Simplify GPTQ-for-LLaMa installation (#122)

This commit is contained in:
jllllll 2023-08-10 11:19:47 -05:00 committed by GitHub
parent fa4a948b38
commit 28e3ce4317
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23

View File

@ -1,8 +1,6 @@
import argparse import argparse
import glob import glob
import os import os
import re
import site
import subprocess import subprocess
import sys import sys
@ -126,10 +124,10 @@ def update_dependencies(initial_installation=False):
os.chdir("text-generation-webui") os.chdir("text-generation-webui")
run_cmd("git pull", assert_success=True, environment=True) run_cmd("git pull", assert_success=True, environment=True)
textgen_requirements = open("requirements.txt").read().splitlines()
# Workaround for git+ packages not updating properly Also store requirements.txt for later use # Workaround for git+ packages not updating properly Also store requirements.txt for later use
with open("requirements.txt") as f: git_requirements = [req for req in textgen_requirements if req.startswith("git+")]
textgen_requirements = f.read()
git_requirements = [req for req in textgen_requirements.splitlines() if req.startswith("git+")]
# Loop through each "git+" requirement and uninstall it # Loop through each "git+" requirement and uninstall it
for req in git_requirements: for req in git_requirements:
@ -165,17 +163,6 @@ def update_dependencies(initial_installation=False):
clear_cache() clear_cache()
return return
# Get GPU CUDA/compute support
if '+cu' in torver:
nvcc_device_query = "__nvcc_device_query" if not sys.platform.startswith("win") else "__nvcc_device_query.exe"
compute_array = run_cmd(os.path.join(conda_env_path, "bin", nvcc_device_query), environment=True, capture_output=True)
else:
compute_array = type('obj', (object,), {'stdout': b'', 'returncode': 1})
# Fix a bitsandbytes compatibility issue with Linux
# if sys.platform.startswith("linux"):
# shutil.copy(os.path.join(site_packages_path, "bitsandbytes", "libbitsandbytes_cuda117.so"), os.path.join(site_packages_path, "bitsandbytes", "libbitsandbytes_cpu.so"))
if not os.path.exists("repositories/"): if not os.path.exists("repositories/"):
os.mkdir("repositories") os.mkdir("repositories")
@ -197,24 +184,10 @@ def update_dependencies(initial_installation=False):
if 'rocm5.4.2-cp310-cp310-linux_x86_64.whl' in exllama_rocm: if 'rocm5.4.2-cp310-cp310-linux_x86_64.whl' in exllama_rocm:
run_cmd("python -m pip install " + exllama_rocm, environment=True) run_cmd("python -m pip install " + exllama_rocm, environment=True)
# Fix build issue with exllama in Linux/WSL # Fix JIT compile issue with exllama in Linux/WSL
if sys.platform.startswith("linux") and not os.path.exists(f"{conda_env_path}/lib64"): if sys.platform.startswith("linux") and not os.path.exists(f"{conda_env_path}/lib64"):
run_cmd(f'ln -s "{conda_env_path}/lib" "{conda_env_path}/lib64"', environment=True) run_cmd(f'ln -s "{conda_env_path}/lib" "{conda_env_path}/lib64"', environment=True)
# oobabooga fork requires min compute of 6.0
gptq_min_compute = 60
gptq_min_compute_check = any(int(compute) >= gptq_min_compute for compute in compute_array.stdout.decode('utf-8').split(',')) if compute_array.returncode == 0 else False
# Install GPTQ-for-LLaMa which enables 4bit CUDA quantization
if not os.path.exists("GPTQ-for-LLaMa/"):
# Install oobabooga fork if min compute met or if failed to check
if '+rocm' in torver:
run_cmd("git clone https://github.com/WapaMario63/GPTQ-for-LLaMa-ROCm.git GPTQ-for-LLaMa -b rocm", assert_success=True, environment=True)
elif gptq_min_compute_check or compute_array.returncode != 0:
run_cmd("git clone https://github.com/oobabooga/GPTQ-for-LLaMa.git -b cuda", assert_success=True, environment=True)
else:
run_cmd("git clone https://github.com/qwopqwop200/GPTQ-for-LLaMa.git -b cuda", assert_success=True, environment=True)
# On some Linux distributions, g++ may not exist or be the wrong version to compile GPTQ-for-LLaMa # On some Linux distributions, g++ may not exist or be the wrong version to compile GPTQ-for-LLaMa
if sys.platform.startswith("linux"): if sys.platform.startswith("linux"):
gxx_output = run_cmd("g++ -dumpfullversion -dumpversion", environment=True, capture_output=True) gxx_output = run_cmd("g++ -dumpfullversion -dumpversion", environment=True, capture_output=True)
@ -224,64 +197,21 @@ def update_dependencies(initial_installation=False):
# Install/Update ROCm AutoGPTQ for AMD GPUs # Install/Update ROCm AutoGPTQ for AMD GPUs
if '+rocm' in torver: if '+rocm' in torver:
if run_cmd("[ -d ./AutoGPTQ-rocm ] && rm -rfd ./AutoGPTQ-rocm; git clone https://github.com/jllllll/AutoGPTQ.git ./AutoGPTQ-rocm -b rocm && cp ./AutoGPTQ-rocm/setup_rocm.py ./AutoGPTQ-rocm/setup.py && python -m pip install ./AutoGPTQ-rocm --force-reinstall --no-deps", environment=True).returncode != 0: auto_gptq_version = [req for req in textgen_requirements if req.startswith('https://github.com/PanQiWei/AutoGPTQ/releases/download/')][0].split('/')[7]
print_big_message("WARNING: AutoGPTQ kernel compilation failed!\n The installer will proceed to install a pre-compiled wheel.") auto_gptq_wheel = run_cmd(f'curl -s https://api.github.com/repos/PanQiWei/AutoGPTQ/releases/tags/{auto_gptq_version} | grep browser_download_url | grep rocm5.4.2-cp310-cp310-linux_x86_64.whl | cut -d : -f 2,3 | tr -d \'"\'', environment=True, capture_output=True).stdout.decode('utf-8')
if run_cmd("python -m pip install https://github.com/jllllll/GPTQ-for-LLaMa-Wheels/raw/Linux-x64/ROCm-5.4.2/auto_gptq-0.3.2%2Brocm5.4.2-cp310-cp310-linux_x86_64.whl --force-reinstall --no-deps", environment=True).returncode != 0: if not auto_gptq_wheel and run_cmd(f"python -m pip install {auto_gptq_wheel} --force-reinstall --no-deps", environment=True).returncode != 0:
print_big_message("ERROR: AutoGPTQ wheel installation failed!\n You will not be able to use GPTQ-based models with AutoGPTQ.") print_big_message("ERROR: AutoGPTQ wheel installation failed!\n You will not be able to use GPTQ-based models with AutoGPTQ.")
# Install GPTQ-for-LLaMa dependencies # Install GPTQ-for-LLaMa for ROCm
os.chdir("GPTQ-for-LLaMa")
run_cmd("git pull", environment=True)
# Finds the path to your dependencies
for sitedir in site.getsitepackages():
if "site-packages" in sitedir:
site_packages_path = sitedir
break
# This path is critical to installing the following dependencies
if site_packages_path is None:
print("Could not find the path to your Python packages. Exiting...")
sys.exit()
# Compile and install GPTQ-for-LLaMa
if '+rocm' in torver: if '+rocm' in torver:
if os.path.exists('setup_rocm.py'): gptq_wheel = run_cmd('curl -s https://api.github.com/repos/jllllll/GPTQ-for-LLaMa-CUDA/releases/latest | grep browser_download_url | grep rocm5.4.2-cp310-cp310-linux_x86_64.whl | cut -d : -f 2,3 | tr -d \'"\'', environment=True, capture_output=True).stdout.decode('utf-8')
os.replace("setup_rocm.py", "setup.py") install_gptq = run_cmd("python -m pip install " + gptq_wheel, environment=True).returncode == 0
elif os.path.exists('setup_cuda.py'): if install_gptq:
os.rename("setup_cuda.py", "setup.py")
build_gptq = run_cmd("python -m pip install .", environment=True).returncode == 0
# Wheel installation can fail while in the build directory of a package with the same name
os.chdir("..")
# If the path does not exist or if command returncode is not 0, then the install failed or was potentially installed outside env
quant_cuda_path_regex = os.path.join(site_packages_path, "quant_cuda*/")
quant_cuda_path = glob.glob(quant_cuda_path_regex)
if not build_gptq:
# Attempt installation via alternative, Windows/Linux-specific method
if sys.platform.startswith("win") or sys.platform.startswith("linux") and not quant_cuda_path:
print_big_message("WARNING: GPTQ-for-LLaMa compilation failed, but this is FINE and can be ignored!\nThe installer will proceed to install a pre-compiled wheel.")
if '+rocm' in torver:
wheel = 'ROCm-5.4.2/quant_cuda-0.0.0-cp310-cp310-linux_x86_64.whl'
else:
wheel = f"{'' if gptq_min_compute_check or compute_array.returncode != 0 else '832e220d6dbf11bec5eaa8b221a52c1c854d2a25/'}quant_cuda-0.0.0-cp310-cp310-{'linux_x86_64' if sys.platform.startswith('linux') else 'win_amd64'}.whl"
url = f"https://github.com/jllllll/GPTQ-for-LLaMa-Wheels/raw/{'Linux-x64' if sys.platform.startswith('linux') else 'main'}/" + wheel
result = run_cmd("python -m pip install " + url, environment=True)
if result.returncode == 0 and glob.glob(quant_cuda_path_regex):
print("Wheel installation success!") print("Wheel installation success!")
else: else:
print("ERROR: GPTQ wheel installation failed. You will not be able to use GPTQ-based models.") print("ERROR: GPTQ wheel installation failed.")
elif quant_cuda_path:
print_big_message("WARNING: GPTQ-for-LLaMa compilation failed, but this is FINE and can be ignored!\nquant_cuda has already been installed.")
else:
print("ERROR: GPTQ CUDA kernel compilation failed.")
print("You will not be able to use GPTQ-based models with GPTQ-for-LLaMa.") print("You will not be able to use GPTQ-based models with GPTQ-for-LLaMa.")
print("Continuing with install..")
clear_cache() clear_cache()