Fix group_lengths for short datasets (#12558)

examples/flax/language-modeling/run_clm_flax.py

@@ -398,6 +398,7 @@ def main():
         total_length = len(concatenated_examples[list(examples.keys())[0]])
         # We drop the small remainder, we could add padding if the model supported it instead of this drop, you can
         # customize this part to your needs.
+        if total_length >= block_size:
             total_length = (total_length // block_size) * block_size
         # Split by chunks of max_len.
         result = {

examples/flax/language-modeling/run_mlm_flax.py

@@ -431,6 +431,7 @@ if __name__ == "__main__":
             total_length = len(concatenated_examples[list(examples.keys())[0]])
             # We drop the small remainder, we could add padding if the model supported it instead of this drop, you can
             # customize this part to your needs.
+            if total_length >= max_seq_length:
                 total_length = (total_length // max_seq_length) * max_seq_length
             # Split by chunks of max_len.
             result = {

examples/flax/language-modeling/run_t5_mlm_flax.py

@@ -541,6 +541,7 @@ if __name__ == "__main__":
         total_length = len(concatenated_examples[list(examples.keys())[0]])
         # We drop the small remainder, we could add padding if the model supported it instead of this drop, you can
         # customize this part to your needs.
+        if total_length >= expanded_inputs_length:
             total_length = (total_length // expanded_inputs_length) * expanded_inputs_length
         # Split by chunks of max_len.
         result = {

examples/pytorch/language-modeling/run_clm.py

@@ -404,6 +404,7 @@ def main():
         total_length = len(concatenated_examples[list(examples.keys())[0]])
         # We drop the small remainder, we could add padding if the model supported it instead of this drop, you can
         # customize this part to your needs.
+        if total_length >= block_size:
             total_length = (total_length // block_size) * block_size
         # Split by chunks of max_len.
         result = {

examples/pytorch/language-modeling/run_clm_no_trainer.py

@@ -343,6 +343,7 @@ def main():
         total_length = len(concatenated_examples[list(examples.keys())[0]])
         # We drop the small remainder, we could add padding if the model supported it instead of this drop, you can
         # customize this part to your needs.
+        if total_length >= block_size:
             total_length = (total_length // block_size) * block_size
         # Split by chunks of max_len.
         result = {

examples/pytorch/language-modeling/run_mlm.py

@@ -433,6 +433,7 @@ def main():
             total_length = len(concatenated_examples[list(examples.keys())[0]])
             # We drop the small remainder, we could add padding if the model supported it instead of this drop, you can
             # customize this part to your needs.
+            if total_length >= max_seq_length:
                 total_length = (total_length // max_seq_length) * max_seq_length
             # Split by chunks of max_len.
             result = {

examples/pytorch/language-modeling/run_mlm_no_trainer.py

@@ -387,6 +387,7 @@ def main():
             total_length = len(concatenated_examples[list(examples.keys())[0]])
             # We drop the small remainder, we could add padding if the model supported it instead of this drop, you can
             # customize this part to your needs.
+            if total_length >= max_seq_length:
                 total_length = (total_length // max_seq_length) * max_seq_length
             # Split by chunks of max_len.
             result = {

examples/pytorch/language-modeling/run_plm.py

@@ -406,6 +406,7 @@ def main():
             total_length = len(concatenated_examples[list(examples.keys())[0]])
             # We drop the small remainder, we could add padding if the model supported it instead of this drop, you can
             # customize this part to your needs.
+            if total_length >= max_seq_length:
                 total_length = (total_length // max_seq_length) * max_seq_length
             # Split by chunks of max_len.
             result = {

examples/tensorflow/language-modeling/run_clm.py

@@ -405,6 +405,7 @@ def main():
         total_length = len(concatenated_examples[list(examples.keys())[0]])
         # We drop the small remainder, we could add padding if the model supported it instead of this drop, you can
         # customize this part to your needs.
+        if total_length >= block_size:
             total_length = (total_length // block_size) * block_size
         # Split by chunks of max_len.
         result = {

examples/tensorflow/language-modeling/run_mlm.py

@@ -466,6 +466,7 @@ def main():
             total_length = len(concatenated_examples[list(examples.keys())[0]])
             # We drop the small remainder, we could add padding if the model supported it instead of this drop, you can
             # customize this part to your needs.
+            if total_length >= max_seq_length:
                 total_length = (total_length // max_seq_length) * max_seq_length
             # Split by chunks of max_len.
             result = {