Fix group_lengths for short datasets (#12558)

examples/flax/language-modeling/run_clm_flax.py

@@ -398,7 +398,8 @@ 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.
-        total_length = (total_length // block_size) * block_size
+        if total_length >= block_size:
+            total_length = (total_length // block_size) * block_size
         # Split by chunks of max_len.
         result = {
             k: [t[i : i + block_size] for i in range(0, total_length, block_size)]

examples/flax/language-modeling/run_mlm_flax.py

@@ -431,7 +431,8 @@ 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.
-            total_length = (total_length // max_seq_length) * max_seq_length
+            if total_length >= max_seq_length:
+                total_length = (total_length // max_seq_length) * max_seq_length
             # Split by chunks of max_len.
             result = {
                 k: [t[i : i + max_seq_length] for i in range(0, total_length, max_seq_length)]

examples/flax/language-modeling/run_t5_mlm_flax.py

@@ -541,7 +541,8 @@ 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.
-        total_length = (total_length // expanded_inputs_length) * expanded_inputs_length
+        if total_length >= expanded_inputs_length:
+            total_length = (total_length // expanded_inputs_length) * expanded_inputs_length
         # Split by chunks of max_len.
         result = {
             k: [t[i : i + expanded_inputs_length] for i in range(0, total_length, expanded_inputs_length)]

examples/pytorch/language-modeling/run_clm.py

@@ -404,7 +404,8 @@ 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.
-        total_length = (total_length // block_size) * block_size
+        if total_length >= block_size:
+            total_length = (total_length // block_size) * block_size
         # Split by chunks of max_len.
         result = {
             k: [t[i : i + block_size] for i in range(0, total_length, block_size)]

examples/pytorch/language-modeling/run_clm_no_trainer.py

@@ -343,7 +343,8 @@ 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.
-        total_length = (total_length // block_size) * block_size
+        if total_length >= block_size:
+            total_length = (total_length // block_size) * block_size
         # Split by chunks of max_len.
         result = {
             k: [t[i : i + block_size] for i in range(0, total_length, block_size)]

examples/pytorch/language-modeling/run_mlm.py

@@ -433,7 +433,8 @@ 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.
-            total_length = (total_length // max_seq_length) * max_seq_length
+            if total_length >= max_seq_length:
+                total_length = (total_length // max_seq_length) * max_seq_length
             # Split by chunks of max_len.
             result = {
                 k: [t[i : i + max_seq_length] for i in range(0, total_length, max_seq_length)]

examples/pytorch/language-modeling/run_mlm_no_trainer.py

@@ -387,7 +387,8 @@ 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.
-            total_length = (total_length // max_seq_length) * max_seq_length
+            if total_length >= max_seq_length:
+                total_length = (total_length // max_seq_length) * max_seq_length
             # Split by chunks of max_len.
             result = {
                 k: [t[i : i + max_seq_length] for i in range(0, total_length, max_seq_length)]

examples/pytorch/language-modeling/run_plm.py

@@ -406,7 +406,8 @@ 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.
-            total_length = (total_length // max_seq_length) * max_seq_length
+            if total_length >= max_seq_length:
+                total_length = (total_length // max_seq_length) * max_seq_length
             # Split by chunks of max_len.
             result = {
                 k: [t[i : i + max_seq_length] for i in range(0, total_length, max_seq_length)]

examples/tensorflow/language-modeling/run_clm.py

@@ -405,7 +405,8 @@ 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.
-        total_length = (total_length // block_size) * block_size
+        if total_length >= block_size:
+            total_length = (total_length // block_size) * block_size
         # Split by chunks of max_len.
         result = {
             k: [t[i : i + block_size] for i in range(0, total_length, block_size)]

examples/tensorflow/language-modeling/run_mlm.py

@@ -466,7 +466,8 @@ 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.
-            total_length = (total_length // max_seq_length) * max_seq_length
+            if total_length >= max_seq_length:
+                total_length = (total_length // max_seq_length) * max_seq_length
             # Split by chunks of max_len.
             result = {
                 k: [t[i : i + max_seq_length] for i in range(0, total_length, max_seq_length)]