diff --git a/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp b/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
--- a/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
+++ b/mlir/lib/Dialect/Tosa/IR/TosaOps.cpp
@@ -100,6 +100,14 @@
 // TOSA Operator Verifiers.
 //===----------------------------------------------------------------------===//
 
+static bool haveZeroDimensions(ShapedType shapedType) {
+  if (!shapedType.hasStaticShape())
+    return false;
+
+  return llvm::count_if(shapedType.getShape(),
+                        [](int64_t dim) { return dim == 0; });
+}
+
 template <typename T> static LogicalResult verifyConvOp(T op) {
   // All TOSA conv ops have an input() and weight().
   auto inputType = llvm::dyn_cast<RankedTensorType>(op.getInput().getType());
@@ -115,6 +123,9 @@
     return failure();
   }
 
+  if (haveZeroDimensions(inputType))
+    return op.emitOpError() << "cannot take tensor with zero dimensions";
+
   auto inputEType = inputType.getElementType();
   auto weightEType = weightType.getElementType();
 
@@ -142,7 +153,11 @@
 }
 
 LogicalResult tosa::AvgPool2dOp::verify() {
-  auto inputETy = llvm::cast<ShapedType>(getInput().getType()).getElementType();
+  auto inputType = llvm::cast<ShapedType>(getInput().getType());
+  if (haveZeroDimensions(inputType))
+    return emitOpError() << "cannot take tensor with zero dimensions";
+
+  auto inputETy = inputType.getElementType();
   auto resultETy = llvm::cast<ShapedType>(getType()).getElementType();
 
   if (auto quantType =
@@ -765,6 +780,9 @@
       return emitOpError() << "Cannot reshape " << inputElementsNum
                            << " elements into " << outputElementsNum;
     }
+
+    if (haveZeroDimensions(inputType) || haveZeroDimensions(outputType))
+      return emitOpError() << "cannot take tensor with zero dimensions";
   }
   return mlir::success();
 }
diff --git a/mlir/test/Dialect/Tosa/invalid.mlir b/mlir/test/Dialect/Tosa/invalid.mlir
--- a/mlir/test/Dialect/Tosa/invalid.mlir
+++ b/mlir/test/Dialect/Tosa/invalid.mlir
@@ -150,3 +150,29 @@
   %0 = "tosa.const"() {value = dense<0.000000e+00> : tensor<1x1xf32>} : () -> tensor<100x100xf32>
   return %0 : tensor<100x100xf32>
 }
+
+// -----
+
+func.func @test_reshape_zero_dim_input(%arg0 : tensor<13x0x3xf32>) -> () {
+  // expected-error@+1 {{'tosa.reshape' op cannot take tensor with zero dimensions}}
+  %0 = "tosa.reshape"(%arg0) {new_shape = array<i64: 13, 21, 3>} : (tensor<13x0x3xf32>) -> tensor<13x0x3xf32>
+  return
+}
+
+// -----
+
+func.func @test_conv2d_zero_dim_input(%arg0: tensor<1x29x0x4xf32>, %arg1: tensor<16x3x3x4xf32>, %arg2: tensor<16xf32>) -> tensor<1x27x27x16xf32> {
+  // expected-error@+1 {{'tosa.conv2d' op cannot take tensor with zero dimensions}}
+  %0 = "tosa.conv2d"(%arg0, %arg1, %arg2) {dilation = array<i64: 1, 1>, pad = array<i64: 0, 0, 0, 0>, stride = array<i64: 1, 1>}
+           : (tensor<1x29x0x4xf32>, tensor<16x3x3x4xf32>, tensor<16xf32>) -> tensor<1x27x27x16xf32>
+  return %0 : tensor<1x27x27x16xf32>
+}
+
+// -----
+
+func.func @test_avg_pool2d_zero_dim_input(%arg0: tensor<1x0x7x9xf32>) -> tensor<1x7x7x9xf32> {
+  // expected-error@+1 {{'tosa.avg_pool2d' op cannot take tensor with zero dimensions}}  
+    %0 = "tosa.avg_pool2d"(%arg0) {acc_type = f32, kernel = array<i64: 2, 2>, pad = array<i64: 0, 1, 0, 1>, stride = array<i64: 1, 1>} 
+      : (tensor<1x0x7x9xf32>) -> tensor<1x7x7x9xf32>
+    return %0 : tensor<1x7x7x9xf32>
+}
\ No newline at end of file