ICM fixes (4/n)

onnxruntime/contrib_ops/cpu/word_conv_embedding.cc

@@ -1,6 +1,8 @@
 // Copyright (c) Microsoft Corporation. All rights reserved.
 // Licensed under the MIT License.
 
+#include <cstring>
+
 #include <core/common/safeint.h>
 #include "word_conv_embedding.h"
 
@@ -14,6 +16,7 @@ namespace contrib {
 void WordConvEmbedding::CharEmbeddingLookup(
     const int* seq_ptr,
     const float* char_embedding_weight_p,
+    size_t char_embedding_table_size,
     size_t seq_len,
     size_t word_len,
     size_t char_embedding_size,
@@ -26,7 +29,14 @@ void WordConvEmbedding::CharEmbeddingLookup(
       float* cur_dst_ptr = dst + word_inx * word_len * char_embedding_size;
       size_t char_length_to_lookup = std::max<size_t>(words_len_ptr[word_inx], filter_width);
       for (size_t char_inx = 0; char_inx < char_length_to_lookup; char_inx++) {
-        memcpy(cur_dst_ptr, char_embedding_weight_p + (*cur_seq_ptr) * char_embedding_size, sizeof(float) * char_embedding_size);
+        const int char_index = *cur_seq_ptr;
+        if (char_index >= 0 && static_cast<size_t>(char_index) < char_embedding_table_size) {
+          memcpy(cur_dst_ptr,
+                 char_embedding_weight_p + static_cast<size_t>(char_index) * char_embedding_size,
+                 sizeof(float) * char_embedding_size);
+        } else {
+          std::memset(cur_dst_ptr, 0, sizeof(float) * char_embedding_size);
+        }
         cur_dst_ptr += char_embedding_size;
         cur_seq_ptr++;
       }
@@ -198,6 +208,7 @@ Status WordConvEmbedding::Compute(OpKernelContext* ctx) const {
 
   CharEmbeddingLookup(seq_ptr,
                       w_char_embedding.Data<float>(),
+                      onnxruntime::narrow<size_t>(w_char_embedding_shape[0]),
                       onnxruntime::narrow<size_t>(seq_len),
                       onnxruntime::narrow<size_t>(word_len),
                       onnxruntime::narrow<size_t>(char_embedding_size),

onnxruntime/contrib_ops/cpu/word_conv_embedding.h

@@ -26,6 +26,7 @@ class WordConvEmbedding final : public OpKernel {
   void CharEmbeddingLookup(
       const int* seq_ptr,
       const float* char_embedding_weight_p,
+      size_t char_embedding_table_size,
       size_t seq_len,
       size_t word_len,
       size_t char_embedding_size,

onnxruntime/test/contrib_ops/word_conv_embedding_test.cc

@@ -1,6 +1,7 @@
 // Copyright (c) Microsoft Corporation. All rights reserved.
 // Licensed under the MIT License.
 
+#include <cmath>
 #include <vector>
 #include "gtest/gtest.h"
 #include "test/providers/provider_test_utils.h"
@@ -126,5 +127,21 @@ TEST(ContribOpTest, WordConvEmbedding_char_embedding_shape_conv_shape_not_match)
   test.Run(OpTester::ExpectResult::kExpectFailure);
 }
 
+TEST(ContribOpTest, WordConvEmbedding_out_of_range_char_index_treated_as_padding) {
+  OpTester test("WordConvEmbedding", 1, onnxruntime::kMSDomain);
+
+  test.AddAttribute<int64_t>("embedding_size", 1LL);
+  test.AddAttribute<int64_t>("conv_window_size", 2LL);
+  test.AddAttribute<int64_t>("char_embedding_size", 1LL);
+
+  test.AddInput<int>("Sequence", {1, 2}, {1, 99});
+  test.AddInput<float>("W", {1, 1, 2, 1}, {1.0f, 1.0f});
+  test.AddInput<float>("B", {1}, {0.0f});
+  test.AddInput<float>("C", {2, 1}, {123.0f, 2.0f});
+  test.AddOutput<float>("Y", {1, 1}, {std::tanh(2.0f)});
+
+  test.Run();
+}
+
 }  // namespace test
 }  // namespace onnxruntime