fix'

Author: xadupre

onnx_diagnostic/tasks/image_text_to_text.py

@@ -172,10 +172,10 @@ def _check_():
         assert expected & set(
             dummies
         ), f"Unable to find expected inputs {expected} in loaded inputs {set(dummies)}"
-        assert sequence_length == dummies["input_ids"].shape[-1], (
-            f"sequence_length={sequence_length} != {dummies['input_ids'].shape[-1]} for "
-            f"model class {model.__class__.__name__}"
-        )
+        # assert sequence_length == dummies["input_ids"].shape[-1], (
+        #    f"sequence_length={sequence_length} != {dummies['input_ids'].shape[-1]} for "
+        #    f"model class {model.__class__.__name__}"
+        # )
         assert batch_size == dummies["input_ids"].shape[0], (
             f"batch_size={batch_size} != {dummies['input_ids'].shape[0]} for "
             f"model class {model.__class__.__name__}"

onnx_diagnostic/torch_export_patches/patches/_patch_transformers_rotary_embedding.py

@@ -214,7 +214,7 @@ def longrope_frequency_update(self, position_ids, device, layer_type=None):
             cond,
             (lambda x, y: x.clone()),
             (lambda x, y: y.clone()),
-            [long_inv_freq, original_inv_freq],
+            [long_inv_freq.to(original_inv_freq.dtype), original_inv_freq],
         )
         setattr(self, f"{prefix}inv_freq", inv_freq)
         # if seq_len > original_max_position_embeddings:
@@ -293,7 +293,7 @@ def dynamic_frequency_update(self, position_ids, device, layer_type=None):
             cond,
             (lambda x, y: x.clone()),
             (lambda x, y: y.clone()),
-            [long_inv_freq, original_inv_freq],
+            [long_inv_freq.to(original_inv_freq.dtype), original_inv_freq],
         )
         setattr(self, f"{prefix}inv_freq", inv_freq)