antypasd commited on
Commit
92be862
1 Parent(s): 5e8f546

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +13 -2
README.md CHANGED
@@ -1,6 +1,6 @@
1
  # tweet-topic-21-multi
2
 
3
- This is a roBERTa-base model trained on ~124M tweets from January 2018 to December 2021 (see [here](https://huggingface.co/cardiffnlp/twitter-roberta-base-2021-124m)), and finetuned for single-label topic classification on a corpus of 11,267 tweets.
4
  The original roBERTa-base model can be found [here](https://huggingface.co/cardiffnlp/twitter-roberta-base-2021-124m) and the original reference paper is [TweetEval](https://github.com/cardiffnlp/tweeteval). This model is suitable for English.
5
 
6
  - Reference Paper: [TimeLMs paper](https://arxiv.org/abs/2202.03829).
@@ -20,7 +20,7 @@ The original roBERTa-base model can be found [here](https://huggingface.co/cardi
20
  ## Full classification example
21
 
22
  ```python
23
- from transformers import AutoModelForSequenceClassification
24
  from transformers import AutoTokenizer
25
  import numpy as np
26
  from scipy.special import expit
@@ -41,6 +41,17 @@ scores = output[0][0].detach().numpy()
41
  scores = expit(scores)
42
  predictions = (scores >= 0.5) * 1
43
 
 
 
 
 
 
 
 
 
 
 
 
44
  # Map to classes
45
  for i in range(len(predictions)):
46
  if predictions[i]:
 
1
  # tweet-topic-21-multi
2
 
3
+ This is a roBERTa-base model trained on ~124M tweets from January 2018 to December 2021 (see [here](https://huggingface.co/cardiffnlp/twitter-roberta-base-2021-124m)), and finetuned for multi-label topic classification on a corpus of 11,267 tweets.
4
  The original roBERTa-base model can be found [here](https://huggingface.co/cardiffnlp/twitter-roberta-base-2021-124m) and the original reference paper is [TweetEval](https://github.com/cardiffnlp/tweeteval). This model is suitable for English.
5
 
6
  - Reference Paper: [TimeLMs paper](https://arxiv.org/abs/2202.03829).
 
20
  ## Full classification example
21
 
22
  ```python
23
+ from transformers import AutoModelForSequenceClassification, TFAutoModelForSequenceClassification
24
  from transformers import AutoTokenizer
25
  import numpy as np
26
  from scipy.special import expit
 
41
  scores = expit(scores)
42
  predictions = (scores >= 0.5) * 1
43
 
44
+
45
+ # TF
46
+ #tf_model = TFAutoModelForSequenceClassification.from_pretrained(MODEL)
47
+ #class_mapping = model.config.id2label
48
+ #text = "It is great to see athletes promoting awareness for climate change."
49
+ #tokens = tokenizer(text, return_tensors='tf')
50
+ #output = tf_model(**tokens)
51
+ #scores = output[0][0]
52
+ #scores = expit(scores)
53
+ #predictions = (scores >= 0.5) * 1
54
+
55
  # Map to classes
56
  for i in range(len(predictions)):
57
  if predictions[i]: