Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

implement review suggestions #1062

Merged
merged 3 commits into from
Apr 26, 2023
Merged

implement review suggestions #1062

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
3 commits
Select commit Hold shift + click to select a range
74ad6e9
implement review suggestions
radekosmulski Apr 19, 2023
6621a89
update
radekosmulski Apr 19, 2023
5ff0765
fix test
radekosmulski Apr 19, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
80 changes: 41 additions & 39 deletions examples/usecases/transformers-next-item-prediction.ipynb
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -89,7 +89,7 @@
"name": "stderr",
"output_type": "stream",
"text": [
"2023-04-17 10:53:46.419725: I tensorflow/core/platform/cpu_feature_guard.cc:194] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: SSE3 SSE4.1 SSE4.2 AVX\n",
"2023-04-19 01:35:02.207465: I tensorflow/core/platform/cpu_feature_guard.cc:194] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: SSE3 SSE4.1 SSE4.2 AVX\n",
"To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n"
]
},
Expand All @@ -106,9 +106,9 @@
"text": [
"/usr/local/lib/python3.8/dist-packages/merlin/dtypes/mappings/torch.py:43: UserWarning: PyTorch dtype mappings did not load successfully due to an error: No module named 'torch'\n",
" warn(f\"PyTorch dtype mappings did not load successfully due to an error: {exc.msg}\")\n",
"2023-04-17 10:53:47.662982: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-17 10:53:47.663430: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-17 10:53:47.663612: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n"
"2023-04-19 01:35:03.422817: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-19 01:35:03.423265: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-19 01:35:03.423447: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n"
]
},
{
Expand All @@ -125,17 +125,17 @@
"name": "stderr",
"output_type": "stream",
"text": [
"2023-04-17 10:53:47.910504: I tensorflow/core/platform/cpu_feature_guard.cc:194] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: SSE3 SSE4.1 SSE4.2 AVX\n",
"2023-04-19 01:35:03.695996: I tensorflow/core/platform/cpu_feature_guard.cc:194] This TensorFlow binary is optimized with oneAPI Deep Neural Network Library (oneDNN) to use the following CPU instructions in performance-critical operations: SSE3 SSE4.1 SSE4.2 AVX\n",
"To enable them in other operations, rebuild TensorFlow with the appropriate compiler flags.\n",
"2023-04-17 10:53:47.911325: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-17 10:53:47.911536: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-17 10:53:47.911695: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-17 10:53:48.671264: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-17 10:53:48.671486: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-17 10:53:48.671654: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-17 10:53:48.671765: W tensorflow/core/common_runtime/gpu/gpu_bfc_allocator.cc:42] Overriding orig_value setting because the TF_FORCE_GPU_ALLOW_GROWTH environment variable is set. Original config value was 0.\n",
"2023-04-17 10:53:48.671774: I tensorflow/core/common_runtime/gpu/gpu_process_state.cc:222] Using CUDA malloc Async allocator for GPU: 0\n",
"2023-04-17 10:53:48.671836: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1637] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 24576 MB memory: -> device: 0, name: Quadro RTX 8000, pci bus id: 0000:08:00.0, compute capability: 7.5\n",
"2023-04-19 01:35:03.696824: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-19 01:35:03.697038: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-19 01:35:03.697196: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-19 01:35:04.438845: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-19 01:35:04.439063: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-19 01:35:04.439222: I tensorflow/stream_executor/cuda/cuda_gpu_executor.cc:996] successful NUMA node read from SysFS had negative value (-1), but there must be at least one NUMA node, so returning NUMA node zero\n",
"2023-04-19 01:35:04.439332: W tensorflow/core/common_runtime/gpu/gpu_bfc_allocator.cc:42] Overriding orig_value setting because the TF_FORCE_GPU_ALLOW_GROWTH environment variable is set. Original config value was 0.\n",
"2023-04-19 01:35:04.439340: I tensorflow/core/common_runtime/gpu/gpu_process_state.cc:222] Using CUDA malloc Async allocator for GPU: 0\n",
"2023-04-19 01:35:04.439400: I tensorflow/core/common_runtime/gpu/gpu_device.cc:1637] Created device /job:localhost/replica:0/task:0/device:GPU:0 with 24576 MB memory: -> device: 0, name: Quadro RTX 8000, pci bus id: 0000:08:00.0, compute capability: 7.5\n",
"/usr/local/lib/python3.8/dist-packages/tqdm/auto.py:22: TqdmWarning: IProgress not found. Please update jupyter and ipywidgets. See https://ipywidgets.readthedocs.io/en/stable/user_install.html\n",
" from .autonotebook import tqdm as notebook_tqdm\n"
]
Expand Down Expand Up @@ -859,7 +859,7 @@
"text": [
"/usr/local/lib/python3.8/dist-packages/keras/initializers/initializers_v2.py:120: UserWarning: The initializer TruncatedNormal is unseeded and being called multiple times, which will return identical values each time (even if the initializer is unseeded). Please update your code to provide a seed to the initializer, or avoid using the same initalizer instance more than once.\n",
" warnings.warn(\n",
"2023-04-17 10:53:56.422412: I tensorflow/stream_executor/cuda/cuda_dnn.cc:424] Loaded cuDNN version 8700\n"
"2023-04-19 01:35:13.112516: I tensorflow/stream_executor/cuda/cuda_dnn.cc:424] Loaded cuDNN version 8700\n"
]
},
{
Expand Down Expand Up @@ -893,28 +893,28 @@
"name": "stderr",
"output_type": "stream",
"text": [
"2023-04-17 10:54:07.156255: W tensorflow/core/grappler/optimizers/loop_optimizer.cc:907] Skipping loop optimization for Merge node with control input: model/xl_net_block/sequential_block_5/replace_masked_embeddings/RaggedWhere/Assert/AssertGuard/branch_executed/_95\n"
"2023-04-19 01:35:23.514177: W tensorflow/core/grappler/optimizers/loop_optimizer.cc:907] Skipping loop optimization for Merge node with control input: model/xl_net_block/sequential_block_5/replace_masked_embeddings/RaggedWhere/Assert/AssertGuard/branch_executed/_95\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"2720/2720 [==============================] - 79s 25ms/step - loss: 7.3254 - recall_at_10: 0.1956 - mrr_at_10: 0.0829 - ndcg_at_10: 0.1092 - map_at_10: 0.0829 - precision_at_10: 0.0196 - regularization_loss: 0.0000e+00 - loss_batch: 7.3244\n",
"2720/2720 [==============================] - 77s 24ms/step - loss: 7.3177 - recall_at_10: 0.2005 - mrr_at_10: 0.0877 - ndcg_at_10: 0.1141 - map_at_10: 0.0877 - precision_at_10: 0.0201 - regularization_loss: 0.0000e+00 - loss_batch: 7.3157\n",
"Epoch 2/5\n",
"2720/2720 [==============================] - 67s 24ms/step - loss: 6.1170 - recall_at_10: 0.3607 - mrr_at_10: 0.1683 - ndcg_at_10: 0.2137 - map_at_10: 0.1683 - precision_at_10: 0.0361 - regularization_loss: 0.0000e+00 - loss_batch: 6.1153\n",
"2720/2720 [==============================] - 67s 24ms/step - loss: 6.1038 - recall_at_10: 0.3626 - mrr_at_10: 0.1703 - ndcg_at_10: 0.2157 - map_at_10: 0.1703 - precision_at_10: 0.0363 - regularization_loss: 0.0000e+00 - loss_batch: 6.1022\n",
"Epoch 3/5\n",
"2720/2720 [==============================] - 68s 25ms/step - loss: 5.5579 - recall_at_10: 0.4363 - mrr_at_10: 0.2080 - ndcg_at_10: 0.2620 - map_at_10: 0.2080 - precision_at_10: 0.0436 - regularization_loss: 0.0000e+00 - loss_batch: 5.5554\n",
"2720/2720 [==============================] - 63s 23ms/step - loss: 5.5767 - recall_at_10: 0.4321 - mrr_at_10: 0.2067 - ndcg_at_10: 0.2600 - map_at_10: 0.2067 - precision_at_10: 0.0432 - regularization_loss: 0.0000e+00 - loss_batch: 5.5751\n",
"Epoch 4/5\n",
"2720/2720 [==============================] - 69s 25ms/step - loss: 5.3061 - recall_at_10: 0.4666 - mrr_at_10: 0.2246 - ndcg_at_10: 0.2818 - map_at_10: 0.2246 - precision_at_10: 0.0467 - regularization_loss: 0.0000e+00 - loss_batch: 5.3038\n",
"2720/2720 [==============================] - 64s 23ms/step - loss: 5.3249 - recall_at_10: 0.4631 - mrr_at_10: 0.2223 - ndcg_at_10: 0.2792 - map_at_10: 0.2223 - precision_at_10: 0.0463 - regularization_loss: 0.0000e+00 - loss_batch: 5.3230\n",
"Epoch 5/5\n",
"2720/2720 [==============================] - 70s 25ms/step - loss: 5.1669 - recall_at_10: 0.4821 - mrr_at_10: 0.2327 - ndcg_at_10: 0.2917 - map_at_10: 0.2327 - precision_at_10: 0.0482 - regularization_loss: 0.0000e+00 - loss_batch: 5.1647\n"
"2720/2720 [==============================] - 68s 25ms/step - loss: 5.1779 - recall_at_10: 0.4796 - mrr_at_10: 0.2318 - ndcg_at_10: 0.2904 - map_at_10: 0.2318 - precision_at_10: 0.0480 - regularization_loss: 0.0000e+00 - loss_batch: 5.1764\n"
]
},
{
"data": {
"text/plain": [
"<keras.callbacks.History at 0x7fb47895b0a0>"
"<keras.callbacks.History at 0x7fa9a45e20a0>"
]
},
"execution_count": 15,
Expand Down Expand Up @@ -961,27 +961,27 @@
"name": "stderr",
"output_type": "stream",
"text": [
"2023-04-17 10:59:53.461458: W tensorflow/core/grappler/optimizers/loop_optimizer.cc:907] Skipping loop optimization for Merge node with control input: model/xl_net_block/sequential_block_5/replace_masked_embeddings/RaggedWhere/Assert/AssertGuard/branch_executed/_71\n"
"2023-04-19 01:40:55.868086: W tensorflow/core/grappler/optimizers/loop_optimizer.cc:907] Skipping loop optimization for Merge node with control input: model/xl_net_block/sequential_block_5/replace_masked_embeddings/RaggedWhere/Assert/AssertGuard/branch_executed/_71\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"340/340 [==============================] - 11s 20ms/step - loss: 4.7125 - recall_at_10: 0.5556 - mrr_at_10: 0.3126 - ndcg_at_10: 0.3704 - map_at_10: 0.3126 - precision_at_10: 0.0556 - regularization_loss: 0.0000e+00 - loss_batch: 4.7121\n"
"340/340 [==============================] - 10s 18ms/step - loss: 4.7317 - recall_at_10: 0.5562 - mrr_at_10: 0.3114 - ndcg_at_10: 0.3695 - map_at_10: 0.3114 - precision_at_10: 0.0556 - regularization_loss: 0.0000e+00 - loss_batch: 4.7313\n"
]
},
{
"data": {
"text/plain": [
"{'loss': 4.712502956390381,\n",
" 'recall_at_10': 0.5555964112281799,\n",
" 'mrr_at_10': 0.3125925064086914,\n",
" 'ndcg_at_10': 0.37040114402770996,\n",
" 'map_at_10': 0.3125925064086914,\n",
" 'precision_at_10': 0.05555963143706322,\n",
"{'loss': 4.731719017028809,\n",
" 'recall_at_10': 0.556216835975647,\n",
" 'mrr_at_10': 0.3114089369773865,\n",
" 'ndcg_at_10': 0.36954960227012634,\n",
" 'map_at_10': 0.3114089369773865,\n",
" 'precision_at_10': 0.0556216724216938,\n",
" 'regularization_loss': 0.0,\n",
" 'loss_batch': 4.561119079589844}"
" 'loss_batch': 4.579279899597168}"
]
},
"execution_count": 16,
Expand Down Expand Up @@ -1091,15 +1091,15 @@
"name": "stdout",
"output_type": "stream",
"text": [
"INFO:tensorflow:Assets written to: /tmp/tmpvdxq5w0m/model.savedmodel/assets\n"
"INFO:tensorflow:Assets written to: /tmp/tmpk2jcmm5x/model.savedmodel/assets\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"INFO:tensorflow:Assets written to: /tmp/tmpvdxq5w0m/model.savedmodel/assets\n",
"/usr/local/lib/python3.8/dist-packages/merlin/models/tf/utils/tf_utils.py:100: CustomMaskWarning: Custom mask layers require a config and must override get_config. When loading, the custom mask layer must be passed to the custom_objects argument.\n",
"INFO:tensorflow:Assets written to: /tmp/tmpk2jcmm5x/model.savedmodel/assets\n",
"/usr/local/lib/python3.8/dist-packages/merlin/models/tf/utils/tf_utils.py:101: CustomMaskWarning: Custom mask layers require a config and must override get_config. When loading, the custom mask layer must be passed to the custom_objects argument.\n",
" config[key] = tf.keras.utils.serialize_keras_object(maybe_value)\n",
"/usr/local/lib/python3.8/dist-packages/merlin/models/tf/core/combinators.py:288: CustomMaskWarning: Custom mask layers require a config and must override get_config. When loading, the custom mask layer must be passed to the custom_objects argument.\n",
" config[i] = tf.keras.utils.serialize_keras_object(layer)\n",
Expand Down Expand Up @@ -1286,7 +1286,7 @@
"output_type": "stream",
"text": [
"INFO:tensorflow:Assets written to: /workspace/ensemble/1_predicttensorflowtriton/1/model.savedmodel/assets\n",
"/usr/local/lib/python3.8/dist-packages/merlin/models/tf/utils/tf_utils.py:100: CustomMaskWarning: Custom mask layers require a config and must override get_config. When loading, the custom mask layer must be passed to the custom_objects argument.\n",
"/usr/local/lib/python3.8/dist-packages/merlin/models/tf/utils/tf_utils.py:101: CustomMaskWarning: Custom mask layers require a config and must override get_config. When loading, the custom mask layer must be passed to the custom_objects argument.\n",
" config[key] = tf.keras.utils.serialize_keras_object(maybe_value)\n",
"/usr/local/lib/python3.8/dist-packages/merlin/models/tf/core/combinators.py:288: CustomMaskWarning: Custom mask layers require a config and must override get_config. When loading, the custom mask layer must be passed to the custom_objects argument.\n",
" config[i] = tf.keras.utils.serialize_keras_object(layer)\n",
Expand Down Expand Up @@ -1397,8 +1397,8 @@
{
"data": {
"text/plain": [
"array([[-3.497796 , -3.5179672 , 2.9564662 , ..., -0.45221543,\n",
" -0.8475621 , -1.217337 ]], dtype=float32)"
"array([[-4.958611 , -4.977809 , 0.66012955, ..., -2.32572 ,\n",
" -0.6495375 , -2.4545593 ]], dtype=float32)"
]
},
"execution_count": 20,
Expand Down Expand Up @@ -1428,7 +1428,8 @@
}
],
"source": [
"response.as_numpy('city_id_list/categorical_output').shape"
"predictions = response.as_numpy('city_id_list/categorical_output')\n",
"predictions.shape"
]
},
{
Expand Down Expand Up @@ -1457,7 +1458,8 @@
}
],
"source": [
"wf.output_schema.select_by_name('city_id_list').to_pandas()['properties.embedding_sizes.cardinality'][0]"
"cardinality = wf.output_schema['city_id_list'].properties['embedding_sizes']['cardinality']\n",
"cardinality"
]
},
{
Expand Down
23 changes: 17 additions & 6 deletions tests/unit/tf/examples/test_usecase_transformers_next_item_prediction.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,9 +18,9 @@
execute=False,
)
@pytest.mark.notebook
def test_next_item_prediction(tb):
def test_next_item_prediction(tb, tmpdir):
tb.inject(
"""
f"""
import os, random
from datetime import datetime, timedelta
from merlin.datasets.synthetic import generate_data
Expand All @@ -33,16 +33,27 @@ def generate_date():
return date
df['checkin'] = [generate_date() for _ in range(df.shape[0])]
df['checkout'] = [generate_date() for _ in range(df.shape[0])]
df.to_csv('/tmp/train_set.csv')
df.to_csv('{tmpdir}/train_set.csv')
"""
)
tb.cells[4].source = tb.cells[4].source.replace("get_booking('/workspace/data')", "")
tb.cells[4].source = tb.cells[4].source.replace(
"read_csv('/workspace/data/train_set.csv'", "read_csv('/tmp/train_set.csv'"
"read_csv('/workspace/data/train_set.csv'", f"read_csv('{tmpdir}/train_set.csv'"
)
tb.cells[31].source = tb.cells[31].source.replace("epochs=5", "epochs=1")
tb.cells[37].source = tb.cells[37].source.replace("/workspace/ensemble", "/tmp/ensemble")
tb.cells[37].source = tb.cells[37].source.replace("/workspace/ensemble", f"{tmpdir}/ensemble")
tb.execute_cell(list(range(0, 38)))

with utils.run_triton_server("/tmp/ensemble", grpc_port=8001):
with utils.run_triton_server(f"{tmpdir}/ensemble", grpc_port=8001):
tb.execute_cell(list(range(38, len(tb.cells))))

tb.inject(
"""
logits_count = predictions.shape[1]
"""
)
tb.execute_cell(len(tb.cells) - 1)

cardinality = tb.ref("cardinality")
logits_count = tb.ref("logits_count")
assert logits_count == cardinality