Total runs:
266.1K
Run Growth:
-79.9K
Growth Rate:
-30.05%
Sponsored by
Vmake Labs Video Enhancer
mrm8488 / t5-base-finetuned-sarcasm-twitter
Model's Last Updated: March 18 2023
text-generation
Introduction of t5-base-finetuned-sarcasm-twitter
Model Details of t5-base-finetuned-sarcasm-twitter
T5-base fine-tuned for Sarcasm Detection 🙄
Google's T5 base fine-tuned on Twitter Sarcasm Dataset for Sequence classification (as text generation) downstream task.
Details of T5
The T5 model was presented in Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformer by Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, Peter J. Liu in Here the abstract:
Transfer learning, where a model is first pre-trained on a data-rich task before being fine-tuned on a downstream task, has emerged as a powerful technique in natural language processing (NLP). The effectiveness of transfer learning has given rise to a diversity of approaches, methodology, and practice. In this paper, we explore the landscape of transfer learning techniques for NLP by introducing a unified framework that converts every language problem into a text-to-text format. Our systematic study compares pre-training objectives, architectures, unlabeled datasets, transfer approaches, and other factors on dozens of language understanding tasks. By combining the insights from our exploration with scale and our new “Colossal Clean Crawled Corpus”, we achieve state-of-the-art results on many benchmarks covering summarization, question answering, text classification, and more. To facilitate future work on transfer learning for NLP, we release our dataset, pre-trained models, and code.
Details of the downstream task (Sequence Classification as Text generation) - Dataset 📚
For Twitter training and testing datasets are provided for sarcasm detection tasks in jsonlines format.
Each line contains a JSON object with the following fields :
-
label
:
SARCASMorNOT_SARCASM- NOT in test data
-
id
: String identifier for sample. This id will be required when making submissions.
- ONLY in test data
- response : the sarcastic response, whether a sarcastic Tweet
-
context
: the conversation context of the
response
-
Note, the context is an ordered list of dialogue, i.e., if the context contains three elements,
c1,c2,c3, in that order, thenc2is a reply toc1andc3is a reply toc2. Further, if the sarcastic response isr, thenris a reply toc3.
-
Note, the context is an ordered list of dialogue, i.e., if the context contains three elements,
For instance, for the following training example :
"label": "SARCASM", "response": "Did Kelly just call someone else messy? Baaaahaaahahahaha", "context": ["X is looking a First Lady should . #classact, "didn't think it was tailored enough it looked messy"]
The response tweet, "Did Kelly..." is a reply to its immediate context "didn't think it was tailored..." which is a reply to "X is looking...". Your goal is to predict the label of the "response" while also using the context (i.e, the immediate or the full context).
Dataset size statistics :
| Train | Val | Test | |
|---|---|---|---|
| 4050 | 450 | 500 |
The datasets was preprocessed to convert it to a text-to-text (classfication as generation task).
Model fine-tuning 🏋️
The training script is a slightly modified version of this Colab Notebook created by Suraj Patil , so all credits to him!
Test set metrics 🧾
| precision | recall | f1-score | support | |
|---|---|---|---|---|
| derison | 0.84 | 0.80 | 0.82 | 246 |
| normal | 0.82 | 0.85 | 0.83 | 254 |
| accuracy | 0.83 | 500 | ||
| macro avg | 0.83 | 0.83 | 0.83 | 500 |
| weighted avg | 0.83 | 0.83 | 0.83 | 500 |
Model in Action 🚀
from transformers import AutoTokenizer, AutoModelWithLMHead
tokenizer = AutoTokenizer.from_pretrained("mrm8488/t5-base-finetuned-sarcasm-twitter")
model = AutoModelWithLMHead.from_pretrained("mrm8488/t5-base-finetuned-sarcasm-twitter")
def eval_conversation(text):
input_ids = tokenizer.encode(text + '</s>', return_tensors='pt')
output = model.generate(input_ids=input_ids, max_length=3)
dec = [tokenizer.decode(ids) for ids in output]
label = dec[0]
return label
# For similarity with the training dataset we should replace users mentions in twits for @USER token and urls for URL token.
twit1 = "Trump just suspended the visa program that allowed me to move to the US to start @USER!" +
" Unfortunately, I won’t be able to vote in a few months but if you can, please vote him out, " +
"he's destroying what made America great in so many different ways!"
twit2 = "@USER @USER @USER We have far more cases than any other country, " +
"so leaving remote workers in would be disastrous. Makes Trump sense."
twit3 = "My worry is that i wouldn’t be surprised if half the country actually agrees with this move..."
me = "Trump doing so??? It must be a mistake... XDDD"
conversation = twit1 + twit2
eval_conversation(conversation) #Output: 'derison'
conversation = twit1 + twit3
eval_conversation(conversation) #Output: 'normal'
conversation = twit1 + me
eval_conversation(conversation) #Output: 'derison'
# We will get 'normal' when sarcasm is not detected and 'derison' when detected
Created by Manuel Romero/@mrm8488 | LinkedIn
Made with ♥ in Spain
Runs of mrm8488 t5-base-finetuned-sarcasm-twitter on huggingface.co
9.5K
Total runs
0
24-hour runs
-34
3-day runs
-30
7-day runs
4.9K
30-day runs
More Information About t5-base-finetuned-sarcasm-twitter huggingface.co Model
Url of t5-base-finetuned-sarcasm-twitter
t5-base-finetuned-sarcasm-twitter huggingface.co Url
Provider of t5-base-finetuned-sarcasm-twitter huggingface.co
Other API from mrm8488
Updated:January 21 2024
Total runs:
65.7K
Run Growth:
49.2K
Growth Rate:
74.89%
Total runs:
33.2K
Run Growth:
947
Growth Rate:
2.85%
Total runs:
13.0K
Run Growth:
-431
Growth Rate:
-3.32%
Total runs:
9.0K
Run Growth:
689
Growth Rate:
7.88%
Total runs:
8.7K
Run Growth:
5.7K
Growth Rate:
65.22%
Total runs:
7.8K
Run Growth:
-2.1K
Growth Rate:
-27.21%
Total runs:
5.9K
Run Growth:
-53
Growth Rate:
-0.91%
Total runs:
4.4K
Run Growth:
1.4K
Growth Rate:
32.28%
Total runs:
3.7K
Run Growth:
-893
Growth Rate:
-24.06%
Total runs:
2.4K
Run Growth:
-1.3K
Growth Rate:
-53.78%
Total runs:
1.5K
Run Growth:
141
Growth Rate:
9.56%
Total runs:
1.4K
Run Growth:
612
Growth Rate:
42.71%
Total runs:
1.1K
Run Growth:
642
Growth Rate:
58.68%
Total runs:
987
Run Growth:
253
Growth Rate:
25.63%
Total runs:
973
Run Growth:
-1.3K
Growth Rate:
-136.90%
Total runs:
844
Run Growth:
3
Growth Rate:
0.37%
Total runs:
779
Run Growth:
35
Growth Rate:
4.49%
Total runs:
682
Run Growth:
335
Growth Rate:
49.12%
Total runs:
671
Run Growth:
-232
Growth Rate:
-34.12%
Total runs:
647
Run Growth:
-320
Growth Rate:
-49.46%
Total runs:
634
Run Growth:
-1.2K
Growth Rate:
-195.27%
Total runs:
619
Run Growth:
63
Growth Rate:
10.21%
Total runs:
481
Run Growth:
150
Growth Rate:
32.47%
Total runs:
471
Run Growth:
-574
Growth Rate:
-120.84%
Total runs:
464
Run Growth:
366
Growth Rate:
78.88%
Total runs:
374
Run Growth:
340
Growth Rate:
90.91%
Total runs:
366
Run Growth:
57
Growth Rate:
15.57%
Total runs:
362
Run Growth:
-143
Growth Rate:
-39.39%
Total runs:
335
Run Growth:
-43
Growth Rate:
-12.80%
Total runs:
326
Run Growth:
199
Growth Rate:
61.42%
Total runs:
326
Run Growth:
88
Growth Rate:
26.99%
Total runs:
300
Run Growth:
-48
Growth Rate:
-16.00%
Total runs:
269
Run Growth:
24
Growth Rate:
8.92%
Total runs:
260
Run Growth:
255
Growth Rate:
98.08%
Total runs:
250
Run Growth:
-122
Growth Rate:
-48.80%
Total runs:
242
Run Growth:
-17
Growth Rate:
-7.02%
Total runs:
227
Run Growth:
-197
Growth Rate:
-86.78%
Total runs:
202
Run Growth:
37
Growth Rate:
18.32%
Total runs:
190
Run Growth:
157
Growth Rate:
82.63%
Total runs:
175
Run Growth:
141
Growth Rate:
80.57%
Total runs:
167
Run Growth:
97
Growth Rate:
57.74%
Total runs:
154
Run Growth:
120
Growth Rate:
77.92%
Total runs:
150
Run Growth:
-96
Growth Rate:
-60.38%
Total runs:
149
Run Growth:
1
Growth Rate:
0.61%
Total runs:
145
Run Growth:
-327
Growth Rate:
-218.00%
Total runs:
131
Run Growth:
69
Growth Rate:
52.67%
Total runs:
130
Run Growth:
30
Growth Rate:
27.03%
Total runs:
127
Run Growth:
57
Growth Rate:
44.88%
Total runs:
127
Run Growth:
44
Growth Rate:
34.65%
Total runs:
122
Run Growth:
-60
Growth Rate:
-49.18%
Total runs:
121
Run Growth:
-20
Growth Rate:
-15.63%
Total runs:
120
Run Growth:
-224
Growth Rate:
-186.67%
Total runs:
118
Run Growth:
26
Growth Rate:
22.03%
Total runs:
116
Run Growth:
38
Growth Rate:
32.76%
Total runs:
116
Run Growth:
89
Growth Rate:
77.39%
Total runs:
113
Run Growth:
19
Growth Rate:
16.81%
Total runs:
112
Run Growth:
90
Growth Rate:
80.36%
Total runs:
112
Run Growth:
46
Growth Rate:
41.07%
Total runs:
111
Run Growth:
31
Growth Rate:
27.93%
Total runs:
110
Run Growth:
51
Growth Rate:
46.36%
Total runs:
106
Run Growth:
71
Growth Rate:
66.36%
Total runs:
104
Run Growth:
71
Growth Rate:
67.62%
Total runs:
103
Run Growth:
60
Growth Rate:
57.69%
Total runs:
101
Run Growth:
66
Growth Rate:
62.86%
Total runs:
100
Run Growth:
8
Growth Rate:
8.00%
Total runs:
99
Run Growth:
-106
Growth Rate:
-107.07%
Total runs:
94
Run Growth:
-18
Growth Rate:
-19.15%
Total runs:
92
Run Growth:
68
Growth Rate:
73.91%
Total runs:
87
Run Growth:
-77
Growth Rate:
-88.51%
Total runs:
85
Run Growth:
-75
Growth Rate:
-88.24%
Total runs:
82
Run Growth:
-40
Growth Rate:
-48.78%
Total runs:
82
Run Growth:
57
Growth Rate:
69.51%
Total runs:
81
Run Growth:
73
Growth Rate:
90.12%
Total runs:
79
Run Growth:
51
Growth Rate:
64.56%
Total runs:
76
Run Growth:
47
Growth Rate:
61.84%
Total runs:
74
Run Growth:
-8
Growth Rate:
-10.26%
Total runs:
73
Run Growth:
31
Growth Rate:
42.47%
Total runs:
67
Run Growth:
29
Growth Rate:
43.28%
Total runs:
66
Run Growth:
52
Growth Rate:
78.79%
Total runs:
65
Run Growth:
24
Growth Rate:
36.92%
Total runs:
63
Run Growth:
31
Growth Rate:
49.21%
Total runs:
60
Run Growth:
60
Growth Rate:
100.00%
Total runs:
58
Run Growth:
-59
Growth Rate:
-103.51%
Total runs:
57
Run Growth:
31
Growth Rate:
51.67%
Total runs:
57
Run Growth:
-27
Growth Rate:
-47.37%
Total runs:
56
Run Growth:
9
Growth Rate:
15.52%
Total runs:
56
Run Growth:
29
Growth Rate:
51.79%
Total runs:
55
Run Growth:
47
Growth Rate:
85.45%
Total runs:
55
Run Growth:
-3
Growth Rate:
-5.36%
Total runs:
54
Run Growth:
-101
Growth Rate:
-174.14%
Total runs:
54
Run Growth:
40
Growth Rate:
71.43%