From 7b50a223656973253db97fd64b843f7c22c95fc4 Mon Sep 17 00:00:00 2001
From: Maxim Melichov <80150303+maxmelichov@users.noreply.github.com>
Date: Sun, 2 Mar 2025 13:48:48 +0200
Subject: [PATCH] Update utils.py

Fixing the following bugs:

- AttributeError: 'Tensor' object has no attribute 'is_integer'
- AttributeError: 'float' object has no attribute 'item'

These errors occur when attempting to convert the model from a PyTorch (.pt) file to an ONNX (.onnx) file.
---
 fairseq/models/wav2vec/utils.py | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/fairseq/models/wav2vec/utils.py b/fairseq/models/wav2vec/utils.py
index dd52d86242..6845c17ba8 100644
--- a/fairseq/models/wav2vec/utils.py
+++ b/fairseq/models/wav2vec/utils.py
@@ -5,6 +5,7 @@
 
 import math
 import torch.nn.functional as F
+import torch
 
 
 def pad_to_multiple(x, multiple, dim=-1, value=0):
@@ -14,7 +15,9 @@ def pad_to_multiple(x, multiple, dim=-1, value=0):
     tsz = x.size(dim)
     m = tsz / multiple
     remainder = math.ceil(m) * multiple - tsz
-    if m.is_integer():
+    if isinstance(m, torch.Tensor):
+        m = m.item()  # Convert tensor to Python float
+    if isinstance(m, float) and m.is_integer():
         return x, 0
     pad_offset = (0,) * (-1 - dim) * 2