diff --git a/.github/workflows/check_quality.yaml b/.github/workflows/check_quality.yaml
new file mode 100644
index 00000000..5f550dc3
--- /dev/null
+++ b/.github/workflows/check_quality.yaml
@@ -0,0 +1,36 @@
+name: Quality checks
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  run_cpu_tests:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v2
+
+      - name: Set up Python 3.8
+        uses: actions/setup-python@v2
+        with:
+          python-version: 3.8
+
+      - name: Install quality requirements
+        run: |
+          pip install --upgrade pip
+          pip install -e .[quality]
+
+      - name: Check style with black
+        run: |
+          black --check .
+
+      - name: Check style with ruff
+        run: |
+          ruff .
diff --git a/.github/workflows/test_cpu_neural_compressor.yaml b/.github/workflows/test_cpu_neural_compressor.yaml
new file mode 100644
index 00000000..46b66c11
--- /dev/null
+++ b/.github/workflows/test_cpu_neural_compressor.yaml
@@ -0,0 +1,32 @@
+name: Intel Neural Compressor CPU Tests
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  run_cpu_tests:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v2
+
+      - name: Set up Python 3.8
+        uses: actions/setup-python@v2
+        with:
+          python-version: 3.8
+
+      - name: Install Intel Neural Compressor CPU requirements
+        run: |
+          pip install --upgrade pip
+          pip install -e .[test,neural-compressor]
+
+      - name: Run Intel Neural Compressor CPU tests
+        run: |
+          pytest -k "cpu_neural_compressor"
diff --git a/.github/workflows/test_cpu.yaml b/.github/workflows/test_cpu_onnxruntime.yaml
similarity index 67%
rename from .github/workflows/test_cpu.yaml
rename to .github/workflows/test_cpu_onnxruntime.yaml
index 90c8d3e9..2d239171 100644
--- a/.github/workflows/test_cpu.yaml
+++ b/.github/workflows/test_cpu_onnxruntime.yaml
@@ -1,4 +1,4 @@
-name: CPU Unit Tests
+name: OnnxRuntime CPU Tests
 
 on:
   push:
@@ -14,7 +14,7 @@ jobs:
   run_cpu_tests:
     runs-on: ubuntu-latest
     steps:
-      - name: Checkout code
+      - name: Checkout
         uses: actions/checkout@v2
 
       - name: Set up Python 3.8
@@ -22,11 +22,11 @@ jobs:
         with:
           python-version: 3.8
 
-      - name: Install CPU requirements
+      - name: Install requirements
         run: |
           pip install --upgrade pip
-          pip install -r cpu_requirements.txt
-          pip install -e .[test]
+          pip install -e .[test,onnxruntime,diffusers]
 
-      - name: Run CPU tests
-        run: pytest -k "cpu"
+      - name: Run tests
+        run: |
+          pytest -k "cpu_onnxruntime"
diff --git a/.github/workflows/test_cpu_openvino.yaml b/.github/workflows/test_cpu_openvino.yaml
new file mode 100644
index 00000000..515e3a4b
--- /dev/null
+++ b/.github/workflows/test_cpu_openvino.yaml
@@ -0,0 +1,32 @@
+name: OpenVINO CPU Tests
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  run_cpu_tests:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v2
+
+      - name: Set up Python 3.8
+        uses: actions/setup-python@v2
+        with:
+          python-version: 3.8
+
+      - name: Install requirements
+        run: |
+          pip install --upgrade pip
+          pip install -e .[test,openvino,diffusers]
+
+      - name: Run tests
+        run: |
+          pytest -k "cpu_openvino"
diff --git a/.github/workflows/test_cpu_pytorch.yaml b/.github/workflows/test_cpu_pytorch.yaml
new file mode 100644
index 00000000..a23e84cd
--- /dev/null
+++ b/.github/workflows/test_cpu_pytorch.yaml
@@ -0,0 +1,32 @@
+name: Pytorch CPU tests
+
+on:
+  push:
+    branches: [main]
+  pull_request:
+    branches: [main]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  run_cpu_tests:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v2
+
+      - name: Set up Python 3.8
+        uses: actions/setup-python@v2
+        with:
+          python-version: 3.8
+
+      - name: Install requirements
+        run: |
+          pip install --upgrade pip
+          pip install -e .[test,diffusers]
+
+      - name: Run tests
+        run: |
+          pytest -k "cpu_pytorch"
diff --git a/.github/workflows/test_gpu.yaml b/.github/workflows/test_cuda_onnxruntime_inference.yaml
similarity index 69%
rename from .github/workflows/test_gpu.yaml
rename to .github/workflows/test_cuda_onnxruntime_inference.yaml
index d76f763f..161a7b6f 100644
--- a/.github/workflows/test_gpu.yaml
+++ b/.github/workflows/test_cuda_onnxruntime_inference.yaml
@@ -1,4 +1,4 @@
-name: GPU Unit Tests
+name: OnnxRuntime CUDA Inference Tests
 
 on:
   pull_request:
@@ -12,7 +12,7 @@ jobs:
   build-and-test:
     runs-on: self-hosted
     steps:
-      - name: Restore docker ownership
+      - name: Restore files ownership
         run: docker run
           --rm
           --entrypoint /bin/bash
@@ -23,13 +23,10 @@ jobs:
           ubuntu
           -c 'chown -R ${HOST_UID}:${HOST_GID} /workspace/optimum-benchmark'
 
-      - name: Checkout code
+      - name: Checkout
         uses: actions/checkout@v2
 
-      - name: Build GPU Docker image
-        run: docker build --no-cache --build-arg CACHEBUST=$(date +%s) -f docker/gpu.dockerfile -t optimum-benchmark-gpu .
-
-      - name: Run GPU tests
+      - name: Run tests
         run: docker run
           --rm
           --entrypoint /bin/bash
@@ -37,4 +34,4 @@ jobs:
           --volume $(pwd):/workspace/optimum-benchmark
           --workdir /workspace/optimum-benchmark
           optimum-benchmark-gpu
-          -c "pip install -e .[test] && pytest -k '(cuda or tensorrt) and not onnxruntime_training' -x"
+          -c "pip install -e .[test,diffusers] && pytest -k '(cuda or tensorrt) and onnxruntime and not training' -x"
diff --git a/.github/workflows/test_ort_training.yaml b/.github/workflows/test_cuda_onnxruntime_training.yaml
similarity index 65%
rename from .github/workflows/test_ort_training.yaml
rename to .github/workflows/test_cuda_onnxruntime_training.yaml
index 6540628e..3aa34330 100644
--- a/.github/workflows/test_ort_training.yaml
+++ b/.github/workflows/test_cuda_onnxruntime_training.yaml
@@ -1,4 +1,4 @@
-name: OnnxRuntime Training Unit Tests
+name: OnnxRuntime CUDA Training Tests
 
 on:
   pull_request:
@@ -12,7 +12,7 @@ jobs:
   build-and-test:
     runs-on: self-hosted
     steps:
-      - name: Restore docker ownership
+      - name: Restore files ownership
         run: docker run
           --rm
           --entrypoint /bin/bash
@@ -23,18 +23,15 @@ jobs:
           ubuntu
           -c 'chown -R ${HOST_UID}:${HOST_GID} /workspace/optimum-benchmark'
 
-      - name: Checkout code
+      - name: Checkout
         uses: actions/checkout@v2
 
-      - name: Build OnnxRuntime Training Docker image
-        run: docker build --no-cache --build-arg CACHEBUST=$(date +%s) -f docker/ort_training.dockerfile -t optimum-benchmark-ort-training .
-
-      - name: Run OnnxRuntime Training tests
+      - name: Run tests
         run: docker run
           --rm
           --entrypoint /bin/bash
           --gpus '"device=0,1"'
           --volume $(pwd):/workspace/optimum-benchmark
           --workdir /workspace/optimum-benchmark
-          optimum-benchmark-ort-training
-          -c "pip install -e .[test] && pytest -k 'cuda_onnxruntime_training' -x"
+          onnxruntime-training
+          -c "pip install -e .[test,peft] && pytest -k 'onnxruntime_training' -x"
diff --git a/.github/workflows/test_cuda_pytorch.yaml b/.github/workflows/test_cuda_pytorch.yaml
new file mode 100644
index 00000000..40c4d5dc
--- /dev/null
+++ b/.github/workflows/test_cuda_pytorch.yaml
@@ -0,0 +1,37 @@
+name: Pytorch CUDA Tests
+
+on:
+  pull_request:
+    types: [opened, reopened, synchronize]
+
+concurrency:
+  group: ${{ github.workflow }}-${{ github.event.pull_request.number || github.ref }}
+  cancel-in-progress: true
+
+jobs:
+  build-and-test:
+    runs-on: self-hosted
+    steps:
+      - name: Restore files ownership
+        run: docker run
+          --rm
+          --entrypoint /bin/bash
+          --env HOST_UID=`id -u`
+          --env HOST_GID=`id -g`
+          --volume $(pwd):/workspace/optimum-benchmark
+          --workdir /workspace/optimum-benchmark
+          ubuntu
+          -c 'chown -R ${HOST_UID}:${HOST_GID} /workspace/optimum-benchmark'
+
+      - name: Checkout
+        uses: actions/checkout@v2
+
+      - name: Run tests
+        run: docker run
+          --rm
+          --entrypoint /bin/bash
+          --gpus '"device=0,1"'
+          --volume $(pwd):/workspace/optimum-benchmark
+          --workdir /workspace/optimum-benchmark
+          optimum-benchmark-gpu
+          -c "pip install -e .[test,peft,diffusers] && pytest -k 'cuda_pytorch' -x"
diff --git a/Makefile b/Makefile
new file mode 100644
index 00000000..bd8fab32
--- /dev/null
+++ b/Makefile
@@ -0,0 +1,36 @@
+#  Copyright 2021 The HuggingFace Team. 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.
+
+SHELL := /bin/bash
+CURRENT_DIR = $(shell pwd)
+DEFAULT_CLONE_URL := https://github.com/huggingface/optimum-benchmark.git
+# If CLONE_URL is empty, revert to DEFAULT_CLONE_URL
+REAL_CLONE_URL = $(if $(CLONE_URL),$(CLONE_URL),$(DEFAULT_CLONE_URL))
+
+# Install the library in development mode
+.PHONY:	style test
+
+# Run code quality checks
+style_check:
+	black --check .
+	ruff .
+
+# Format the code
+style:
+	black .
+	ruff . --fix
+
+# Run tests for the library
+test:
+	python -m pytest tests
\ No newline at end of file
diff --git a/docker/gpu.dockerfile b/docker/gpu.dockerfile
index f01d2bbe..42dfa69f 100644
--- a/docker/gpu.dockerfile
+++ b/docker/gpu.dockerfile
@@ -27,11 +27,3 @@ RUN apt-get install -y software-properties-common wget apt-utils patchelf git li
     apt-get clean
 RUN unattended-upgrade
 RUN apt-get autoremove -y
-RUN pip install --upgrade pip
-
-# this line forces the docker build to rebuild from this point on
-ARG CACHEBUST=1
-
-# Install optimum-benchmark dependencies
-COPY gpu_requirements.txt /tmp/gpu_requirements.txt
-RUN pip install -r /tmp/gpu_requirements.txt
\ No newline at end of file
diff --git a/docker/ort_training.dockerfile b/docker/onnxruntime_training.dockerfile
similarity index 92%
rename from docker/ort_training.dockerfile
rename to docker/onnxruntime_training.dockerfile
index c40654d8..3a469a0a 100644
--- a/docker/ort_training.dockerfile
+++ b/docker/onnxruntime_training.dockerfile
@@ -63,10 +63,3 @@ RUN $PYTHON_EXE -m pip install torch-ort
 ENV TORCH_CUDA_ARCH_LIST="5.2 6.0 6.1 7.0 7.5 8.0 8.6+PTX"
 RUN $PYTHON_EXE -m pip install --upgrade protobuf==3.20.2
 RUN $PYTHON_EXE -m torch_ort.configure
-
-# this line forces the docker build to rebuild from this point on
-ARG CACHEBUST=1
-
-# Install optimum-benchmark dependencies
-COPY requirements.txt /tmp/requirements.txt
-RUN pip install -r /tmp/requirements.txt
\ No newline at end of file
diff --git a/docker/scripts/build_gpu.sh b/docker/scripts/build_gpu.sh
deleted file mode 100644
index 5ecfdb6a..00000000
--- a/docker/scripts/build_gpu.sh
+++ /dev/null
@@ -1 +0,0 @@
-docker build --no-cache --build-arg CACHEBUST=$(date +%s) -f docker/gpu.dockerfile -t optimum-benchmark-gpu .
\ No newline at end of file
diff --git a/docker/scripts/build_ort_training.sh b/docker/scripts/build_ort_training.sh
deleted file mode 100644
index d2efd477..00000000
--- a/docker/scripts/build_ort_training.sh
+++ /dev/null
@@ -1 +0,0 @@
-docker build --no-cache --build-arg CACHEBUST=$(date +%s) -f docker/ort_training.dockerfile -t optimum-benchmark-ort-training .
\ No newline at end of file
diff --git a/optimum_benchmark/backends/base.py b/optimum_benchmark/backends/base.py
index 774223e0..6727f855 100644
--- a/optimum_benchmark/backends/base.py
+++ b/optimum_benchmark/backends/base.py
@@ -64,30 +64,30 @@ def __init__(self, model: str, task: str, device: str, hub_kwargs: Dict[str, Any
 
         if self.is_diffusion_pipeline():
             # for pipelines
+            self.library = "diffusers"
+            self.model_type = self.task
             self.pretrained_config = None
             self.pretrained_processor = None
-            self.model_type = self.task
         else:
             # for models
+            self.library = "transformers"
             self.pretrained_config = AutoConfig.from_pretrained(
                 pretrained_model_name_or_path=self.model, **self.hub_kwargs
             )
             self.model_type = self.pretrained_config.model_type
 
             try:
-                # the processor sometimes contains information about the model's
-                # input shapes that's not available in the config
+                # the processor sometimes contains information about the model's input shapes that's not available in the config
                 self.pretrained_processor = AutoProcessor.from_pretrained(
                     pretrained_model_name_or_path=self.model, **self.hub_kwargs
                 )
             except ValueError:
+                # sometimes the processor is not available or can't be determined/detected
                 LOGGER.warning("Could not find the model's preprocessor")
                 self.pretrained_processor = None
 
         self.automodel_class = TasksManager.get_model_class_for_task(
-            task=self.task,
-            framework="pt",
-            model_type=self.model_type,
+            task=self.task, library=self.library, model_type=self.model_type
         )
 
     def is_text_generation_model(self) -> bool:
diff --git a/setup.py b/setup.py
index cd83b6f7..5471e59f 100644
--- a/setup.py
+++ b/setup.py
@@ -1,12 +1,55 @@
 from setuptools import find_packages, setup
 
+OPTIMUM_VERSION = "1.13.0"
+
+INSTALL_REQUIRES = [
+    # Mandatory HF dependencies
+    f"optimum=={OPTIMUM_VERSION}",  # backends, tasks and input generation
+    "accelerate",  # distributed inference and no weights init
+    # Hydra
+    "omegaconf==2.3.0",
+    "hydra-core==1.3.2",
+    "hydra_colorlog==1.2.0",
+    "hydra-joblib-launcher==1.2.0",
+    # Other
+    "codecarbon==2.3.1",
+    "psutil==5.9.0",
+    "pandas>=2.0.0",
+]
+
+# add py3nvml if nvidia driver is installed
+try:
+    import subprocess
+
+    subprocess.run(["nvidia-smi"], stdout=subprocess.DEVNULL)
+    INSTALL_REQUIRES.append("py3nvml==0.2.7")
+except FileNotFoundError:
+    pass
+
+
+EXTRAS_REQUIRE = {
+    "test": ["pytest"],
+    "quality": ["black", "ruff"],
+    "report": ["flatten_dict", "matplotlib", "seaborn", "rich"],
+    # cpu backends
+    "openvino": [f"optimum[openvino,nncf]=={OPTIMUM_VERSION}"],
+    "onnxruntime": [f"optimum[onnxruntime]=={OPTIMUM_VERSION}"],
+    "onnxruntime-gpu": [f"optimum[onnxruntime-gpu]=={OPTIMUM_VERSION}"],
+    "neural-compressor": [f"optimum[neural-compressor]=={OPTIMUM_VERSION}"],
+    # server-like backend
+    "text-generation-inference": ["docker==6.1.3"],
+    # specific settings
+    "diffusers": ["diffusers"],
+    "peft": ["peft"],
+}
+
+
 setup(
     name="optimum-benchmark",
     version="0.0.1",
     packages=find_packages(),
-    extras_require={
-        "test": ["pytest"],
-    },
+    install_requires=INSTALL_REQUIRES,
+    extras_require=EXTRAS_REQUIRE,
     entry_points={
         "console_scripts": [
             "optimum-benchmark=optimum_benchmark.experiment:run_experiment",