diff --git a/.github/workflows/build.yaml b/.github/workflows/build.yaml
index 6694c01c..862aef82 100644
--- a/.github/workflows/build.yaml
+++ b/.github/workflows/build.yaml
@@ -32,22 +32,18 @@ jobs:
             const script = require('.github/workflows/scripts/github_create_release.js')
             await script(github, context, core)
 
-  build_cuda_wheels:
-    name: Build AWQ with CUDA
+  build_wheels:
+    name: Build AWQ
     runs-on: ${{ matrix.os }}
     needs: release
     
     strategy:
       matrix:
         os: [ubuntu-20.04, windows-latest]
-        pyver: ["3.8", "3.9", "3.10", "3.11"]
-        cuda: ["12.1.1"]
+        pyver: ["3.9", "3.10", "3.11", "3.12"]
     defaults:
       run:
         shell: pwsh
-    env:
-      PYPI_CUDA_VERSION: "12.1.1"
-      CUDA_VERSION: ${{ matrix.cuda }}
     
     steps:
       - name: Free Disk Space
@@ -81,38 +77,16 @@ jobs:
           
       - name: Install Dependencies
         run: |
-          # Install CUDA toolkit
-          mamba install -y 'cuda' -c "nvidia/label/cuda-${env:CUDA_VERSION}"
-
-          # Env variables
-          $env:CUDA_PATH = $env:CONDA_PREFIX
-          $env:CUDA_HOME = $env:CONDA_PREFIX
-          
           # Install torch
-          $cudaVersion = $env:CUDA_VERSION.Replace('.', '')
-          $cudaVersionPytorch = $cudaVersion.Substring(0, $cudaVersion.Length - 1)
-          $pytorchVersion = "torch==2.3.1"
-          python -m pip install --upgrade --no-cache-dir $pytorchVersion+cu$cudaVersionPytorch --index-url https://download.pytorch.org/whl/cu$cudaVersionPytorch
-          python -m pip install build setuptools wheel ninja requests
+          python -m pip install --upgrade --no-cache-dir torch==2.5.1
+          python -m pip install build setuptools wheel
 
           # Print version information
           python --version
           python -c "import torch; print('PyTorch:', torch.__version__)"
-          python -c "import torch; print('CUDA:', torch.version.cuda)"
-          python -c "import os; print('CUDA_HOME:', os.getenv('CUDA_HOME', None))"
-          python -c "from torch.utils import cpp_extension; print (cpp_extension.CUDA_HOME)"
 
       - name: Build Wheel
         run: |
-          $env:CUDA_PATH = $env:CONDA_PREFIX
-          $env:CUDA_HOME = $env:CONDA_PREFIX
-
-          # Only add +cu118 to wheel if not releasing on PyPi
-          if ( $env:CUDA_VERSION -eq $env:PYPI_CUDA_VERSION ){
-            $env:PYPI_BUILD = 1
-          }
-          $env:PYPI_FORCE_TAGS = 1
-
           python setup.py sdist bdist_wheel
 
       - name: Upload Assets
diff --git a/awq/__init__.py b/awq/__init__.py
index 453b94c8..a65a5304 100644
--- a/awq/__init__.py
+++ b/awq/__init__.py
@@ -1,2 +1,2 @@
-__version__ = "0.2.6"
+__version__ = "0.2.7"
 from awq.models.auto import AutoAWQForCausalLM
diff --git a/scripts/download_wheels.sh b/scripts/download_wheels.sh
index 7a5ec313..4bca9409 100644
--- a/scripts/download_wheels.sh
+++ b/scripts/download_wheels.sh
@@ -1,7 +1,7 @@
 #!/bin/bash
 
 # Set variables
-AWQ_VERSION="0.2.6"
+AWQ_VERSION="0.2.7"
 RELEASE_URL="https://api.github.com/repos/casper-hansen/AutoAWQ/releases/tags/v${AWQ_VERSION}"
 
 # Create a directory to download the wheels
diff --git a/setup.py b/setup.py
index 3b631933..687c6134 100644
--- a/setup.py
+++ b/setup.py
@@ -2,10 +2,8 @@
 import torch
 from pathlib import Path
 from setuptools import setup, find_packages
-from torch.utils.cpp_extension import CUDAExtension
 
-AUTOAWQ_VERSION = "0.2.6"
-PYPI_BUILD = os.getenv("PYPI_BUILD", "0") == "1"
+AUTOAWQ_VERSION = "0.2.7"
 INSTALL_KERNELS = os.getenv("INSTALL_KERNELS", "0") == "1"
 IS_CPU_ONLY = not torch.backends.mps.is_available() and not torch.cuda.is_available()
 TORCH_VERSION = str(os.getenv("TORCH_VERSION", None) or torch.__version__).split('+', maxsplit=1)[0]
@@ -14,23 +12,6 @@
 if CUDA_VERSION:
     CUDA_VERSION = "".join(CUDA_VERSION.split("."))[:3]
 
-ROCM_VERSION = os.getenv("ROCM_VERSION", None) or torch.version.hip
-if ROCM_VERSION:
-    ROCM_VERSION_LEN = min(len(ROCM_VERSION.split(".")), 3)
-    ROCM_VERSION = "".join(ROCM_VERSION.split("."))[:ROCM_VERSION_LEN]
-
-if not PYPI_BUILD:
-    if IS_CPU_ONLY:
-        AUTOAWQ_VERSION += "+cpu"
-    elif CUDA_VERSION:
-        AUTOAWQ_VERSION += f"+cu{CUDA_VERSION}"
-    elif ROCM_VERSION:
-        AUTOAWQ_VERSION += f"+rocm{ROCM_VERSION}"
-    else:
-        raise RuntimeError(
-            "Your system must have either Nvidia or AMD GPU to build this package."
-        )
-
 common_setup_kwargs = {
     "version": AUTOAWQ_VERSION,
     "name": "autoawq",
@@ -50,10 +31,10 @@
         "Environment :: GPU :: NVIDIA CUDA :: 12",
         "License :: OSI Approved :: MIT License",
         "Natural Language :: English",
-        "Programming Language :: Python :: 3.8",
         "Programming Language :: Python :: 3.9",
         "Programming Language :: Python :: 3.10",
         "Programming Language :: Python :: 3.11",
+        "Programming Language :: Python :: 3.12",
         "Programming Language :: C++",
     ],
 }
@@ -82,18 +63,6 @@
 elif IS_CPU_ONLY:
     requirements.append("intel-extension-for-pytorch>=2.4.0")
 
-force_extension = os.getenv("PYPI_FORCE_TAGS", "0")
-if force_extension == "1":
-    # NOTE: We create an empty CUDAExtension because torch helps us with
-    # creating the right boilerplate to enable correct targeting of
-    # the autoawq-kernels package
-    common_setup_kwargs["ext_modules"] = [
-        CUDAExtension(
-            name="test_kernel",
-            sources=[],
-        )
-    ]
-
 setup(
     packages=find_packages(),
     install_requires=requirements,