44 23.798 1643 150 1727 0 0.916
45 23.834 1644 152 1731 0 0.915
46 23.917 1645 158 1739 0 0.912
47 23.775 1639 159 1728 0 0.912
48 23.805 1634 164 1723 0 0.909
49 23.998 1641 159 1732 0 0.912
50 23.814 1645 155 1736 0 0.914
Correct 1653
Incorrect 138
Baseline 0.005
Accuracy 0.923
Model: C:\Users\moha\Documents\Prodigy\Last_version_data\model_merged_all_02_03
Training data: C:\Users\moha\Documents\Prodigy\Last_version_data\model_merged_all_02_03\training.jsonl
Evaluation data: C:\Users\moha\Documents\Prodigy\Last_version_data\model_merged_all_02_03\evaluation.jsonl
I am wondering how can I have access to deep learning model in order to draw los and accuracy in python?
Many many thanks