From d024048d74c8faa8a64548dd7938531e93815c31 Mon Sep 17 00:00:00 2001
From: Dan Saunders <danjsaund@gmail.com>
Date: Wed, 17 Sep 2025 14:50:49 -0400
Subject: [PATCH] logs + fix

---
 src/axolotl/kernels/moe/torch_grouped.py | 27 +++++++++++++++---------
 1 file changed, 17 insertions(+), 10 deletions(-)

diff --git a/src/axolotl/kernels/moe/torch_grouped.py b/src/axolotl/kernels/moe/torch_grouped.py
index f44684abd..8f1b149fe 100644
--- a/src/axolotl/kernels/moe/torch_grouped.py
+++ b/src/axolotl/kernels/moe/torch_grouped.py
@@ -249,9 +249,9 @@ def moe_ffn_forward_grouped(
     for i in range(E):
         sel = flat_idx == i
         if sel.any():
-            Xi = x_rep[sel]
+            Xi = x_rep[sel].contiguous()
             As.append(Xi)
-            Bs.append(W13[i])
+            Bs.append(W13[i].contiguous())
             expert_slices.append((i, sel))
 
     if not As:
@@ -263,13 +263,20 @@ def moe_ffn_forward_grouped(
         b_tensors: List[torch.Tensor],
         target_dtype: torch.dtype,
     ) -> Optional[List[torch.Tensor]]:
-        if target_dtype != dt:
-            a_tensors = [t.to(target_dtype) for t in a_tensors]
-            b_tensors = [t.to(target_dtype) for t in b_tensors]
-        outputs = _call_grouped_mm(a_tensors, b_tensors)
-        if outputs is not None and target_dtype != dt:
-            outputs = [t.to(dt) for t in outputs]
-        return outputs
+        global LAST_ERROR
+        try:
+            if target_dtype != dt:
+                a_tensors = [t.to(target_dtype).contiguous() for t in a_tensors]
+                b_tensors = [t.to(target_dtype).contiguous() for t in b_tensors]
+            outputs = _call_grouped_mm(a_tensors, b_tensors)
+            if outputs is not None and target_dtype != dt:
+                outputs = [t.to(dt).contiguous() for t in outputs]
+            return outputs
+        except RuntimeError as err:
+            LAST_ERROR = f"grouped_mm cast failure: {err}"  # type: ignore[assignment]
+            if torch.cuda.is_available():  # pragma: no cover - defensive
+                torch.cuda.synchronize()
+            return None
 
     def _try_grouped_mm(
         a_tensors: List[torch.Tensor], b_tensors: List[torch.Tensor]
@@ -305,7 +312,7 @@ def moe_ffn_forward_grouped(
         I2 = Yi.shape[-1] // 2
         Yi_hidden = F.silu(Yi[:, :I2]) * Yi[:, I2:]
         As2.append(Yi_hidden)
-        Bs2.append(W2[i])
+        Bs2.append(W2[i].contiguous())
 
     Y2_list, _cast_used_down = _try_grouped_mm(As2, Bs2)
     if Y2_list is None: