update generation params
Browse files
README.md
CHANGED
@@ -1,22 +1,23 @@
|
|
1 |
---
|
2 |
-
license:
|
3 |
tags:
|
4 |
- grammar
|
5 |
- spelling
|
6 |
- punctuation
|
7 |
- error-correction
|
8 |
-
|
|
|
9 |
widget:
|
10 |
-
- text: "
|
11 |
-
example_title: "
|
|
|
|
|
12 |
- text: "so em if we have an now so with fito ringina know how to estimate the tren given the ereafte mylite trend we can also em an estimate is nod s
|
13 |
i again tort watfettering an we have estimated the trend an
|
14 |
called wot to be called sthat of exty right now we can and look at
|
15 |
wy this should not hare a trend i becan we just remove the trend an and we can we now estimate
|
16 |
tesees ona effect of them exty"
|
17 |
example_title: "Transcribed Audio Example 2"
|
18 |
-
- text: "I would like a peice of pie."
|
19 |
-
example_title: "miss-spelling"
|
20 |
- text: "My coworker said he used a financial planner to help choose his stocks so he wouldn't loose money."
|
21 |
example_title: "incorrect word choice (context)"
|
22 |
- text: "good so hve on an tadley i'm not able to make it to the exla session on monday this week e which is why i am e recording pre recording
|
@@ -25,26 +26,66 @@ ta ohow to remove trents in these nalitives from time series"
|
|
25 |
example_title: "lowercased audio transcription output"
|
26 |
- text: "Frustrated, the chairs took me forever to set up."
|
27 |
example_title: "dangling modifier"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
28 |
- text: "There car broke down so their hitching a ride to they're class."
|
29 |
example_title: "compound-1"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
30 |
- text: "Which part of Zurich was you going to go hiking in when we were there for the first time together? ! ?"
|
31 |
example_title: "chatbot on Zurich"
|
32 |
|
33 |
-
|
34 |
-
|
35 |
-
|
36 |
-
|
37 |
-
|
38 |
-
|
39 |
-
|
40 |
-
length_penalty: 1
|
41 |
-
early_stopping: True
|
42 |
---
|
43 |
|
|
|
44 |
|
45 |
# t5-v1_1-base-ft-jflAUG
|
46 |
|
47 |
-
|
48 |
|
49 |
- this model _(at least from preliminary testing)_ can handle large amounts of errors in the source text (i.e. from audio transcription) and still produce cohesive results.
|
50 |
- a fine-tuned version of [google/t5-v1_1-base](https://huggingface.co/google/t5-v1_1-base) on an expanded version of the [JFLEG dataset](https://aclanthology.org/E17-2037/).
|
|
|
1 |
---
|
2 |
+
license: cc-by-nc-sa-4.0
|
3 |
tags:
|
4 |
- grammar
|
5 |
- spelling
|
6 |
- punctuation
|
7 |
- error-correction
|
8 |
+
datasets:
|
9 |
+
- jfleg
|
10 |
widget:
|
11 |
+
- text: "i can has cheezburger"
|
12 |
+
example_title: "cheezburger"
|
13 |
+
- text: "There car broke down so their hitching a ride to they're class."
|
14 |
+
example_title: "compound-1"
|
15 |
- text: "so em if we have an now so with fito ringina know how to estimate the tren given the ereafte mylite trend we can also em an estimate is nod s
|
16 |
i again tort watfettering an we have estimated the trend an
|
17 |
called wot to be called sthat of exty right now we can and look at
|
18 |
wy this should not hare a trend i becan we just remove the trend an and we can we now estimate
|
19 |
tesees ona effect of them exty"
|
20 |
example_title: "Transcribed Audio Example 2"
|
|
|
|
|
21 |
- text: "My coworker said he used a financial planner to help choose his stocks so he wouldn't loose money."
|
22 |
example_title: "incorrect word choice (context)"
|
23 |
- text: "good so hve on an tadley i'm not able to make it to the exla session on monday this week e which is why i am e recording pre recording
|
|
|
26 |
example_title: "lowercased audio transcription output"
|
27 |
- text: "Frustrated, the chairs took me forever to set up."
|
28 |
example_title: "dangling modifier"
|
29 |
+
- text: "I would like a peice of pie."
|
30 |
+
example_title: "miss-spelling"
|
31 |
+
- text: "Which part of Zurich was you going to go hiking in when we were there for the first time together? ! ?"
|
32 |
+
example_title: "chatbot on Zurich"
|
33 |
+
|
34 |
+
parameters:
|
35 |
+
max_length: 128
|
36 |
+
min_length: 4
|
37 |
+
num_beams: 4
|
38 |
+
repetition_penalty: 1.21
|
39 |
+
length_penalty: 1
|
40 |
+
early_stopping: True
|
41 |
+
---
|
42 |
+
---
|
43 |
+
license: cc-by-nc-sa-4.0
|
44 |
+
tags:
|
45 |
+
- grammar
|
46 |
+
- spelling
|
47 |
+
- punctuation
|
48 |
+
- error-correction
|
49 |
+
datasets:
|
50 |
+
- jfleg
|
51 |
+
widget:
|
52 |
+
- text: "i can has cheezburger"
|
53 |
+
example_title: "cheezburger"
|
54 |
- text: "There car broke down so their hitching a ride to they're class."
|
55 |
example_title: "compound-1"
|
56 |
+
- text: "so em if we have an now so with fito ringina know how to estimate the tren given the ereafte mylite trend we can also em an estimate is nod s
|
57 |
+
i again tort watfettering an we have estimated the trend an
|
58 |
+
called wot to be called sthat of exty right now we can and look at
|
59 |
+
wy this should not hare a trend i becan we just remove the trend an and we can we now estimate
|
60 |
+
tesees ona effect of them exty"
|
61 |
+
example_title: "Transcribed Audio Example 2"
|
62 |
+
- text: "My coworker said he used a financial planner to help choose his stocks so he wouldn't loose money."
|
63 |
+
example_title: "incorrect word choice (context)"
|
64 |
+
- text: "good so hve on an tadley i'm not able to make it to the exla session on monday this week e which is why i am e recording pre recording
|
65 |
+
an this excelleision and so to day i want e to talk about two things and first of all em i wont em wene give a summary er about
|
66 |
+
ta ohow to remove trents in these nalitives from time series"
|
67 |
+
example_title: "lowercased audio transcription output"
|
68 |
+
- text: "Frustrated, the chairs took me forever to set up."
|
69 |
+
example_title: "dangling modifier"
|
70 |
+
- text: "I would like a peice of pie."
|
71 |
+
example_title: "miss-spelling"
|
72 |
- text: "Which part of Zurich was you going to go hiking in when we were there for the first time together? ! ?"
|
73 |
example_title: "chatbot on Zurich"
|
74 |
|
75 |
+
parameters:
|
76 |
+
max_length: 128
|
77 |
+
min_length: 2
|
78 |
+
num_beams: 4
|
79 |
+
repetition_penalty: 1.21
|
80 |
+
length_penalty: 1
|
81 |
+
early_stopping: True
|
|
|
|
|
82 |
---
|
83 |
|
84 |
+
> A more recent version can be found [here](https://huggingface.co/pszemraj/grammar-synthesis-large). Training smaller and/or comparably sized models is a WIP.
|
85 |
|
86 |
# t5-v1_1-base-ft-jflAUG
|
87 |
|
88 |
+
**GOAL:** a more robust and generalized grammar and spelling correction model that corrects everything in a single shot. It should have a minimal impact on the semantics of correct sentences (i.e. it does not change things that do not need to be changed).
|
89 |
|
90 |
- this model _(at least from preliminary testing)_ can handle large amounts of errors in the source text (i.e. from audio transcription) and still produce cohesive results.
|
91 |
- a fine-tuned version of [google/t5-v1_1-base](https://huggingface.co/google/t5-v1_1-base) on an expanded version of the [JFLEG dataset](https://aclanthology.org/E17-2037/).
|