refactor, fix checkpoint compute for tuple and add tests

apache · Oct 24, 2019 · 19b68f3 · 19b68f3
1 parent 7bc21ca
commit 19b68f3
Show file tree

Hide file tree

Showing 4 changed files with 88 additions and 23 deletions.
diff --git a/src/relay/op/annotation/annotation.cc b/src/relay/op/annotation/annotation.cc
@@ -164,7 +164,11 @@ Mark a checkpoint for checkpointing memory optimization.
 .set_attr<FTVMCompute>("FTVMCompute",
                        [](const Attrs& attrs, const Array<Tensor>& inputs,
                           const Type& out_dtype, const Target& target) -> Array<Tensor> {
-                         return {topi::identity(inputs[0])};
+                         Array<Tensor> outputs;
+                         for (size_t i = 0; i < inputs.size(); ++i) {
+                           outputs.push_back(topi::identity(inputs[i]));
+                         }
+                         return outputs;
                        });
 
 }  // namespace relay

diff --git a/src/relay/pass/gradient.cc b/src/relay/pass/gradient.cc
@@ -400,33 +400,41 @@ struct ReverseAD : ExprMutator {
     throw;
   }
 
+  Expr VisitCheckpoint(const CallNode *call) {
+    const OpNode* op_node = call->op.as<OpNode>();
+    CHECK(op_node) << "expected op in call";
+    Op op_ref = GetRef<Op>(op_node);
+    CHECK(op_ref->name == "annotation.checkpoint") << "expected checkpoint annotation";
+    auto x = call->args[0];
+    return LetList::With([&](LetList* ll) {
+      auto x_var = ll->Push(x);
+      auto ret = ll->Push(GetRev(call->checked_type(), x_var, ll));
+      auto bpv = ll->Push(RefReadNode::make(bp));
+      Expr nbp = FunctionNode::make(
+        {},
+        LetList::With([&](LetList* ll) {
+          // we need a new ReverseAD visitor to avoid clobbering the bp local var
+          auto dup_bp = ll->Push(BPEmpty());
+          ReverseAD dup_diff(dup_bp, ad_vars);
+          auto dup_ad = ll->Push(dup_diff.VisitExpr(DeDup(x)));
+
+          TransferGrads(call->checked_type(), ret, dup_ad, ll);
+          ll->Push(CallNode::make(RefReadNode::make(dup_bp), {}));
+          return CallNode::make(bpv, {});
+        }),
+        TupleTypeNode::make({}),
+        {});
+      ll->Push(RefWriteNode::make(bp, nbp));
+      return ret;
+    });
+  }
+
   Expr VisitExpr_(const CallNode* call) final {
     if (const OpNode* op_node = call->op.as<OpNode>()) {
       Op op_ref = GetRef<Op>(op_node);
 
       if (op_ref->name == "annotation.checkpoint") {
-        auto x = call->args[0];
-        return LetList::With([&](LetList* ll) {
-          auto x_var = ll->Push(x);
-          auto ret = ll->Push(GetRev(call->checked_type(), x_var, ll));
-          auto bpv = ll->Push(RefReadNode::make(bp));
-          Expr nbp = FunctionNode::make(
-            {},
-            LetList::With([&](LetList* ll) {
-              // we need a new ReverseAD visitor to avoid clobbering the bp local var
-              auto dup_bp = ll->Push(BPEmpty());
-              ReverseAD dup_diff(dup_bp, ad_vars);
-              auto dup_ad = ll->Push(dup_diff.VisitExpr(DeDup(x)));
-
-              TransferGrads(call->checked_type(), ret, dup_ad, ll);
-              ll->Push(CallNode::make(RefReadNode::make(dup_bp), {}));
-              return CallNode::make(bpv, {});
-            }),
-            TupleTypeNode::make({}),
-            {});
-          ll->Push(RefWriteNode::make(bp, nbp));
-          return ret;
-        });
+        return VisitCheckpoint(call);
       }
 
       CHECK(rev_map.count(op_ref))

diff --git a/tests/python/relay/test_op_grad_level10.py b/tests/python/relay/test_op_grad_level10.py
@@ -30,6 +30,12 @@ def test_checkpoint():
                             relay.add(inputs[2], inputs[3]))
     check_grad(relay.Function(inputs, relay.annotation.checkpoint(output)))
 
+    out_tuple = relay.Tuple([relay.add(inputs[0], inputs[1]),
+                             relay.multiply(inputs[2], inputs[3])])
+    out_single = relay.subtract(relay.TupleGetItem(relay.annotation.checkpoint(out_tuple), 0),
+                                relay.TupleGetItem(out_tuple, 1))
+    check_grad(relay.Function(inputs, out_single))
+
 
 if __name__ == "__main__":
     test_cross_entropy_grad()
diff --git a/tests/python/relay/test_op_level10.py b/tests/python/relay/test_op_level10.py
@@ -105,6 +105,53 @@ def test_checkpoint_alpha_equal():
 
     relay.analysis.assert_alpha_equal(df, df_parsed)
 
+def test_checkpoint_alpha_equal_tuple():
+    xs = [relay.var("x{}".format(i), relay.TensorType((1,), "float32")) for i in range(4)]
+    f = relay.Function(xs, relay.annotation.checkpoint(
+        relay.Tuple([relay.add(xs[0], xs[1]), relay.add(xs[2], xs[3])])
+    ))
+    df = transform.gradient(run_infer_type(f))
+
+    # run PE and DCE
+    with transform.PassContext(opt_level=3):
+        passes = [transform.PartialEvaluate(),
+                  transform.DeadCodeElimination(inline_once=True)]
+        mod = transform.Sequential(passes)(relay.Module.from_expr(df))
+        df = mod["main"]
+
+    df_parsed = relay.parser.fromtext(
+        """
+        v0.0.4
+        fn (%x: Tensor[(1), float32], %y: Tensor[(1), float32],
+            %z: Tensor[(1), float32], %w: Tensor[(1), float32])
+            -> ((Tensor[(1), float32], Tensor[(1), float32]),
+                (Tensor[(1), float32], Tensor[(1), float32],
+                 Tensor[(1), float32], Tensor[(1), float32])) {
+        let %x1: Tensor[(1), float32] = add(%x, %y) /* ty=Tensor[(1), float32] */;
+        let %x2: Tensor[(1), float32] = add(%z, %w) /* ty=Tensor[(1), float32] */;
+        let %x3: Tensor[(1), float32] = zeros_like(%x2) /* ty=Tensor[(1), float32] */;
+        let %x4: Tensor[(1), float32] = ones_like(%x1) /* ty=Tensor[(1), float32] */;
+        %0 = (%x1, %x2);
+        %1 = zeros_like(%x) /* ty=Tensor[(1), float32] */;
+        %2 = collapse_sum_like(%x4, %x) /* ty=Tensor[(1), float32] */;
+        %3 = add(%1, %2) /* ty=Tensor[(1), float32] */;
+        %4 = zeros_like(%y) /* ty=Tensor[(1), float32] */;
+        %5 = collapse_sum_like(%x4, %y) /* ty=Tensor[(1), float32] */;
+        %6 = add(%4, %5) /* ty=Tensor[(1), float32] */;
+        %7 = zeros_like(%z) /* ty=Tensor[(1), float32] */;
+        %8 = collapse_sum_like(%x3, %z) /* ty=Tensor[(1), float32] */;
+        %9 = add(%7, %8) /* ty=Tensor[(1), float32] */;
+        %10 = zeros_like(%w) /* ty=Tensor[(1), float32] */;
+        %11 = collapse_sum_like(%x3, %w) /* ty=Tensor[(1), float32] */;
+        %12 = add(%10, %11) /* ty=Tensor[(1), float32] */;
+        %13 = (%3, %6, %9, %12);
+        (%0, %13)
+        }
+        """
+    )
+
+    relay.analysis.assert_alpha_equal(df, df_parsed)
+
 def test_collapse_sum_like():
     shape = (3, 4, 5, 6)
     shape_like = (4, 5, 6)