Multi Process Scheduler Implementation, Benchmarker, and Report Generation Refactor (#96)

Full refactor of GuideLLM enabling better overall performance to ensure
minimal overhead for benchmarking with a new multiprocess and threaded
scheduler along with significant updates to the output formats enabling
better analysis, visibility, and clarity.

<img width="668" alt="Screenshot 2025-04-11 at 2 26 13 PM"
src="https://github.com/user-attachments/assets/a723854a-7fe0-4eb2-9408-f632e747c3c2"
/>

Fixes:
- #92 
- #77 
- #47 
- #79

---------

Co-authored-by: Alexandre Marques <alexandre@neuralmagic.com>
Co-authored-by: Samuel Monson <smonson@redhat.com>
Co-authored-by: David Gray <40244437+dagrayvid@users.noreply.github.com>

Author: 4 people

.github/workflows/development.yml

@@ -11,8 +11,8 @@ jobs:
     strategy:
       matrix:
         python:
-          - "3.12"
-          - "3.8"
+          - "3.13"
+          - "3.9"
     steps:
       - uses: actions/checkout@v4
       - name: Set up Python
@@ -29,8 +29,8 @@ jobs:
     strategy:
       matrix:
         python:
-          - "3.12"
-          - "3.8"
+          - "3.13"
+          - "3.9"
     steps:
       - uses: actions/checkout@v4
       - name: Set up Python

.github/workflows/nightly.yml

@@ -10,11 +10,11 @@ jobs:
     strategy:
       matrix:
         python:
+          - "3.13"
           - "3.12"
           - "3.11"
           - "3.10"
           - "3.9"
-          - "3.8"
     steps:
       - uses: actions/checkout@v4
       - name: Set up Python
@@ -31,11 +31,11 @@ jobs:
     strategy:
       matrix:
         python:
+          - "3.13"
           - "3.12"
           - "3.11"
           - "3.10"
           - "3.9"
-          - "3.8"
     steps:
       - uses: actions/checkout@v4
       - name: Set up Python
@@ -52,11 +52,11 @@ jobs:
     strategy:
       matrix:
         python:
+          - "3.13"
           - "3.12"
           - "3.11"
           - "3.10"
           - "3.9"
-          - "3.8"
     steps:
       - uses: actions/checkout@v4
       - name: Set up Python

.github/workflows/quality.yml

@@ -14,8 +14,8 @@ jobs:
     strategy:
       matrix:
         python:
-          - "3.12"
-          - "3.8"
+          - "3.13"
+          - "3.9"
     steps:
       - uses: actions/checkout@v4
       - name: Set up Python
@@ -32,8 +32,8 @@ jobs:
     strategy:
       matrix:
         python:
-          - "3.12"
-          - "3.8"
+          - "3.13"
+          - "3.9"
     steps:
       - uses: actions/checkout@v4
       - name: Set up Python
@@ -52,7 +52,7 @@ jobs:
       - name: Set up Python
         uses: actions/setup-python@v5
         with:
-          python-version: "3.8"
+          python-version: "3.9"
       - name: Install pre-commit
         run: pip install pre-commit
       - name: Run pre-commit checks

.github/workflows/release.yml

@@ -11,11 +11,11 @@ jobs:
     strategy:
       matrix:
         python:
+          - "3.13"
           - "3.12"
           - "3.11"
           - "3.10"
           - "3.9"
-          - "3.8"
     steps:
       - uses: actions/checkout@v4
       - name: Set up Python
@@ -32,11 +32,11 @@ jobs:
     strategy:
       matrix:
         python:
+          - "3.13"
           - "3.12"
           - "3.11"
           - "3.10"
           - "3.9"
-          - "3.8"
     steps:
       - uses: actions/checkout@v4
       - name: Set up Python
@@ -53,11 +53,11 @@ jobs:
     strategy:
       matrix:
         python:
+          - "3.13"
           - "3.12"
           - "3.11"
           - "3.10"
           - "3.9"
-          - "3.8"
     steps:
       - uses: actions/checkout@v4
       - name: Set up Python

pyproject.toml

@@ -8,7 +8,7 @@ where = ["src"]
 include = ["*"]
 
 [tool.setuptools.package-data]
-guidellm = ["*"]
+"guidellm.data" = ["*.gz"]
 
 
 # ************************************************
@@ -32,6 +32,7 @@ dependencies = [
     "loguru",
     "numpy",
     "pillow",
+    "protobuf",
     "pydantic>=2.0.0",
     "pydantic-settings>=2.0.0",
     "pyyaml>=6.0.0",
@@ -77,7 +78,7 @@ dev = [
 
 
 [project.entry-points.console_scripts]
-guidellm = "guidellm.main:generate_benchmark_report_cli"
+guidellm = "guidellm.__main__:cli"
 guidellm-config = "guidellm.config:print_config"
 
 

src/guidellm/__init__.py

@@ -6,14 +6,22 @@
 # flake8: noqa
 
 import os
-import transformers  # type: ignore
+import logging
+import contextlib
 
-os.environ["TOKENIZERS_PARALLELISM"] = "false"  # Silence warnings for tokenizers
-transformers.logging.set_verbosity_error()  # Silence warnings for transformers
 
+with open(os.devnull, "w") as devnull, contextlib.redirect_stderr(
+    devnull
+), contextlib.redirect_stdout(devnull):
+    from transformers.utils import logging as hf_logging  # type: ignore[import]
+
+    # Set the log level for the transformers library to ERROR
+    # to ignore None of PyTorch, TensorFlow found
+    os.environ["TOKENIZERS_PARALLELISM"] = "false"  # Silence warnings for tokenizers
+    hf_logging.set_verbosity_error()
+    logging.getLogger("transformers").setLevel(logging.ERROR)
 
 from .config import settings
 from .logger import configure_logger, logger
-from .main import generate_benchmark_report
 
 __all__ = ["configure_logger", "logger", "settings", "generate_benchmark_report"]