init

experiments/model_finetuning_emoji.py

@@ -95,7 +95,7 @@ def main(
         validate_index = validate_index[:N_VALIDATE_SIZE]
 
         trainer = Trainer(
-            model=AutoModelForSequenceClassification.from_pretrained(model, num_labels=100),
+            model=AutoModelForSequenceClassification.from_pretrained(model, num_labels=99),
             args=TrainingArguments(
                 output_dir=output_dir,
                 evaluation_strategy="steps",
@@ -106,7 +106,7 @@ def main(
             eval_dataset=tokenized_datasets["validation"].select(validate_index),
             compute_metrics=compute_metric,
             model_init=lambda x: AutoModelForSequenceClassification.from_pretrained(
-                model, return_dict=True, num_labels=100,
+                model, return_dict=True, num_labels=99,
             )
         )
 
@@ -141,7 +141,7 @@ def main(
             metric = {}
         for single_test in test_split:
             trainer = Trainer(
-                model=AutoModelForSequenceClassification.from_pretrained(best_model_path, num_labels=100),
+                model=AutoModelForSequenceClassification.from_pretrained(best_model_path, num_labels=99),
                 args=TrainingArguments(
                     output_dir=output_dir,
                     evaluation_strategy="no",
@@ -159,7 +159,7 @@ def main(
     if not skip_upload:
         logging.info("uploading to huggingface")
         model_organization = "tweettemposhift"
-        model_instance = AutoModelForSequenceClassification.from_pretrained(best_model_path, num_labels=100)
+        model_instance = AutoModelForSequenceClassification.from_pretrained(best_model_path, num_labels=99)
         model_instance.push_to_hub(f"{model_organization}/{model_alias}", use_auth_token=True)
         tokenizer.push_to_hub(f"{model_organization}/{model_alias}", use_auth_token=True)
         repo = Repository(model_alias, f"{model_organization}/{model_alias}")

tweet_temporal_shift.py

@@ -2,7 +2,7 @@
 import json
 import datasets
 
-_VERSION = "1.1.1"
+_VERSION = "1.1.2"
 _TWEET_TEMPORAL_DESCRIPTION = """"""
 _TWEET_TEMPORAL_CITATION = """"""
 _TWEET_TOPIC_DESCRIPTION = """
@@ -232,6 +232,7 @@ class TweetTemporalShift(datasets.GeneratorBasedBuilder):
             with open(dl_manager.download(url_map)) as f:
                 label_classes = f.readlines()
             label_classes = [x.strip('\n') for x in label_classes]
+            label_classes = [x for n, x in enumerate(label_classes) if n != 68]
             features['gold_label'] = datasets.features.ClassLabel(names=label_classes)
         elif "sentiment" in self.config.name:
             features["text"] = datasets.Value("string")