refine

oneflow/user/ops/fused_group_norm_min_max_observer_op.cpp

@@ -37,26 +37,26 @@ namespace oneflow {
   // check input shape
   CHECK_GT_OR_RETURN(x.shape().NumAxes(), 2)
       << "fused_group_norm_min_max_observer: input should have at least 2 dims, got "
-      << x.shape().NumAxes() << " dims";
+      << x.shape().NumAxes() << " dims.";
 
   // check num_group
   int32_t num_groups = ctx->Attr<int32_t>("num_groups");
-  CHECK_EQ_OR_RETURN(channel_size % num_groups, 0) << "Channels should be divisble by num_groups. ";
+  CHECK_EQ_OR_RETURN(channel_size % num_groups, 0) << "fused_group_norm_min_max_observer: Channels should be divisble by num_groups.";
 
   // check gamma and beta size
   if (affine) {
     const user_op::TensorDesc& gamma = ctx->InputTensorDesc("gamma", 0);
     CHECK_EQ_OR_RETURN(gamma.shape().elem_cnt(), channel_size)
-        << "The size of `gamma` must equal to channel_size, expected " << channel_size
-        << " but got " << gamma.shape().elem_cnt();
+        << "fused_group_norm_min_max_observer: The size of `gamma` must be equal to channel_size, expected " << channel_size
+        << " but got " << gamma.shape().elem_cnt() << ".";
     const user_op::TensorDesc& beta = ctx->InputTensorDesc("beta", 0);
     CHECK_EQ_OR_RETURN(beta.shape().elem_cnt(), channel_size)
-        << "The size of `beta` must equal to channel_size, expected " << channel_size << " but got "
-        << beta.shape().elem_cnt();
+        << "fused_group_norm_min_max_observer: The size of `beta` must be equal to channel_size, expected " << channel_size << " but got "
+        << beta.shape().elem_cnt() << ".";
   }
 
   CHECK_OR_RETURN(ctx->Attr<bool>("per_layer_quantization"))
-      << "dynamic quantization only supports per-layer quantization";
+      << "fused_group_norm_min_max_observer: dynamic quantization only supports per-layer quantization.";
   ctx->SetOutputShape("y", 0, x.shape());
   ctx->SetOutputShape("y_scale", 0, Shape({1}));
   ctx->SetOutputShape("y_zero_point", 0, Shape({1}));