sjrhuschlee commited on
Commit
9f7b3ec
1 Parent(s): 3368621

Update README.md

Browse files
Files changed (1) hide show
  1. README.md +5 -2
README.md CHANGED
@@ -141,6 +141,8 @@ model-index:
141
 
142
  This is the [flan-t5-base](https://huggingface.co/google/flan-t5-base) model, fine-tuned using the [SQuAD2.0](https://huggingface.co/datasets/squad_v2) dataset. It's been trained on question-answer pairs, including unanswerable questions, for the task of Extractive Question Answering.
143
 
 
 
144
  **NOTE:** The `<cls>` token must be manually added to the beginning of the question for this model to work properly.
145
  It uses the `<cls>` token to be able to make "no answer" predictions.
146
  The t5 tokenizer does not automatically add this special token which is why it is added manually.
@@ -168,7 +170,7 @@ nlp = pipeline(
168
  'question-answering',
169
  model=model_name,
170
  tokenizer=model_name,
171
- trust_remote_code=True,
172
  )
173
  qa_input = {
174
  'question': f'{nlp.tokenizer.cls_token}Where do I live?', # '<cls>Where do I live?'
@@ -179,7 +181,8 @@ res = nlp(qa_input)
179
 
180
  # b) Load model & tokenizer
181
  model = AutoModelForQuestionAnswering.from_pretrained(
182
- model_name, trust_remote_code=True
 
183
  )
184
  tokenizer = AutoTokenizer.from_pretrained(model_name)
185
 
 
141
 
142
  This is the [flan-t5-base](https://huggingface.co/google/flan-t5-base) model, fine-tuned using the [SQuAD2.0](https://huggingface.co/datasets/squad_v2) dataset. It's been trained on question-answer pairs, including unanswerable questions, for the task of Extractive Question Answering.
143
 
144
+ **UPDATE:** With transformers version 4.31.0 the `use_remote_code=True` is no longer necessary and will cause `AutoModelForQuestionAnswering.from_pretrained()` to fail.
145
+
146
  **NOTE:** The `<cls>` token must be manually added to the beginning of the question for this model to work properly.
147
  It uses the `<cls>` token to be able to make "no answer" predictions.
148
  The t5 tokenizer does not automatically add this special token which is why it is added manually.
 
170
  'question-answering',
171
  model=model_name,
172
  tokenizer=model_name,
173
+ # trust_remote_code=True, # Do not use if version transformers>=4.31.0
174
  )
175
  qa_input = {
176
  'question': f'{nlp.tokenizer.cls_token}Where do I live?', # '<cls>Where do I live?'
 
181
 
182
  # b) Load model & tokenizer
183
  model = AutoModelForQuestionAnswering.from_pretrained(
184
+ model_name,
185
+ # trust_remote_code=True # Do not use if version transformers>=4.31.0
186
  )
187
  tokenizer = AutoTokenizer.from_pretrained(model_name)
188