File size: 6,473 Bytes
efb03ac |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 |
from typing import Any, Dict, List, Optional, Text
import tensorflow_model_analysis as tfma
from tfx import v1 as tfx
from ml_metadata.proto import metadata_store_pb2
from tfx.proto import example_gen_pb2
import absl
import tensorflow_model_analysis as tfma
from tfx.components import Evaluator
from tfx.components import ExampleValidator
from tfx.components import ImportExampleGen
from tfx.components import Pusher
from tfx.components import SchemaGen
from tfx.components import StatisticsGen
from tfx.components import Trainer
from tfx.components import Tuner
from tfx.extensions.google_cloud_ai_platform.trainer.component import (
Trainer as VertexTrainer,
)
from tfx.extensions.google_cloud_ai_platform.pusher.component import (
Pusher as VertexPusher,
)
from tfx.components import Transform
from tfx.dsl.components.common import resolver
from tfx.dsl.experimental import latest_blessed_model_resolver
from tfx.orchestration import pipeline
from tfx.proto import example_gen_pb2
from tfx.proto import trainer_pb2
from tfx.types import Channel
from tfx.types.standard_artifacts import Model
from tfx.types.standard_artifacts import ModelBlessing
from tfx.orchestration.data_types import RuntimeParameter
from components.pusher.GHReleasePusher.component import Pusher as GHPusher
def create_pipeline(
pipeline_name: Text,
pipeline_root: Text,
data_path: Text,
modules: Dict[Text, Text],
train_args: trainer_pb2.TrainArgs,
eval_args: trainer_pb2.EvalArgs,
serving_model_dir: Text,
metadata_connection_config: Optional[metadata_store_pb2.ConnectionConfig] = None,
) -> tfx.dsl.Pipeline:
components = []
input_config = example_gen_pb2.Input(
splits=[
example_gen_pb2.Input.Split(name="train", pattern="train/*.tfrecord"),
example_gen_pb2.Input.Split(name="eval", pattern="test/*.tfrecord"),
]
)
example_gen = ImportExampleGen(input_base=data_path, input_config=input_config)
components.append(example_gen)
statistics_gen = StatisticsGen(examples=example_gen.outputs["examples"])
components.append(statistics_gen)
schema_gen = SchemaGen(
statistics=statistics_gen.outputs["statistics"], infer_feature_shape=True
)
components.append(schema_gen)
# example_validator = tfx.components.ExampleValidator(
# statistics=statistics_gen.outputs['statistics'],
# schema=schema_gen.outputs['schema'])
# components.append(example_validator)
transform = Transform(
examples=example_gen.outputs["examples"],
schema=schema_gen.outputs["schema"],
preprocessing_fn=modules["preprocessing_fn"],
)
components.append(transform)
tuner = Tuner(
tuner_fn=modules["tuner_fn"],
examples=transform.outputs["transformed_examples"],
schema=schema_gen.outputs["schema"],
transform_graph=transform.outputs["transform_graph"],
train_args=train_args,
eval_args=eval_args,
)
components.append(tuner)
trainer_args = {
"run_fn": modules["training_fn"],
"transformed_examples": transform.outputs["transformed_examples"],
"schema": schema_gen.outputs["schema"],
"hyperparameters": tuner.outputs["best_hyperparameters"],
"transform_graph": transform.outputs["transform_graph"],
"train_args": train_args,
"eval_args": eval_args,
}
trainer = Trainer(**trainer_args)
components.append(trainer)
model_resolver = resolver.Resolver(
strategy_class=latest_blessed_model_resolver.LatestBlessedModelResolver,
model=Channel(type=Model),
model_blessing=Channel(type=ModelBlessing),
).with_id("latest_blessed_model_resolver")
components.append(model_resolver)
# Uses TFMA to compute evaluation statistics over features of a model and
# perform quality validation of a candidate model (compare to a baseline).
eval_config = tfma.EvalConfig(
model_specs=[tfma.ModelSpec(label_key="label_xf", prediction_key="label_xf")],
slicing_specs=[tfma.SlicingSpec()],
metrics_specs=[
tfma.MetricsSpec(
metrics=[
tfma.MetricConfig(
class_name="SparseCategoricalAccuracy",
threshold=tfma.MetricThreshold(
value_threshold=tfma.GenericValueThreshold(
lower_bound={"value": 0.55}
),
# Change threshold will be ignored if there is no
# baseline model resolved from MLMD (first run).
change_threshold=tfma.GenericChangeThreshold(
direction=tfma.MetricDirection.HIGHER_IS_BETTER,
absolute={"value": -1e-3},
),
),
)
]
)
],
)
evaluator = Evaluator(
examples=transform.outputs["transformed_examples"],
model=trainer.outputs["model"],
baseline_model=model_resolver.outputs["model"],
eval_config=eval_config,
)
components.append(evaluator)
pusher_args = {
"model": trainer.outputs["model"],
"model_blessing": evaluator.outputs["blessing"],
"push_destination": tfx.proto.PushDestination(
filesystem=tfx.proto.PushDestination.Filesystem(
base_directory=serving_model_dir
)
),
}
pusher = Pusher(**pusher_args) # pylint: disable=unused-variable
components.append(pusher)
pusher_args = {
"model": trainer.outputs["model"],
"model_blessing": evaluator.outputs["blessing"],
"custom_config": {
"GH_RELEASE": {
"ACCESS_TOKEN": "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",
"USERNAME": "deep-diver",
"REPONAME": "PyGithubTest",
"BRANCH": "main",
"ASSETNAME": "saved_model.tar.gz",
}
},
}
gh_pusher = GHPusher(**pusher_args).with_id("gh_release_pusher")
components.append(gh_pusher)
return pipeline.Pipeline(
pipeline_name=pipeline_name,
pipeline_root=pipeline_root,
components=components,
enable_cache=False,
metadata_connection_config=metadata_connection_config,
)
|