Hugging Face's logo

werner1hugo
/
lid-framework-models

Text Classification
fastText
language-identification
short-text
scikit-learn
model-selection
Model card Files
xet
 Community
werner1hugo commited on
Commit
9743d10
·
verified ·
1 Parent(s): b7ba0db

Add files using upload-large-folder tool

Browse files
Files changed (50) hide show
  1. fasttext/amazon_reviews_multi/fasttext_subword/classification_report.txt +30 -0
  2. fasttext/amazon_reviews_multi/fasttext_subword/param_search_results.csv +10 -0
  3. fasttext/amazon_reviews_multi/fasttext_subword/run_metadata.json +24 -0
  4. fasttext/amazon_reviews_multi/fasttext_word/classification_report.txt +30 -0
  5. fasttext/amazon_reviews_multi/fasttext_word/param_search_results.csv +10 -0
  6. fasttext/amazon_reviews_multi/fasttext_word/run_metadata.json +24 -0
  7. fasttext/europarl/fasttext_subword/classification_report.txt +30 -0
  8. fasttext/europarl/fasttext_subword/param_search_results.csv +10 -0
  9. fasttext/europarl/fasttext_subword/run_metadata.json +24 -0
  10. fasttext/europarl/fasttext_word/classification_report.txt +30 -0
  11. fasttext/europarl/fasttext_word/param_search_results.csv +10 -0
  12. fasttext/europarl/fasttext_word/run_metadata.json +24 -0
  13. fasttext/exorde-social-media-december-2024-week1/fasttext_subword/classification_report.txt +30 -0
  14. fasttext/exorde-social-media-december-2024-week1/fasttext_subword/param_search_results.csv +10 -0
  15. fasttext/exorde-social-media-december-2024-week1/fasttext_subword/run_metadata.json +24 -0
  16. fasttext/exorde-social-media-december-2024-week1/fasttext_word/classification_report.txt +30 -0
  17. fasttext/exorde-social-media-december-2024-week1/fasttext_word/param_search_results.csv +10 -0
  18. fasttext/exorde-social-media-december-2024-week1/fasttext_word/run_metadata.json +24 -0
  19. fasttext/flores_plus/fasttext_subword/classification_report.txt +30 -0
  20. fasttext/flores_plus/fasttext_subword/param_search_results.csv +10 -0
  21. fasttext/flores_plus/fasttext_subword/run_metadata.json +24 -0
  22. fasttext/flores_plus/fasttext_word/classification_report.txt +30 -0
  23. fasttext/flores_plus/fasttext_word/param_search_results.csv +10 -0
  24. fasttext/flores_plus/fasttext_word/run_metadata.json +24 -0
  25. fasttext/language-identification/fasttext_subword/classification_report.txt +30 -0
  26. fasttext/language-identification/fasttext_subword/param_search_results.csv +10 -0
  27. fasttext/language-identification/fasttext_subword/run_metadata.json +24 -0
  28. fasttext/language-identification/fasttext_word/classification_report.txt +30 -0
  29. fasttext/language-identification/fasttext_word/param_search_results.csv +10 -0
  30. fasttext/language-identification/fasttext_word/run_metadata.json +24 -0
  31. fasttext/massive/fasttext_subword/classification_report.txt +30 -0
  32. fasttext/massive/fasttext_subword/param_search_results.csv +10 -0
  33. fasttext/massive/fasttext_subword/run_metadata.json +24 -0
  34. fasttext/massive/fasttext_word/classification_report.txt +30 -0
  35. fasttext/massive/fasttext_word/param_search_results.csv +10 -0
  36. fasttext/massive/fasttext_word/run_metadata.json +24 -0
  37. fasttext/mmarco/fasttext_subword/classification_report.txt +30 -0
  38. fasttext/mmarco/fasttext_subword/param_search_results.csv +10 -0
  39. fasttext/mmarco/fasttext_subword/run_metadata.json +24 -0
  40. fasttext/mmarco/fasttext_word/classification_report.txt +30 -0
  41. fasttext/mmarco/fasttext_word/param_search_results.csv +10 -0
  42. fasttext/mmarco/fasttext_word/run_metadata.json +24 -0
  43. fasttext/multilingual_cc_news/fasttext_subword/classification_report.txt +30 -0
  44. fasttext/multilingual_cc_news/fasttext_subword/param_search_results.csv +10 -0
  45. fasttext/multilingual_cc_news/fasttext_subword/run_metadata.json +24 -0
  46. fasttext/multilingual_cc_news/fasttext_word/classification_report.txt +30 -0
  47. fasttext/multilingual_cc_news/fasttext_word/param_search_results.csv +10 -0
  48. fasttext/multilingual_cc_news/fasttext_word/run_metadata.json +24 -0
  49. fasttext/multilingual_toxicity_dataset/fasttext_subword/classification_report.txt +30 -0
  50. fasttext/multilingual_toxicity_dataset/fasttext_subword/param_search_results.csv +10 -0
fasttext/amazon_reviews_multi/fasttext_subword/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.00 0.00 0.00 0
4
+ da 0.00 0.00 0.00 0
5
+ de 1.00 1.00 1.00 20022
6
+ el 0.00 0.00 0.00 0
7
+ en 1.00 1.00 1.00 19795
8
+ es 1.00 1.00 1.00 19841
9
+ fi 0.00 0.00 0.00 0
10
+ fr 1.00 1.00 1.00 19900
11
+ hr 0.00 0.00 0.00 0
12
+ it 0.00 0.00 0.00 0
13
+ ja 1.00 1.00 1.00 19990
14
+ ko 0.00 0.00 0.00 0
15
+ lt 0.00 0.00 0.00 0
16
+ mk 0.00 0.00 0.00 0
17
+ nb 0.00 0.00 0.00 0
18
+ nl 0.00 0.00 0.00 0
19
+ pl 0.00 0.00 0.00 0
20
+ pt 0.00 0.00 0.00 0
21
+ ro 0.00 0.00 0.00 0
22
+ ru 0.00 0.00 0.00 0
23
+ sl 0.00 0.00 0.00 0
24
+ sv 0.00 0.00 0.00 0
25
+ uk 0.00 0.00 0.00 0
26
+ zh 1.00 1.00 1.00 19993
27
+
28
+ accuracy 1.00 119541
29
+ macro avg 0.25 0.25 0.25 119541
30
+ weighted avg 1.00 1.00 1.00 119541
fasttext/amazon_reviews_multi/fasttext_subword/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.999603,11.32
3
+ 5,0.1,0.999665,11.4
4
+ 5,0.5,0.999833,11.29
5
+ 10,0.05,0.999707,19.9
6
+ 10,0.1,0.999791,19.93
7
+ 10,0.5,0.999833,19.65
8
+ 25,0.05,0.999833,44.87
9
+ 25,0.1,0.999812,45.65
10
+ 25,0.5,0.999854,45.07
fasttext/amazon_reviews_multi/fasttext_subword/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_subword",
3
+ "tags": {
4
+ "feature_type": "subword",
5
+ "wordNgrams": "2",
6
+ "minn": "2",
7
+ "maxn": "4",
8
+ "loss": "softmax",
9
+ "dataset_version": "amazon_reviews_multi"
10
+ },
11
+ "params": {
12
+ "epoch": "25",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.999832693385533,
17
+ "f1_macro": 0.9998326634729101,
18
+ "precision_macro": 0.9998323692172443,
19
+ "recall_macro": 0.9998329695159586,
20
+ "inference_time_total_s": 2.9018,
21
+ "inference_time_ms_per_sample": 0.024275,
22
+ "throughput_samples_per_sec": 41195.4
23
+ }
24
+ }
fasttext/amazon_reviews_multi/fasttext_word/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.00 0.00 0.00 0
4
+ da 0.00 0.00 0.00 0
5
+ de 1.00 1.00 1.00 20022
6
+ el 0.00 0.00 0.00 0
7
+ en 1.00 1.00 1.00 19795
8
+ es 1.00 1.00 1.00 19841
9
+ fi 0.00 0.00 0.00 0
10
+ fr 1.00 1.00 1.00 19900
11
+ hr 0.00 0.00 0.00 0
12
+ it 0.00 0.00 0.00 0
13
+ ja 0.94 0.06 0.11 19990
14
+ ko 0.00 0.00 0.00 0
15
+ lt 0.00 0.00 0.00 0
16
+ mk 0.00 0.00 0.00 0
17
+ nb 0.00 0.00 0.00 0
18
+ nl 0.00 0.00 0.00 0
19
+ pl 0.00 0.00 0.00 0
20
+ pt 0.00 0.00 0.00 0
21
+ ro 0.00 0.00 0.00 0
22
+ ru 0.00 0.00 0.00 0
23
+ sl 0.00 0.00 0.00 0
24
+ sv 0.00 0.00 0.00 0
25
+ uk 0.00 0.00 0.00 0
26
+ zh 0.51 1.00 0.68 19993
27
+
28
+ accuracy 0.84 119541
29
+ macro avg 0.23 0.21 0.20 119541
30
+ weighted avg 0.91 0.84 0.80 119541
fasttext/amazon_reviews_multi/fasttext_word/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.797777,4.5
3
+ 5,0.1,0.797788,4.55
4
+ 5,0.5,0.797734,4.81
5
+ 10,0.05,0.797729,7.3
6
+ 10,0.1,0.797818,7.42
7
+ 10,0.5,0.797837,7.14
8
+ 25,0.05,0.797636,18.96
9
+ 25,0.1,0.797611,17.49
10
+ 25,0.5,0.79789,20.36
fasttext/amazon_reviews_multi/fasttext_word/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_word",
3
+ "tags": {
4
+ "feature_type": "word_ngram",
5
+ "wordNgrams": "1",
6
+ "minn": "0",
7
+ "maxn": "0",
8
+ "loss": "softmax",
9
+ "dataset_version": "amazon_reviews_multi"
10
+ },
11
+ "params": {
12
+ "epoch": "25",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.8411674655557507,
17
+ "f1_macro": 0.7968306004142348,
18
+ "precision_macro": 0.9092107304710257,
19
+ "recall_macro": 0.8416939311995351,
20
+ "inference_time_total_s": 0.8704,
21
+ "inference_time_ms_per_sample": 0.007281,
22
+ "throughput_samples_per_sec": 137346.04
23
+ }
24
+ }
fasttext/europarl/fasttext_subword/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.00 0.00 0.00 0
4
+ da 1.00 1.00 1.00 19996
5
+ de 1.00 1.00 1.00 20016
6
+ el 1.00 1.00 1.00 19956
7
+ en 1.00 1.00 1.00 20082
8
+ es 1.00 1.00 1.00 19992
9
+ fi 1.00 1.00 1.00 19935
10
+ fr 1.00 1.00 1.00 19942
11
+ hr 0.00 0.00 0.00 0
12
+ it 1.00 1.00 1.00 20036
13
+ ja 0.00 0.00 0.00 0
14
+ ko 0.00 0.00 0.00 0
15
+ lt 0.00 0.00 0.00 0
16
+ mk 0.00 0.00 0.00 0
17
+ nb 0.00 0.00 0.00 0
18
+ nl 1.00 1.00 1.00 20086
19
+ pl 1.00 1.00 1.00 20006
20
+ pt 1.00 1.00 1.00 19998
21
+ ro 1.00 1.00 1.00 19916
22
+ ru 0.00 0.00 0.00 0
23
+ sl 0.00 0.00 0.00 0
24
+ sv 1.00 1.00 1.00 19996
25
+ uk 0.00 0.00 0.00 0
26
+ zh 0.00 0.00 0.00 0
27
+
28
+ accuracy 1.00 259957
29
+ macro avg 0.54 0.54 0.54 259957
30
+ weighted avg 1.00 1.00 1.00 259957
fasttext/europarl/fasttext_subword/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.999106,28.6
3
+ 5,0.1,0.999317,28.82
4
+ 5,0.5,0.999471,28.87
5
+ 10,0.05,0.999298,54.14
6
+ 10,0.1,0.999375,54.15
7
+ 10,0.5,0.9995,54.14
8
+ 25,0.05,0.999385,129.2
9
+ 25,0.1,0.999462,128.75
10
+ 25,0.5,0.999519,129.88
fasttext/europarl/fasttext_subword/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_subword",
3
+ "tags": {
4
+ "feature_type": "subword",
5
+ "wordNgrams": "2",
6
+ "minn": "2",
7
+ "maxn": "4",
8
+ "loss": "softmax",
9
+ "dataset_version": "europarl"
10
+ },
11
+ "params": {
12
+ "epoch": "25",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.9993729732224945,
17
+ "f1_macro": 0.9993729867417601,
18
+ "precision_macro": 0.9993731983961379,
19
+ "recall_macro": 0.9993728972464199,
20
+ "inference_time_total_s": 6.7203,
21
+ "inference_time_ms_per_sample": 0.025852,
22
+ "throughput_samples_per_sec": 38682.27
23
+ }
24
+ }
fasttext/europarl/fasttext_word/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.00 0.00 0.00 0
4
+ da 1.00 1.00 1.00 19996
5
+ de 1.00 1.00 1.00 20016
6
+ el 1.00 1.00 1.00 19956
7
+ en 1.00 1.00 1.00 20082
8
+ es 1.00 1.00 1.00 19992
9
+ fi 0.99 1.00 1.00 19935
10
+ fr 1.00 1.00 1.00 19942
11
+ hr 0.00 0.00 0.00 0
12
+ it 1.00 1.00 1.00 20036
13
+ ja 0.00 0.00 0.00 0
14
+ ko 0.00 0.00 0.00 0
15
+ lt 0.00 0.00 0.00 0
16
+ mk 0.00 0.00 0.00 0
17
+ nb 0.00 0.00 0.00 0
18
+ nl 1.00 1.00 1.00 20086
19
+ pl 1.00 1.00 1.00 20006
20
+ pt 1.00 1.00 1.00 19998
21
+ ro 1.00 1.00 1.00 19916
22
+ ru 0.00 0.00 0.00 0
23
+ sl 0.00 0.00 0.00 0
24
+ sv 1.00 1.00 1.00 19996
25
+ uk 0.00 0.00 0.00 0
26
+ zh 0.00 0.00 0.00 0
27
+
28
+ accuracy 1.00 259957
29
+ macro avg 0.54 0.54 0.54 259957
30
+ weighted avg 1.00 1.00 1.00 259957
fasttext/europarl/fasttext_word/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.998269,14.95
3
+ 5,0.1,0.998519,15.2
4
+ 5,0.5,0.998721,15.08
5
+ 10,0.05,0.998451,26.6
6
+ 10,0.1,0.998557,28.96
7
+ 10,0.5,0.998576,25.75
8
+ 25,0.05,0.998586,62.89
9
+ 25,0.1,0.998702,65.83
10
+ 25,0.5,0.998759,63.87
fasttext/europarl/fasttext_word/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_word",
3
+ "tags": {
4
+ "feature_type": "word_ngram",
5
+ "wordNgrams": "1",
6
+ "minn": "0",
7
+ "maxn": "0",
8
+ "loss": "softmax",
9
+ "dataset_version": "europarl"
10
+ },
11
+ "params": {
12
+ "epoch": "25",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.9985997684232393,
17
+ "f1_macro": 0.9985992859536525,
18
+ "precision_macro": 0.9986004248910461,
19
+ "recall_macro": 0.9985994130953136,
20
+ "inference_time_total_s": 2.4701,
21
+ "inference_time_ms_per_sample": 0.009502,
22
+ "throughput_samples_per_sec": 105243.52
23
+ }
24
+ }
fasttext/exorde-social-media-december-2024-week1/fasttext_subword/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.96 0.94 0.95 3501
4
+ da 0.91 0.89 0.90 3175
5
+ de 0.97 0.97 0.97 19988
6
+ el 1.00 0.99 0.99 2671
7
+ en 0.94 0.96 0.95 19955
8
+ es 0.98 0.98 0.98 20014
9
+ fi 0.91 0.93 0.92 4341
10
+ fr 0.99 0.99 0.99 19964
11
+ hr 0.94 0.93 0.93 1723
12
+ it 0.97 0.97 0.97 19940
13
+ ja 0.97 0.97 0.97 19926
14
+ ko 0.98 0.97 0.97 9596
15
+ lt 0.96 0.90 0.93 385
16
+ mk 0.99 0.94 0.97 369
17
+ nb 0.85 0.82 0.84 1657
18
+ nl 0.98 0.97 0.98 10778
19
+ pl 0.97 0.97 0.97 7536
20
+ pt 0.98 0.99 0.99 19948
21
+ ro 0.97 0.96 0.97 2756
22
+ ru 0.98 0.98 0.98 19824
23
+ sl 0.84 0.77 0.80 713
24
+ sv 0.97 0.95 0.96 6856
25
+ uk 0.97 0.93 0.95 4517
26
+ zh 0.95 0.96 0.96 19968
27
+
28
+ accuracy 0.97 240101
29
+ macro avg 0.96 0.94 0.95 240101
30
+ weighted avg 0.97 0.97 0.97 240101
fasttext/exorde-social-media-december-2024-week1/fasttext_subword/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.905079,28.26
3
+ 5,0.1,0.94221,28.16
4
+ 5,0.5,0.949372,28.39
5
+ 10,0.05,0.942726,49.96
6
+ 10,0.1,0.948625,50.42
7
+ 10,0.5,0.949611,49.85
8
+ 25,0.05,0.949354,113.56
9
+ 25,0.1,0.948747,113.67
10
+ 25,0.5,0.949866,113.26
fasttext/exorde-social-media-december-2024-week1/fasttext_subword/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_subword",
3
+ "tags": {
4
+ "feature_type": "subword",
5
+ "wordNgrams": "2",
6
+ "minn": "2",
7
+ "maxn": "4",
8
+ "loss": "softmax",
9
+ "dataset_version": "exorde-social-media-december-2024-week1"
10
+ },
11
+ "params": {
12
+ "epoch": "25",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.9682342014402272,
17
+ "f1_macro": 0.9491932783934312,
18
+ "precision_macro": 0.9559232010596661,
19
+ "recall_macro": 0.9428664563019996,
20
+ "inference_time_total_s": 7.319,
21
+ "inference_time_ms_per_sample": 0.030483,
22
+ "throughput_samples_per_sec": 32805.16
23
+ }
24
+ }
fasttext/exorde-social-media-december-2024-week1/fasttext_word/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.95 0.92 0.93 3501
4
+ da 0.89 0.87 0.88 3175
5
+ de 0.96 0.95 0.96 19988
6
+ el 1.00 0.97 0.98 2671
7
+ en 0.92 0.95 0.94 19955
8
+ es 0.97 0.97 0.97 20014
9
+ fi 0.88 0.82 0.85 4341
10
+ fr 0.99 0.99 0.99 19964
11
+ hr 0.93 0.90 0.92 1723
12
+ it 0.97 0.96 0.96 19940
13
+ ja 0.56 0.95 0.70 19926
14
+ ko 0.97 0.86 0.91 9596
15
+ lt 0.95 0.82 0.88 385
16
+ mk 0.97 0.91 0.94 369
17
+ nb 0.84 0.76 0.80 1657
18
+ nl 0.98 0.96 0.97 10778
19
+ pl 0.96 0.95 0.95 7536
20
+ pt 0.98 0.98 0.98 19948
21
+ ro 0.96 0.95 0.96 2756
22
+ ru 0.97 0.96 0.97 19824
23
+ sl 0.82 0.71 0.76 713
24
+ sv 0.96 0.95 0.95 6856
25
+ uk 0.97 0.86 0.91 4517
26
+ zh 0.81 0.37 0.51 19968
27
+
28
+ accuracy 0.90 240101
29
+ macro avg 0.92 0.89 0.90 240101
30
+ weighted avg 0.92 0.90 0.90 240101
fasttext/exorde-social-media-december-2024-week1/fasttext_word/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.894855,20.1
3
+ 5,0.1,0.898782,19.43
4
+ 5,0.5,0.899593,19.6
5
+ 10,0.05,0.898107,35.54
6
+ 10,0.1,0.899117,36.32
7
+ 10,0.5,0.899015,36.91
8
+ 25,0.05,0.89845,85.09
9
+ 25,0.1,0.899848,82.77
10
+ 25,0.5,0.898883,87.97
fasttext/exorde-social-media-december-2024-week1/fasttext_word/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_word",
3
+ "tags": {
4
+ "feature_type": "word_ngram",
5
+ "wordNgrams": "1",
6
+ "minn": "0",
7
+ "maxn": "0",
8
+ "loss": "softmax",
9
+ "dataset_version": "exorde-social-media-december-2024-week1"
10
+ },
11
+ "params": {
12
+ "epoch": "25",
13
+ "lr": "0.1"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.8991632687910505,
17
+ "f1_macro": 0.8985305506686693,
18
+ "precision_macro": 0.9231642533621339,
19
+ "recall_macro": 0.8866139538959009,
20
+ "inference_time_total_s": 2.8393,
21
+ "inference_time_ms_per_sample": 0.011825,
22
+ "throughput_samples_per_sec": 84564.44
23
+ }
24
+ }
fasttext/flores_plus/fasttext_subword/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.99 0.97 0.98 80
4
+ da 0.90 0.84 0.87 94
5
+ de 0.98 0.98 0.98 61
6
+ el 1.00 1.00 1.00 62
7
+ en 1.00 1.00 1.00 94
8
+ es 0.98 0.97 0.97 58
9
+ fi 0.99 1.00 0.99 84
10
+ fr 0.96 0.98 0.97 55
11
+ hr 0.97 0.98 0.97 98
12
+ it 1.00 1.00 1.00 55
13
+ ja 0.99 0.99 0.99 199
14
+ ko 1.00 1.00 1.00 205
15
+ lt 0.98 1.00 0.99 91
16
+ mk 1.00 0.98 0.99 81
17
+ nb 0.86 0.92 0.89 103
18
+ nl 0.99 0.97 0.98 72
19
+ pl 1.00 1.00 1.00 83
20
+ pt 0.98 0.99 0.98 80
21
+ ro 1.00 0.99 0.99 69
22
+ ru 0.98 1.00 0.99 83
23
+ sl 0.99 0.96 0.97 100
24
+ sv 0.99 0.99 0.99 98
25
+ uk 1.00 1.00 1.00 87
26
+ zh 0.00 0.00 0.00 0
27
+
28
+ accuracy 0.98 2092
29
+ macro avg 0.94 0.94 0.94 2092
30
+ weighted avg 0.98 0.98 0.98 2092
fasttext/flores_plus/fasttext_subword/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.148222,0.72
3
+ 5,0.1,0.052794,0.73
4
+ 5,0.5,0.839754,0.73
5
+ 10,0.05,0.052794,0.83
6
+ 10,0.1,0.071336,0.83
7
+ 10,0.5,0.961842,0.83
8
+ 25,0.05,0.200145,1.33
9
+ 25,0.1,0.828753,1.23
10
+ 25,0.5,0.974251,1.34
fasttext/flores_plus/fasttext_subword/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_subword",
3
+ "tags": {
4
+ "feature_type": "subword",
5
+ "wordNgrams": "2",
6
+ "minn": "2",
7
+ "maxn": "4",
8
+ "loss": "softmax",
9
+ "dataset_version": "flores_plus"
10
+ },
11
+ "params": {
12
+ "epoch": "25",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.9789674952198852,
17
+ "f1_macro": 0.9787420891328592,
18
+ "precision_macro": 0.9791241929467931,
19
+ "recall_macro": 0.9786400130801987,
20
+ "inference_time_total_s": 0.0755,
21
+ "inference_time_ms_per_sample": 0.036074,
22
+ "throughput_samples_per_sec": 27720.71
23
+ }
24
+ }
fasttext/flores_plus/fasttext_word/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.97 0.94 0.96 80
4
+ da 0.85 0.87 0.86 94
5
+ de 1.00 0.98 0.99 61
6
+ el 1.00 1.00 1.00 62
7
+ en 0.99 0.97 0.98 94
8
+ es 0.95 0.97 0.96 58
9
+ fi 0.99 0.88 0.93 84
10
+ fr 1.00 0.96 0.98 55
11
+ hr 0.92 0.89 0.90 98
12
+ it 0.96 0.98 0.97 55
13
+ ja 0.73 0.99 0.84 199
14
+ ko 1.00 0.84 0.91 205
15
+ lt 1.00 0.92 0.96 91
16
+ mk 1.00 0.98 0.99 81
17
+ nb 0.88 0.82 0.85 103
18
+ nl 0.95 0.99 0.97 72
19
+ pl 0.93 0.90 0.91 83
20
+ pt 0.99 0.99 0.99 80
21
+ ro 0.97 0.93 0.95 69
22
+ ru 0.94 0.89 0.91 83
23
+ sl 0.97 0.93 0.95 100
24
+ sv 0.94 0.96 0.95 98
25
+ uk 0.97 0.95 0.96 87
26
+ zh 0.00 0.00 0.00 0
27
+
28
+ accuracy 0.93 2092
29
+ macro avg 0.91 0.90 0.90 2092
30
+ weighted avg 0.94 0.93 0.93 2092
fasttext/flores_plus/fasttext_word/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.033162,0.21
3
+ 5,0.1,0.255358,0.31
4
+ 5,0.5,0.952137,0.21
5
+ 10,0.05,0.26373,0.41
6
+ 10,0.1,0.886777,0.31
7
+ 10,0.5,0.96238,0.31
8
+ 25,0.05,0.920455,0.71
9
+ 25,0.1,0.954555,0.61
10
+ 25,0.5,0.960734,0.61
fasttext/flores_plus/fasttext_word/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_word",
3
+ "tags": {
4
+ "feature_type": "word_ngram",
5
+ "wordNgrams": "1",
6
+ "minn": "0",
7
+ "maxn": "0",
8
+ "loss": "softmax",
9
+ "dataset_version": "flores_plus"
10
+ },
11
+ "params": {
12
+ "epoch": "10",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.9292543021032504,
17
+ "f1_macro": 0.9418987015368867,
18
+ "precision_macro": 0.9510548514899856,
19
+ "recall_macro": 0.9360289085752775,
20
+ "inference_time_total_s": 0.0252,
21
+ "inference_time_ms_per_sample": 0.012046,
22
+ "throughput_samples_per_sec": 83016.1
23
+ }
24
+ }
fasttext/language-identification/fasttext_subword/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.00 0.00 0.00 0
4
+ da 0.00 0.00 0.00 0
5
+ de 1.00 1.00 1.00 433
6
+ el 1.00 1.00 1.00 418
7
+ en 0.99 1.00 1.00 441
8
+ es 1.00 0.99 0.99 436
9
+ fi 0.00 0.00 0.00 0
10
+ fr 1.00 1.00 1.00 417
11
+ hr 0.00 0.00 0.00 0
12
+ it 0.99 0.99 0.99 428
13
+ ja 1.00 1.00 1.00 449
14
+ ko 0.00 0.00 0.00 0
15
+ lt 0.00 0.00 0.00 0
16
+ mk 0.00 0.00 0.00 0
17
+ nb 0.00 0.00 0.00 0
18
+ nl 0.99 0.99 0.99 399
19
+ pl 1.00 0.98 0.99 392
20
+ pt 0.98 0.99 0.99 423
21
+ ro 0.00 0.00 0.00 0
22
+ ru 1.00 1.00 1.00 425
23
+ sl 0.00 0.00 0.00 0
24
+ sv 0.00 0.00 0.00 0
25
+ uk 0.00 0.00 0.00 0
26
+ zh 0.99 1.00 1.00 452
27
+
28
+ accuracy 1.00 5113
29
+ macro avg 0.50 0.50 0.50 5113
30
+ weighted avg 1.00 1.00 1.00 5113
fasttext/language-identification/fasttext_subword/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.356591,0.88
3
+ 5,0.1,0.875076,0.9
4
+ 5,0.5,0.997567,0.89
5
+ 10,0.05,0.876053,1.19
6
+ 10,0.1,0.996574,1.19
7
+ 10,0.5,0.997567,1.19
8
+ 25,0.05,0.996574,2.19
9
+ 25,0.1,0.997567,2.1
10
+ 25,0.5,0.998513,2.19
fasttext/language-identification/fasttext_subword/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_subword",
3
+ "tags": {
4
+ "feature_type": "subword",
5
+ "wordNgrams": "2",
6
+ "minn": "2",
7
+ "maxn": "4",
8
+ "loss": "softmax",
9
+ "dataset_version": "language-identification"
10
+ },
11
+ "params": {
12
+ "epoch": "25",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.9951105026403285,
17
+ "f1_macro": 0.9950397893392807,
18
+ "precision_macro": 0.9951121403974614,
19
+ "recall_macro": 0.9949879324949041,
20
+ "inference_time_total_s": 0.1392,
21
+ "inference_time_ms_per_sample": 0.027218,
22
+ "throughput_samples_per_sec": 36740.42
23
+ }
24
+ }
fasttext/language-identification/fasttext_word/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.00 0.00 0.00 0
4
+ da 0.00 0.00 0.00 0
5
+ de 1.00 0.99 1.00 433
6
+ el 1.00 1.00 1.00 418
7
+ en 1.00 1.00 1.00 441
8
+ es 1.00 0.99 0.99 436
9
+ fi 0.00 0.00 0.00 0
10
+ fr 0.99 1.00 0.99 417
11
+ hr 0.00 0.00 0.00 0
12
+ it 0.99 0.98 0.98 428
13
+ ja 0.55 0.01 0.03 449
14
+ ko 0.00 0.00 0.00 0
15
+ lt 0.00 0.00 0.00 0
16
+ mk 0.00 0.00 0.00 0
17
+ nb 0.00 0.00 0.00 0
18
+ nl 0.98 0.95 0.96 399
19
+ pl 0.98 0.93 0.96 392
20
+ pt 0.98 0.97 0.97 423
21
+ ro 0.00 0.00 0.00 0
22
+ ru 1.00 0.98 0.99 425
23
+ sl 0.00 0.00 0.00 0
24
+ sv 0.00 0.00 0.00 0
25
+ uk 0.00 0.00 0.00 0
26
+ zh 0.48 1.00 0.65 452
27
+
28
+ accuracy 0.90 5113
29
+ macro avg 0.46 0.45 0.44 5113
30
+ weighted avg 0.91 0.90 0.87 5113
fasttext/language-identification/fasttext_word/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.860653,0.31
3
+ 5,0.1,0.871508,0.32
4
+ 5,0.5,0.876016,0.32
5
+ 10,0.05,0.871487,0.52
6
+ 10,0.1,0.872706,0.52
7
+ 10,0.5,0.877418,0.52
8
+ 25,0.05,0.874294,1.02
9
+ 25,0.1,0.875165,1.22
10
+ 25,0.5,0.876685,1.12
fasttext/language-identification/fasttext_word/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_word",
3
+ "tags": {
4
+ "feature_type": "word_ngram",
5
+ "wordNgrams": "1",
6
+ "minn": "0",
7
+ "maxn": "0",
8
+ "loss": "softmax",
9
+ "dataset_version": "language-identification"
10
+ },
11
+ "params": {
12
+ "epoch": "10",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.8951691766086446,
17
+ "f1_macro": 0.8759973255488999,
18
+ "precision_macro": 0.910714604186456,
19
+ "recall_macro": 0.8988870618046478,
20
+ "inference_time_total_s": 0.0438,
21
+ "inference_time_ms_per_sample": 0.008571,
22
+ "throughput_samples_per_sec": 116666.59
23
+ }
24
+ }
fasttext/massive/fasttext_subword/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.00 0.00 0.00 0
4
+ da 0.99 0.99 0.99 957
5
+ de 1.00 1.00 1.00 1142
6
+ el 1.00 1.00 1.00 1141
7
+ en 0.99 1.00 1.00 963
8
+ es 0.99 0.99 0.99 1097
9
+ fi 1.00 1.00 1.00 1044
10
+ fr 1.00 1.00 1.00 1221
11
+ hr 0.00 0.00 0.00 0
12
+ it 1.00 1.00 1.00 1093
13
+ ja 1.00 1.00 1.00 1337
14
+ ko 1.00 1.00 1.00 1324
15
+ lt 0.00 0.00 0.00 0
16
+ mk 0.00 0.00 0.00 0
17
+ nb 0.00 0.00 0.00 0
18
+ nl 1.00 1.00 1.00 1129
19
+ pl 1.00 1.00 1.00 1002
20
+ pt 0.99 0.99 0.99 1070
21
+ ro 1.00 1.00 1.00 1010
22
+ ru 1.00 1.00 1.00 1023
23
+ sl 1.00 1.00 1.00 998
24
+ sv 1.00 0.99 1.00 957
25
+ uk 0.00 0.00 0.00 0
26
+ zh 1.00 1.00 1.00 874
27
+
28
+ accuracy 1.00 19382
29
+ macro avg 0.75 0.75 0.75 19382
30
+ weighted avg 1.00 1.00 1.00 19382
fasttext/massive/fasttext_subword/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.977225,1.74
3
+ 5,0.1,0.991856,1.75
4
+ 5,0.5,0.996541,1.75
5
+ 10,0.05,0.99133,2.95
6
+ 10,0.1,0.995319,2.75
7
+ 10,0.5,0.997328,2.85
8
+ 25,0.05,0.9957,6.56
9
+ 25,0.1,0.99683,6.66
10
+ 25,0.5,0.997896,6.36
fasttext/massive/fasttext_subword/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_subword",
3
+ "tags": {
4
+ "feature_type": "subword",
5
+ "wordNgrams": "2",
6
+ "minn": "2",
7
+ "maxn": "4",
8
+ "loss": "softmax",
9
+ "dataset_version": "massive"
10
+ },
11
+ "params": {
12
+ "epoch": "25",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.9974202868641007,
17
+ "f1_macro": 0.9973576195821323,
18
+ "precision_macro": 0.9973036655049542,
19
+ "recall_macro": 0.997414755322694,
20
+ "inference_time_total_s": 0.3027,
21
+ "inference_time_ms_per_sample": 0.015618,
22
+ "throughput_samples_per_sec": 64029.03
23
+ }
24
+ }
fasttext/massive/fasttext_word/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.00 0.00 0.00 0
4
+ da 0.99 0.99 0.99 957
5
+ de 1.00 1.00 1.00 1142
6
+ el 1.00 1.00 1.00 1141
7
+ en 0.99 1.00 0.99 963
8
+ es 0.99 0.99 0.99 1097
9
+ fi 1.00 0.97 0.98 1044
10
+ fr 1.00 1.00 1.00 1221
11
+ hr 0.00 0.00 0.00 0
12
+ it 1.00 0.99 0.99 1093
13
+ ja 0.58 0.98 0.73 1337
14
+ ko 1.00 0.98 0.99 1324
15
+ lt 0.00 0.00 0.00 0
16
+ mk 0.00 0.00 0.00 0
17
+ nb 0.00 0.00 0.00 0
18
+ nl 1.00 0.99 1.00 1129
19
+ pl 1.00 1.00 1.00 1002
20
+ pt 0.99 0.99 0.99 1070
21
+ ro 1.00 1.00 1.00 1010
22
+ ru 1.00 0.99 0.99 1023
23
+ sl 1.00 0.99 0.99 998
24
+ sv 0.99 0.99 0.99 957
25
+ uk 0.00 0.00 0.00 0
26
+ zh 0.25 0.02 0.03 874
27
+
28
+ accuracy 0.95 19382
29
+ macro avg 0.70 0.70 0.69 19382
30
+ weighted avg 0.93 0.95 0.93 19382
fasttext/massive/fasttext_word/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.922506,1.17
3
+ 5,0.1,0.925959,1.19
4
+ 5,0.5,0.927383,1.19
5
+ 10,0.05,0.926964,2.29
6
+ 10,0.1,0.92742,2.29
7
+ 10,0.5,0.929578,2.29
8
+ 25,0.05,0.928807,5.7
9
+ 25,0.1,0.928194,5.7
10
+ 25,0.5,0.92868,5.6
fasttext/massive/fasttext_word/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_word",
3
+ "tags": {
4
+ "feature_type": "word_ngram",
5
+ "wordNgrams": "1",
6
+ "minn": "0",
7
+ "maxn": "0",
8
+ "loss": "softmax",
9
+ "dataset_version": "massive"
10
+ },
11
+ "params": {
12
+ "epoch": "10",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.9462903725105768,
17
+ "f1_macro": 0.925781433897811,
18
+ "precision_macro": 0.9319867882832494,
19
+ "recall_macro": 0.9363424072842795,
20
+ "inference_time_total_s": 0.1498,
21
+ "inference_time_ms_per_sample": 0.007729,
22
+ "throughput_samples_per_sec": 129385.1
23
+ }
24
+ }
fasttext/mmarco/fasttext_subword/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.00 0.00 0.00 0
4
+ da 0.00 0.00 0.00 0
5
+ de 1.00 1.00 1.00 20033
6
+ el 0.00 0.00 0.00 0
7
+ en 1.00 1.00 1.00 20044
8
+ es 0.99 1.00 1.00 20016
9
+ fi 0.00 0.00 0.00 0
10
+ fr 1.00 1.00 1.00 19956
11
+ hr 0.00 0.00 0.00 0
12
+ it 1.00 1.00 1.00 19992
13
+ ja 1.00 1.00 1.00 19935
14
+ ko 0.00 0.00 0.00 0
15
+ lt 0.00 0.00 0.00 0
16
+ mk 0.00 0.00 0.00 0
17
+ nb 0.00 0.00 0.00 0
18
+ nl 1.00 1.00 1.00 19942
19
+ pl 0.00 0.00 0.00 0
20
+ pt 0.99 1.00 0.99 20036
21
+ ro 0.00 0.00 0.00 0
22
+ ru 1.00 1.00 1.00 20086
23
+ sl 0.00 0.00 0.00 0
24
+ sv 0.00 0.00 0.00 0
25
+ uk 0.00 0.00 0.00 0
26
+ zh 1.00 1.00 1.00 20005
27
+
28
+ accuracy 1.00 200045
29
+ macro avg 0.42 0.42 0.42 200045
30
+ weighted avg 1.00 1.00 1.00 200045
fasttext/mmarco/fasttext_subword/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.996202,12.13
3
+ 5,0.1,0.996976,11.92
4
+ 5,0.5,0.997588,12.07
5
+ 10,0.05,0.997001,22.21
6
+ 10,0.1,0.997388,22.1
7
+ 10,0.5,0.997613,22.43
8
+ 25,0.05,0.997401,51.92
9
+ 25,0.1,0.997526,51.68
10
+ 25,0.5,0.997551,51.37
fasttext/mmarco/fasttext_subword/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_subword",
3
+ "tags": {
4
+ "feature_type": "subword",
5
+ "wordNgrams": "2",
6
+ "minn": "2",
7
+ "maxn": "4",
8
+ "loss": "softmax",
9
+ "dataset_version": "mmarco"
10
+ },
11
+ "params": {
12
+ "epoch": "10",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.9974555724961883,
17
+ "f1_macro": 0.997456340339345,
18
+ "precision_macro": 0.9974583357050643,
19
+ "recall_macro": 0.9974550318130501,
20
+ "inference_time_total_s": 2.8056,
21
+ "inference_time_ms_per_sample": 0.014025,
22
+ "throughput_samples_per_sec": 71302.41
23
+ }
24
+ }
fasttext/mmarco/fasttext_word/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.00 0.00 0.00 0
4
+ da 0.00 0.00 0.00 0
5
+ de 1.00 0.98 0.99 20033
6
+ el 0.00 0.00 0.00 0
7
+ en 0.99 1.00 0.99 20044
8
+ es 0.99 0.99 0.99 20016
9
+ fi 0.00 0.00 0.00 0
10
+ fr 1.00 0.99 1.00 19956
11
+ hr 0.00 0.00 0.00 0
12
+ it 1.00 0.99 1.00 19992
13
+ ja 0.54 1.00 0.70 19935
14
+ ko 0.00 0.00 0.00 0
15
+ lt 0.00 0.00 0.00 0
16
+ mk 0.00 0.00 0.00 0
17
+ nb 0.00 0.00 0.00 0
18
+ nl 1.00 0.99 0.99 19942
19
+ pl 0.00 0.00 0.00 0
20
+ pt 0.99 0.99 0.99 20036
21
+ ro 0.00 0.00 0.00 0
22
+ ru 1.00 0.99 1.00 20086
23
+ sl 0.00 0.00 0.00 0
24
+ sv 0.00 0.00 0.00 0
25
+ uk 0.00 0.00 0.00 0
26
+ zh 0.92 0.18 0.30 20005
27
+
28
+ accuracy 0.91 200045
29
+ macro avg 0.39 0.38 0.37 200045
30
+ weighted avg 0.94 0.91 0.89 200045
fasttext/mmarco/fasttext_word/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.89358,8.71
3
+ 5,0.1,0.894014,8.62
4
+ 5,0.5,0.894589,8.84
5
+ 10,0.05,0.894002,15.85
6
+ 10,0.1,0.894355,17.07
7
+ 10,0.5,0.894588,16.88
8
+ 25,0.05,0.894251,39.96
9
+ 25,0.1,0.894636,39.23
10
+ 25,0.5,0.894333,44.92
fasttext/mmarco/fasttext_word/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_word",
3
+ "tags": {
4
+ "feature_type": "word_ngram",
5
+ "wordNgrams": "1",
6
+ "minn": "0",
7
+ "maxn": "0",
8
+ "loss": "softmax",
9
+ "dataset_version": "mmarco"
10
+ },
11
+ "params": {
12
+ "epoch": "25",
13
+ "lr": "0.1"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.9102751880826814,
17
+ "f1_macro": 0.8945281297744344,
18
+ "precision_macro": 0.9426291281301638,
19
+ "recall_macro": 0.9102787400266008,
20
+ "inference_time_total_s": 1.3177,
21
+ "inference_time_ms_per_sample": 0.006587,
22
+ "throughput_samples_per_sec": 151808.34
23
+ }
24
+ }
fasttext/multilingual_cc_news/fasttext_subword/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 1.00 1.00 1.00 20015
4
+ da 0.99 0.99 0.99 20013
5
+ de 1.00 1.00 1.00 20008
6
+ el 1.00 1.00 1.00 20050
7
+ en 0.99 1.00 0.99 20005
8
+ es 0.00 0.00 0.00 0
9
+ fi 1.00 1.00 1.00 20084
10
+ fr 0.00 0.00 0.00 0
11
+ hr 1.00 1.00 1.00 19914
12
+ it 0.00 0.00 0.00 0
13
+ ja 1.00 1.00 1.00 20022
14
+ ko 1.00 1.00 1.00 20000
15
+ lt 1.00 0.99 1.00 4950
16
+ mk 1.00 1.00 1.00 18680
17
+ nb 0.99 0.99 0.99 19959
18
+ nl 1.00 1.00 1.00 20009
19
+ pl 1.00 1.00 1.00 20033
20
+ pt 1.00 1.00 1.00 19949
21
+ ro 1.00 1.00 1.00 20113
22
+ ru 1.00 1.00 1.00 20065
23
+ sl 1.00 1.00 1.00 19951
24
+ sv 1.00 1.00 1.00 20044
25
+ uk 1.00 1.00 1.00 20033
26
+ zh 1.00 1.00 1.00 19979
27
+
28
+ accuracy 1.00 403876
29
+ macro avg 0.87 0.87 0.87 403876
30
+ weighted avg 1.00 1.00 1.00 403876
fasttext/multilingual_cc_news/fasttext_subword/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.99585,52.53
3
+ 5,0.1,0.99649,52.76
4
+ 5,0.5,0.996956,52.74
5
+ 10,0.05,0.996493,94.23
6
+ 10,0.1,0.996703,93.75
7
+ 10,0.5,0.996912,94.22
8
+ 25,0.05,0.996813,214.21
9
+ 25,0.1,0.996912,212.85
10
+ 25,0.5,0.996966,213.97
fasttext/multilingual_cc_news/fasttext_subword/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_subword",
3
+ "tags": {
4
+ "feature_type": "subword",
5
+ "wordNgrams": "2",
6
+ "minn": "2",
7
+ "maxn": "4",
8
+ "loss": "softmax",
9
+ "dataset_version": "multilingual_cc_news"
10
+ },
11
+ "params": {
12
+ "epoch": "25",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.9970882151947628,
17
+ "f1_macro": 0.9970508016807916,
18
+ "precision_macro": 0.9971120291073209,
19
+ "recall_macro": 0.9969945619392637,
20
+ "inference_time_total_s": 14.7611,
21
+ "inference_time_ms_per_sample": 0.036548,
22
+ "throughput_samples_per_sec": 27360.9
23
+ }
24
+ }
fasttext/multilingual_cc_news/fasttext_word/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 1.00 1.00 1.00 20015
4
+ da 0.99 0.98 0.99 20013
5
+ de 1.00 1.00 1.00 20008
6
+ el 1.00 1.00 1.00 20050
7
+ en 0.98 1.00 0.99 20005
8
+ es 0.00 0.00 0.00 0
9
+ fi 1.00 0.99 1.00 20084
10
+ fr 0.00 0.00 0.00 0
11
+ hr 1.00 1.00 1.00 19914
12
+ it 0.00 0.00 0.00 0
13
+ ja 0.82 0.98 0.89 20022
14
+ ko 1.00 0.98 0.99 20000
15
+ lt 1.00 0.99 0.99 4950
16
+ mk 1.00 1.00 1.00 18680
17
+ nb 0.98 0.99 0.99 19959
18
+ nl 1.00 1.00 1.00 20009
19
+ pl 1.00 1.00 1.00 20033
20
+ pt 1.00 1.00 1.00 19949
21
+ ro 1.00 1.00 1.00 20113
22
+ ru 1.00 1.00 1.00 20065
23
+ sl 1.00 1.00 1.00 19951
24
+ sv 1.00 1.00 1.00 20044
25
+ uk 1.00 0.99 1.00 20033
26
+ zh 0.98 0.80 0.88 19979
27
+
28
+ accuracy 0.98 403876
29
+ macro avg 0.86 0.86 0.86 403876
30
+ weighted avg 0.99 0.98 0.98 403876
fasttext/multilingual_cc_news/fasttext_word/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.984308,31.82
3
+ 5,0.1,0.984619,32.82
4
+ 5,0.5,0.985011,32.56
5
+ 10,0.05,0.984564,56.96
6
+ 10,0.1,0.984768,59.4
7
+ 10,0.5,0.98506,57.74
8
+ 25,0.05,0.984702,131.39
9
+ 25,0.1,0.984856,132.89
10
+ 25,0.5,0.985055,133.91
fasttext/multilingual_cc_news/fasttext_word/run_metadata.json ADDED
@@ -0,0 +1,24 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "run_name": "fasttext_word",
3
+ "tags": {
4
+ "feature_type": "word_ngram",
5
+ "wordNgrams": "1",
6
+ "minn": "0",
7
+ "maxn": "0",
8
+ "loss": "softmax",
9
+ "dataset_version": "multilingual_cc_news"
10
+ },
11
+ "params": {
12
+ "epoch": "10",
13
+ "lr": "0.5"
14
+ },
15
+ "metrics": {
16
+ "accuracy": 0.9847775059671781,
17
+ "f1_macro": 0.985102850565851,
18
+ "precision_macro": 0.9868653316205467,
19
+ "recall_macro": 0.9849375987593945,
20
+ "inference_time_total_s": 5.1684,
21
+ "inference_time_ms_per_sample": 0.012797,
22
+ "throughput_samples_per_sec": 78142.71
23
+ }
24
+ }
fasttext/multilingual_toxicity_dataset/fasttext_subword/classification_report.txt ADDED
@@ -0,0 +1,30 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ precision recall f1-score support
2
+
3
+ ca 0.00 0.00 0.00 0
4
+ da 0.00 0.00 0.00 0
5
+ de 1.00 1.00 1.00 506
6
+ el 0.00 0.00 0.00 0
7
+ en 0.99 1.00 0.99 478
8
+ es 1.00 0.99 0.99 495
9
+ fi 0.00 0.00 0.00 0
10
+ fr 1.00 0.99 0.99 472
11
+ hr 0.00 0.00 0.00 0
12
+ it 0.99 1.00 0.99 478
13
+ ja 1.00 0.99 0.99 493
14
+ ko 0.00 0.00 0.00 0
15
+ lt 0.00 0.00 0.00 0
16
+ mk 0.00 0.00 0.00 0
17
+ nb 0.00 0.00 0.00 0
18
+ nl 0.00 0.00 0.00 0
19
+ pl 0.00 0.00 0.00 0
20
+ pt 0.00 0.00 0.00 0
21
+ ro 0.00 0.00 0.00 0
22
+ ru 1.00 1.00 1.00 492
23
+ sl 0.00 0.00 0.00 0
24
+ sv 0.00 0.00 0.00 0
25
+ uk 1.00 1.00 1.00 455
26
+ zh 0.98 1.00 0.99 455
27
+
28
+ accuracy 0.99 4324
29
+ macro avg 0.37 0.37 0.37 4324
30
+ weighted avg 0.99 0.99 0.99 4324
fasttext/multilingual_toxicity_dataset/fasttext_subword/param_search_results.csv ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ epoch,lr,val_f1_macro,train_time_s
2
+ 5,0.05,0.196243,0.87
3
+ 5,0.1,0.716703,0.87
4
+ 5,0.5,0.990663,0.88
5
+ 10,0.05,0.722254,1.07
6
+ 10,0.1,0.98601,1.08
7
+ 10,0.5,0.991235,1.08
8
+ 25,0.05,0.987171,1.78
9
+ 25,0.1,0.991235,1.88
10
+ 25,0.5,0.991823,1.89