Demo for visualize

.github/workflows/python-package-genai.yml

@@ -39,7 +39,7 @@ jobs:
       fail-fast: false
       matrix:
         os: ["ubuntu-22.04"]
-        python-version: ["3.8", "3.10"]
+        python-version: ["3.10"]
 
     steps:
     - uses: actions/checkout@v3

genai-perf/README.md

@@ -89,7 +89,7 @@ genai-perf --help
 Since GenAI-Perf depends on Perf Analyzer,
 you'll need to install the Perf Analyzer binary:
 
-### Install Perf Analyzer (Ubuntu, Python 3.8+)
+### Install Perf Analyzer (Ubuntu, Python 3.10+)
 
 **NOTE**: you must already have CUDA 12 installed
 (checkout the [CUDA installation guide](https://docs.nvidia.com/cuda/cuda-installation-guide-linux/index.html)).

genai-perf/genai_perf/checkpoint/checkpoint.py

@@ -0,0 +1,98 @@
+# Copyright 2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+import json
+import os
+from dataclasses import dataclass
+
+from genai_perf.config.input.config_command import ConfigCommand
+from genai_perf.config.run.results import Results
+from genai_perf.exceptions import GenAIPerfException
+from genai_perf.types import CheckpointObject
+
+
+@dataclass(frozen=True)
+class CheckpointDefaults:
+    FILENAME = "checkpoint.json"
+
+
+@dataclass
+class Checkpoint:
+    """
+    Contains the methods necessary for reading and writing GenAI-Perf
+    state to a file so that stateful subcommands (such as Optimize or
+    Analyze) can resume or continue (ex: running Analyze then Visualize)
+    """
+
+    config: ConfigCommand
+
+    # Every top-level class that needs to store state is passed in
+    results: Results
+
+    def __post_init__(self):
+        self._create_class_from_checkpoint()
+
+    ###########################################################################
+    # Read/Write Methods
+    ###########################################################################
+    def create_checkpoint_object(self) -> None:
+        state_dict = {"Results": self.results.create_checkpoint_object()}
+
+        checkpoint_file_path = self._create_checkpoint_file_path()
+        with open(checkpoint_file_path, "w") as checkpoint_file:
+            json.dump(state_dict, checkpoint_file, default=checkpoint_encoder)
+
+    def _create_class_from_checkpoint(self) -> None:
+        checkpoint_file_path = self._create_checkpoint_file_path()
+
+        if os.path.isfile(checkpoint_file_path):
+            self.checkpoint_exists = True
+            try:
+                with open(checkpoint_file_path, "r") as checkpoint_file:
+                    checkpoint_json = json.load(checkpoint_file)
+                    self._state: CheckpointObject = {
+                        "Results": Results.create_class_from_checkpoint(
+                            checkpoint_json["Results"]
+                        )
+                    }
+
+            except EOFError:
+                raise (
+                    GenAIPerfException(
+                        f"Checkpoint file {checkpoint_file} is"
+                        " empty or corrupted. Delete it and rerun GAP"
+                    )
+                )
+        else:
+            self.checkpoint_exists = False
+            self._state = {}
+
+    def _create_checkpoint_file_path(self) -> str:
+        checkpoint_file_path = os.path.join(
+            self.config.checkpoint_directory, CheckpointDefaults.FILENAME
+        )
+
+        return checkpoint_file_path
+
+
+###########################################################################
+# Encoder
+###########################################################################
+def checkpoint_encoder(obj):
+    if isinstance(obj, bytes):
+        return obj.decode("utf-8")
+    elif hasattr(obj, "create_checkpoint_object"):
+        return obj.create_checkpoint_object()
+    else:
+        return obj.__dict__

genai-perf/genai_perf/config/generate/genai_perf_config.py

@@ -0,0 +1,94 @@
+# Copyright 2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from copy import deepcopy
+from dataclasses import dataclass
+
+from genai_perf.config.generate.search_parameter import SearchUsage
+from genai_perf.config.input.config_command import (
+    ConfigCommand,
+    ConfigInput,
+    ConfigOutputTokens,
+    ConfigSyntheticTokens,
+)
+from genai_perf.types import CheckpointObject, ModelObjectiveParameters
+
+
+@dataclass
+class GenAIPerfConfig:
+    """
+    Creates a config that can be used by be
+    used by other modules to configure the various
+    options of GenAI-Perf
+    """
+
+    def __init__(
+        self,
+        config: ConfigCommand,
+        model_objective_parameters: ModelObjectiveParameters,
+    ):
+        self._set_options_based_on_config(config)
+        self._set_options_based_on_objective(model_objective_parameters)
+
+    ###########################################################################
+    # Set Options Methods
+    ###########################################################################
+    def _set_options_based_on_config(self, config: ConfigCommand) -> None:
+        self.input: ConfigInput = config.input
+        self.output_tokens: ConfigOutputTokens = config.output_tokens
+
+    def _set_options_based_on_objective(
+        self, model_objective_parameters: ModelObjectiveParameters
+    ) -> None:
+        for objective in model_objective_parameters.values():
+            for name, parameter in objective.items():
+                if parameter.usage == SearchUsage.RUNTIME_GAP:
+                    if hasattr(self.input, name):
+                        self.input.__setattr__(
+                            name, parameter.get_value_based_on_category()
+                        )
+
+    ###########################################################################
+    # Checkpoint Methods
+    ###########################################################################
+    def create_checkpoint_object(self) -> CheckpointObject:
+        """
+        Converts the class data into a dictionary that can be written to
+        the checkpoint file
+        """
+        genai_perf_config_dict = deepcopy(self.__dict__)
+
+        return genai_perf_config_dict
+
+    @classmethod
+    def create_class_from_checkpoint(
+        cls, genai_perf_config_dict: CheckpointObject
+    ) -> "GenAIPerfConfig":
+        """
+        Takes the checkpoint's representation of the class and creates (and populates)
+        a new instance of a GenAIPerfConfig
+        """
+        genai_perf_config = GenAIPerfConfig(
+            config=ConfigCommand([""]),
+            model_objective_parameters={},
+        )
+        genai_perf_config.input = ConfigInput(**genai_perf_config_dict["input"])
+        genai_perf_config.input.synthetic_tokens = ConfigSyntheticTokens(
+            **genai_perf_config_dict["input"]["synthetic_tokens"]
+        )
+        genai_perf_config.output_tokens = ConfigOutputTokens(
+            **genai_perf_config_dict["output_tokens"]
+        )
+
+        return genai_perf_config

genai-perf/genai_perf/config/generate/objective_parameter.py

@@ -0,0 +1,51 @@
+# Copyright 2024, NVIDIA CORPORATION & AFFILIATES. All rights reserved.
+#
+# Licensed under the Apache License, Version 2.0 (the "License");
+# you may not use this file except in compliance with the License.
+# You may obtain a copy of the License at
+#
+#     http://www.apache.org/licenses/LICENSE-2.0
+#
+# Unless required by applicable law or agreed to in writing, software
+# distributed under the License is distributed on an "AS IS" BASIS,
+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+
+from dataclasses import dataclass
+from enum import Enum, auto
+from typing import Any, Dict, TypeAlias
+
+from genai_perf.config.generate.search_parameter import SearchUsage
+from genai_perf.exceptions import GenAIPerfException
+
+
+class ObjectiveCategory(Enum):
+    INTEGER = auto()
+    EXPONENTIAL = auto()
+    STR = auto()
+
+
+ObjectiveParameters: TypeAlias = Dict[str, "ObjectiveParameter"]
+
+
+@dataclass
+class ObjectiveParameter:
+    """
+    A dataclass that holds information about a configuration's objective parameter
+    """
+
+    usage: SearchUsage
+    category: ObjectiveCategory
+    value: Any
+
+    def get_value_based_on_category(self) -> Any:
+        if (
+            self.category == ObjectiveCategory.INTEGER
+            or self.category == ObjectiveCategory.STR
+        ):
+            return self.value
+        elif self.category == ObjectiveCategory.EXPONENTIAL:
+            return 2**self.value
+
+        raise GenAIPerfException(f"{self.category} is not a known ObjectiveCategory")