[GraphBolt][CUDA] Make dataloader pickleable.

dmlc · May 10, 2024 · 727ed54 · 727ed54
1 parent 7a2e873
commit 727ed54
Show file tree

Hide file tree

Showing 3 changed files with 22 additions and 12 deletions.
diff --git a/examples/sampling/graphbolt/link_prediction.py b/examples/sampling/graphbolt/link_prediction.py
@@ -328,8 +328,6 @@ def train(args, model, graph, features, train_set):
 
             total_loss += loss.item()
             if step + 1 == args.early_stop:
-                # Early stopping requires a new dataloader to reset its state.
-                dataloader = create_dataloader(args, graph, features, train_set)
                 break
 
         end_epoch_time = time.time()

diff --git a/notebooks/stochastic_training/link_prediction.ipynb b/notebooks/stochastic_training/link_prediction.ipynb
@@ -129,14 +129,13 @@
       "outputs": [],
       "source": [
         "from functools import partial\n",
-        "def create_train_dataloader():\n",
-        "    datapipe = gb.ItemSampler(train_set, batch_size=256, shuffle=True)\n",
-        "    datapipe = datapipe.copy_to(device)\n",
-        "    datapipe = datapipe.sample_uniform_negative(graph, 5)\n",
-        "    datapipe = datapipe.sample_neighbor(graph, [5, 5])\n",
-        "    datapipe = datapipe.transform(partial(gb.exclude_seed_edges, include_reverse_edges=True))\n",
-        "    datapipe = datapipe.fetch_feature(feature, node_feature_keys=[\"feat\"])\n",
-        "    return gb.DataLoader(datapipe)"
+        "datapipe = gb.ItemSampler(train_set, batch_size=256, shuffle=True)\n",
+        "datapipe = datapipe.copy_to(device)\n",
+        "datapipe = datapipe.sample_uniform_negative(graph, 5)\n",
+        "datapipe = datapipe.sample_neighbor(graph, [5, 5])\n",
+        "datapipe = datapipe.transform(partial(gb.exclude_seed_edges, include_reverse_edges=True))\n",
+        "datapipe = datapipe.fetch_feature(feature, node_feature_keys=[\"feat\"])\n",
+        "train_dataloader = gb.DataLoader(datapipe)"
       ]
     },
     {
@@ -157,7 +156,7 @@
       },
       "outputs": [],
       "source": [
-        "data = next(iter(create_train_dataloader()))\n",
+        "data = next(iter(train_dataloader))\n",
         "print(f\"MiniBatch: {data}\")"
       ]
     },
@@ -253,7 +252,7 @@
         "for epoch in range(3):\n",
         "    model.train()\n",
         "    total_loss = 0\n",
-        "    for step, data in tqdm(enumerate(create_train_dataloader())):\n",
+        "    for step, data in tqdm(enumerate(train_dataloader)):\n",
         "        # Get node pairs with labels for loss calculation.\n",
         "        compacted_seeds = data.compacted_seeds.T\n",
         "        labels = data.labels\n",

diff --git a/python/dgl/graphbolt/base.py b/python/dgl/graphbolt/base.py
@@ -313,6 +313,19 @@ def __iter__(self):
         while len(self.buffer) > 0:
             yield self.buffer.popleft()
 
+    def __getstate__(self):
+        state = (self.datapipe, self.buffer.maxlen)
+        if IterDataPipe.getstate_hook is not None:
+            return IterDataPipe.getstate_hook(state)
+        return state
+
+    def __setstate__(self, state):
+        self.datapipe, buffer_size = state
+        self.buffer = deque(maxlen=buffer_size)
+
+    def reset(self):
+        self.buffer.clear()
+
 
 @functional_datapipe("wait")
 class Waiter(IterDataPipe):