Update lm_finetuning.py

lm_finetuning.py

@@ -2,8 +2,12 @@
 wandb offline
 export WANDB_DISABLED='true'
 export RAY_RESULTS='ray_results'
-python lm_finetuning.py -m "roberta-large" -c "ckpt/roberta_large" --push-to-hub --hf-organization "cardiffnlp" -a "roberta-large-tweet-topic-multi" --split-train "train_all" --split-valid "validation_2021" --split-test "test_2021"
-python lm_finetuning.py -m "roberta-base" -c "ckpt/roberta_base" --push-to-hub --hf-organization "cardiffnlp" -a "twitter-roberta-base-2019-90m-tweet-topic-multi"
+python lm_finetuning.py -m "roberta-large" -o "ckpt/roberta_large" --push-to-hub --hf-organization "cardiffnlp" -a "roberta-large-tweet-topic-multi-all" --split-train "train_all" --split-valid "validation_2021" --split-test "test_2021"
+python lm_finetuning.py -m "roberta-large" -o "ckpt/roberta_large" --push-to-hub --hf-organization "cardiffnlp" -a "roberta-large-tweet-topic-multi-2020" --split-train "train_2020" --split-valid "validation_2020" --split-test "test_2021"
+
+
+python lm_finetuning.py -m "roberta-base" -c "ckpt/roberta_base" --push-to-hub --hf-organization "cardiffnlp" -a "roberta-large-tweet-topic-multi" --split-train "train_all" --split-valid "validation_2021" --split-test "test_2021"
+
 python lm_finetuning.py -m "cardiffnlp/twitter-roberta-base-2019-90m" -c "ckpt/twitter-roberta-base-2019-90m" --push-to-hub --hf-organization "cardiffnlp" -a "twitter-roberta-base-2019-90m-tweet-topic-multi"
 python lm_finetuning.py -m "cardiffnlp/twitter-roberta-base-dec2020" -c "ckpt/twitter-roberta-base-dec2020"
 python lm_finetuning.py -m "cardiffnlp/twitter-roberta-base-dec2021" -c "ckpt/twitter-roberta-base-dec2021"
@@ -90,7 +94,7 @@ def main():
     tokenizer = AutoTokenizer.from_pretrained(opt.model, local_files_only=not network)
     model = AutoModelForSequenceClassification.from_pretrained(
         opt.model,
-        num_labels=len(dataset[opt.split_train][0]['label']),
+        num_labels=len(dataset[opt.split_train]['label'][0]),
         local_files_only=not network,
         problem_type="multi_label_classification"
     )
@@ -114,7 +118,7 @@ def main():
             eval_dataset=tokenized_datasets[opt.split_validation],
             compute_metrics=compute_metric_search,
             model_init=lambda x: AutoModelForSequenceClassification.from_pretrained(
-                opt.model, return_dict=True, num_labels=dataset[opt.split_train].features['label'].num_classes)
+                opt.model, return_dict=True, num_labels=len(data[opt.split_train]['label'][0]))
         )
         # parameter search
         if PARALLEL:
@@ -149,7 +153,7 @@ def main():
     # evaluation
     model = AutoModelForSequenceClassification.from_pretrained(
         best_model_path,
-        num_labels=dataset[opt.split_train].features['label'].num_classes,
+        num_labels=len(data[opt.split_train]['label'][0]),
         local_files_only=not network)
     trainer = Trainer(
         model=model,
@@ -162,7 +166,7 @@ def main():
         eval_dataset=tokenized_datasets[opt.split_test],
         compute_metrics=compute_metric_all,
         model_init=lambda x: AutoModelForSequenceClassification.from_pretrained(
-            opt.model, return_dict=True, num_labels=dataset[opt.split_train].features['label'].num_classes)
+            opt.model, return_dict=True, num_labels=len(data[opt.split_train]['label'][0]))
     )
     summary_file = pj(opt.output_dir, opt.summary_file)
     if not opt.skip_eval: