Lakoc commited on
Commit
b5f5a07
1 Parent(s): 90c11f1

End of training

Browse files
README.md CHANGED
@@ -17,15 +17,15 @@ should probably proofread and complete it, then remove this comment. -->
17
 
18
  This model is a fine-tuned version of [](https://huggingface.co/) on the fleurs dataset.
19
  It achieves the following results on the evaluation set:
20
- - Loss: 7.9877
21
- - Cer: 1.0
22
- - Wer: 1.0
23
- - Mer: 1.0
24
- - Wil: 1
25
- - Wip: 0
26
- - Hits: 0
27
- - Substitutions: 0
28
- - Deletions: 37
29
  - Insertions: 0
30
 
31
  ## Model description
@@ -55,10 +55,10 @@ The following hyperparameters were used during training:
55
 
56
  ### Training results
57
 
58
- | Training Loss | Epoch | Step | Validation Loss | Cer | Wer | Mer | Wil | Wip | Hits | Substitutions | Deletions | Insertions |
59
- |:-------------:|:-----:|:----:|:---------------:|:---:|:---:|:---:|:---:|:---:|:----:|:-------------:|:---------:|:----------:|
60
- | 7.9999 | 0.0 | 4 | 7.9908 | 1.0 | 1.0 | 1.0 | 1 | 0 | 0 | 0 | 37 | 0 |
61
- | 7.9169 | 0.01 | 8 | 7.9877 | 1.0 | 1.0 | 1.0 | 1 | 0 | 0 | 0 | 37 | 0 |
62
 
63
 
64
  ### Framework versions
 
17
 
18
  This model is a fine-tuned version of [](https://huggingface.co/) on the fleurs dataset.
19
  It achieves the following results on the evaluation set:
20
+ - Loss: 7.9999
21
+ - Cer: 0.9756
22
+ - Wer: 0.975
23
+ - Mer: 0.975
24
+ - Wil: 0.9875
25
+ - Wip: 0.0125
26
+ - Hits: 1
27
+ - Substitutions: 1
28
+ - Deletions: 38
29
  - Insertions: 0
30
 
31
  ## Model description
 
55
 
56
  ### Training results
57
 
58
+ | Training Loss | Epoch | Step | Validation Loss | Cer | Wer | Mer | Wil | Wip | Hits | Substitutions | Deletions | Insertions |
59
+ |:-------------:|:-----:|:----:|:---------------:|:------:|:-----:|:-----:|:------:|:------:|:----:|:-------------:|:---------:|:----------:|
60
+ | 8.1019 | 0.0 | 4 | 8.0334 | 0.9756 | 0.975 | 0.975 | 0.9875 | 0.0125 | 1 | 1 | 38 | 0 |
61
+ | 8.0279 | 0.01 | 8 | 7.9999 | 0.9756 | 0.975 | 0.975 | 0.9875 | 0.0125 | 1 | 1 | 38 | 0 |
62
 
63
 
64
  ### Framework versions
model.safetensors CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:c153c19d30e5e6bd473b126f82415f6ce389cf5e3c5d2ce34c5bd9f7f9064e20
3
  size 155523256
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:27215ad9b82fccbd5b92ac34d87682763113f31ad585259c32e6fd17147c816e
3
  size 155523256
predictions_test_wer95.00.csv ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ label,prediction
2
+ however due to the slow communication channels styles in the west could lag behind by 25 to 30 year,the
3
+ all nouns alongside the word sie for you always begin with a capital letter even in the middle of a sentence,the
predictions_test_wer95.00_hyp.trn ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ the (utterance_0)
2
+ the (utterance_1)
predictions_test_wer95.00_hyp.trn.dtl ADDED
@@ -0,0 +1,107 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ DETAILED OVERALL REPORT FOR THE SYSTEM: experiment/predictions_test_wer95.00_hyp.trn
2
+
3
+ SENTENCE RECOGNITION PERFORMANCE
4
+
5
+ sentences 2
6
+ with errors 100.0% ( 2)
7
+
8
+ with substitutions 0.0% ( 0)
9
+ with deletions 100.0% ( 2)
10
+ with insertions 0.0% ( 0)
11
+
12
+
13
+ WORD RECOGNITION PERFORMANCE
14
+
15
+ Percent Total Error = 95.0% ( 38)
16
+
17
+ Percent Correct = 5.0% ( 2)
18
+
19
+ Percent Substitution = 0.0% ( 0)
20
+ Percent Deletions = 95.0% ( 38)
21
+ Percent Insertions = 0.0% ( 0)
22
+ Percent Word Accuracy = 5.0%
23
+
24
+
25
+ Ref. words = ( 40)
26
+ Hyp. words = ( 2)
27
+ Aligned words = ( 40)
28
+
29
+ CONFUSION PAIRS Total (0)
30
+ With >= 1 occurrences (0)
31
+
32
+ -------
33
+ 0
34
+
35
+
36
+
37
+ INSERTIONS Total (0)
38
+ With >= 1 occurrences (0)
39
+
40
+ -------
41
+ 0
42
+
43
+
44
+
45
+ DELETIONS Total (34)
46
+ With >= 1 occurrences (34)
47
+
48
+ 1: 2 -> a
49
+ 2: 2 -> in
50
+ 3: 2 -> the
51
+ 4: 2 -> to
52
+ 5: 1 -> 25
53
+ 6: 1 -> 30
54
+ 7: 1 -> all
55
+ 8: 1 -> alongside
56
+ 9: 1 -> always
57
+ 10: 1 -> begin
58
+ 11: 1 -> behind
59
+ 12: 1 -> by
60
+ 13: 1 -> capital
61
+ 14: 1 -> channels
62
+ 15: 1 -> communication
63
+ 16: 1 -> could
64
+ 17: 1 -> due
65
+ 18: 1 -> even
66
+ 19: 1 -> for
67
+ 20: 1 -> however
68
+ 21: 1 -> lag
69
+ 22: 1 -> letter
70
+ 23: 1 -> middle
71
+ 24: 1 -> nouns
72
+ 25: 1 -> of
73
+ 26: 1 -> sentence
74
+ 27: 1 -> sie
75
+ 28: 1 -> slow
76
+ 29: 1 -> styles
77
+ 30: 1 -> west
78
+ 31: 1 -> with
79
+ 32: 1 -> word
80
+ 33: 1 -> year
81
+ 34: 1 -> you
82
+ -------
83
+ 38
84
+
85
+
86
+
87
+ SUBSTITUTIONS Total (0)
88
+ With >= 1 occurrences (0)
89
+
90
+ -------
91
+ 0
92
+
93
+
94
+ * NOTE: The 'Substitution' words are those reference words
95
+ for which the recognizer supplied an incorrect word.
96
+
97
+
98
+ FALSELY RECOGNIZED Total (0)
99
+ With >= 1 occurrences (0)
100
+
101
+ -------
102
+ 0
103
+
104
+
105
+ * NOTE: The 'Falsely Recognized' words are those hypothesis words
106
+ which the recognizer incorrectly substituted for a reference word.
107
+
predictions_test_wer95.00_hyp.trn.snt.utt ADDED
@@ -0,0 +1,47 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ===============================================================================
2
+
3
+ SENTENCE LEVEL REPORT FOR THE SYSTEM:
4
+ Name: experiment/predictions_test_wer95.00_hyp.trn
5
+
6
+ ===============================================================================
7
+
8
+
9
+ SPEAKER utt
10
+ id: (utterance_0)
11
+ Scores: (#C #S #D #I) 1 0 18 0
12
+ REF: HOWEVER DUE TO THE SLOW COMMUNICATION CHANNELS STYLES IN the WEST COULD LAG BEHIND BY 25 TO 30 YEAR
13
+ HYP: ******* *** ** *** **** ************* ******** ****** ** the **** ***** *** ****** ** ** ** ** ****
14
+ Eval: D D D D D D D D D D D D D D D D D D
15
+
16
+ Correct = 5.3% 1 ( 1)
17
+ Substitutions = 0.0% 0 ( 0)
18
+ Deletions = 94.7% 18 ( 18)
19
+ Insertions = 0.0% 0 ( 0)
20
+
21
+ Errors = 94.7% 18 ( 18)
22
+
23
+ Ref. words = 19 ( 19)
24
+ Hyp. words = 1 ( 1)
25
+ Aligned words = 19 ( 19)
26
+
27
+ -------------------------------------------------------------------------------
28
+
29
+ id: (utterance_1)
30
+ Scores: (#C #S #D #I) 1 0 20 0
31
+ REF: ALL NOUNS ALONGSIDE THE WORD SIE FOR YOU ALWAYS BEGIN WITH A CAPITAL LETTER EVEN IN the MIDDLE OF A SENTENCE
32
+ HYP: *** ***** ********* *** **** *** *** *** ****** ***** **** * ******* ****** **** ** the ****** ** * ********
33
+ Eval: D D D D D D D D D D D D D D D D D D D D
34
+
35
+ Correct = 4.8% 1 ( 2)
36
+ Substitutions = 0.0% 0 ( 0)
37
+ Deletions = 95.2% 20 ( 38)
38
+ Insertions = 0.0% 0 ( 0)
39
+
40
+ Errors = 95.2% 20 ( 38)
41
+
42
+ Ref. words = 21 ( 40)
43
+ Hyp. words = 1 ( 2)
44
+ Aligned words = 21 ( 40)
45
+
46
+ -------------------------------------------------------------------------------
47
+
predictions_test_wer95.00_hyp.trn.sys ADDED
@@ -0,0 +1,18 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+
2
+
3
+
4
+ SYSTEM SUMMARY PERCENTAGES by SPEAKER
5
+
6
+ ,----------------------------------------------------------------.
7
+ | experiment/predictions_test_wer95.00_hyp.trn |
8
+ |----------------------------------------------------------------|
9
+ | SPKR | # Snt # Wrd | Corr Sub Del Ins Err S.Err |
10
+ |--------+-------------+-----------------------------------------|
11
+ | utt | 2 40 | 5.0 0.0 95.0 0.0 95.0 100.0 |
12
+ |================================================================|
13
+ | Sum/Avg| 2 40 | 5.0 0.0 95.0 0.0 95.0 100.0 |
14
+ |================================================================|
15
+ | Mean | 2.0 40.0 | 5.0 0.0 95.0 0.0 95.0 100.0 |
16
+ | S.D. | 0.0 0.0 | 0.0 0.0 0.0 0.0 0.0 0.0 |
17
+ | Median | 2.0 40.0 | 5.0 0.0 95.0 0.0 95.0 100.0 |
18
+ `----------------------------------------------------------------'
predictions_test_wer95.00_ref.trn ADDED
@@ -0,0 +1,2 @@
 
 
 
1
+ however due to the slow communication channels styles in the west could lag behind by 25 to 30 year (utterance_0)
2
+ all nouns alongside the word sie for you always begin with a capital letter even in the middle of a sentence (utterance_1)
training_args.bin CHANGED
@@ -1,3 +1,3 @@
1
  version https://git-lfs.github.com/spec/v1
2
- oid sha256:7cc4ba9fb8a0dd76d3b6f2eeb454c5b8d63b06194a670d8e85e9bd7b07ecf810
3
  size 5496
 
1
  version https://git-lfs.github.com/spec/v1
2
+ oid sha256:559d421ca1bec408255b36b49256003405d3c8ab9bc317575f8fe66f3767e497
3
  size 5496