Rename embedding variable to weight (#9720)

Summary:
I renamed the variable in the `Embedding` module from `weight` to `table` a few months ago, because it seemed like a more meaningful name. Turns out it's not such a good idea because it deviates from PyTorch, which unnecessarily breaks C++->Python translated code.

ebetica ezyang apaszke
Pull Request resolved: https://github.com/pytorch/pytorch/pull/9720

Differential Revision: D8955647

Pulled By: goldsborough

fbshipit-source-id: 77228b07d2b733866e8cdecaa6d0686eef4cc3ea
This commit is contained in:
Peter Goldsborough
2018-07-23 14:43:25 -07:00
committed by Facebook Github Bot
parent 431415adc4
commit 31ba2f15e1
3 changed files with 12 additions and 7 deletions
+7 -2
View File
@@ -142,8 +142,13 @@ TEST_CASE("modules") {
SECTION("embedding") {
SECTION("basic") {
int dict_size = 10;
const int64_t dict_size = 10;
Embedding model(dict_size, 2);
REQUIRE(model->parameters().contains("weight"));
REQUIRE(model->weight.ndimension() == 2);
REQUIRE(model->weight.size(0) == dict_size);
REQUIRE(model->weight.size(1) == 2);
// Cannot get gradients to change indices (input) - only for embedding
// params
auto x = torch::full({10}, dict_size - 1, torch::kInt64);
@@ -156,7 +161,7 @@ TEST_CASE("modules") {
REQUIRE(y.size(0) == 10);
REQUIRE(y.size(1) == 2);
REQUIRE(model->parameters()["table"].grad().numel() == 2 * dict_size);
REQUIRE(model->parameters()["weight"].grad().numel() == 2 * dict_size);
}
SECTION("list") {