End of training
Browse files- README.md +13 -13
- model.safetensors +1 -1
- predictions_test_wer95.00.csv +3 -0
- predictions_test_wer95.00_hyp.trn +2 -0
- predictions_test_wer95.00_hyp.trn.dtl +107 -0
- predictions_test_wer95.00_hyp.trn.snt.utt +47 -0
- predictions_test_wer95.00_hyp.trn.sys +18 -0
- predictions_test_wer95.00_ref.trn +2 -0
- training_args.bin +1 -1
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.
|
| 21 |
-
- Cer:
|
| 22 |
-
- Wer:
|
| 23 |
-
- Mer:
|
| 24 |
-
- Wil:
|
| 25 |
-
- Wip: 0
|
| 26 |
-
- Hits:
|
| 27 |
-
- Substitutions:
|
| 28 |
-
- Deletions:
|
| 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
|
| 59 |
-
|
| 60 |
-
|
|
| 61 |
-
|
|
| 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:
|
| 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:
|
| 3 |
size 5496
|
|
|
|
| 1 |
version https://git-lfs.github.com/spec/v1
|
| 2 |
+
oid sha256:559d421ca1bec408255b36b49256003405d3c8ab9bc317575f8fe66f3767e497
|
| 3 |
size 5496
|