Add option for CUDA 13.1 llama.cpp in installer

one_click.py

@@ -47,6 +47,10 @@ def is_x86_64():
     return platform.machine() == "x86_64"
 
 
+def is_nvidia(gpu_choice):
+    return gpu_choice.startswith("NVIDIA_CUDA")
+
+
 def is_installed():
     site_packages_path = None
     for sitedir in site.getsitepackages():
@@ -81,6 +85,12 @@ def get_gpu_choice():
     """Get GPU choice from state file or ask user"""
     state = load_state()
     gpu_choice = state.get('gpu_choice')
+
+    # Migrate old state values
+    if gpu_choice == 'NVIDIA_CUDA128':
+        gpu_choice = 'NVIDIA_CUDA12'
+        state['gpu_choice'] = gpu_choice
+        save_state(state)
 
     if not gpu_choice:
         if "GPU_CHOICE" in os.environ:
@@ -99,7 +109,23 @@ def get_gpu_choice():
             )
 
         # Convert choice to GPU name
-        gpu_choice = {"A": "NVIDIA_CUDA128", "B": "AMD", "C": "APPLE", "D": "INTEL", "N": "NONE"}[choice]
+        if choice == "A":
+            # Ask for NVIDIA CUDA version
+            if "NVIDIA_CUDA_CHOICE" in os.environ:
+                cuda_choice = os.environ["NVIDIA_CUDA_CHOICE"].upper()
+                print_big_message(f"Selected NVIDIA CUDA choice \"{cuda_choice}\" based on the NVIDIA_CUDA_CHOICE environment variable.")
+            else:
+                cuda_choice = get_user_choice(
+                    "Which CUDA version do you want for llama.cpp?",
+                    {
+                        'A': 'CUDA 12.4 - Supports all GPUs',
+                        'B': 'CUDA 13.1 - Faster on RTX 50xx Blackwell, but drops GTX 750/900/1000 and Titan V (needs driver >= 580)',
+                    },
+                )
+
+            gpu_choice = {"A": "NVIDIA_CUDA12", "B": "NVIDIA_CUDA13"}[cuda_choice]
+        else:
+            gpu_choice = {"B": "AMD", "C": "APPLE", "D": "INTEL", "N": "NONE"}[choice]
 
         # Save choice to state
         state['gpu_choice'] = gpu_choice
@@ -113,7 +139,7 @@ def get_pytorch_install_command(gpu_choice):
     base_cmd = f"python -m pip install torch=={TORCH_VERSION} "
     pypi_fallback = " --extra-index-url https://pypi.org/simple/"
 
-    if gpu_choice == "NVIDIA_CUDA128":
+    if is_nvidia(gpu_choice):
         return base_cmd + "--index-url https://download.pytorch.org/whl/cu128" + pypi_fallback
     elif gpu_choice == "AMD":
         py_tag = f"cp{PYTHON_VERSION.replace('.', '')}"
@@ -131,7 +157,7 @@ def get_pytorch_update_command(gpu_choice):
     base_cmd = f"python -m pip install --upgrade torch=={TORCH_VERSION} "
     pypi_fallback = " --extra-index-url https://pypi.org/simple/"
 
-    if gpu_choice == "NVIDIA_CUDA128":
+    if is_nvidia(gpu_choice):
         return f"{base_cmd}--index-url https://download.pytorch.org/whl/cu128" + pypi_fallback
     elif gpu_choice == "AMD":
         py_tag = f"cp{PYTHON_VERSION.replace('.', '')}"
@@ -148,8 +174,10 @@ def get_requirements_file(gpu_choice):
     """Get requirements file path based on GPU choice"""
     requirements_base = os.path.join("requirements", "full")
 
-    if gpu_choice == "NVIDIA_CUDA128":
+    if gpu_choice == "NVIDIA_CUDA12":
         file_name = "requirements.txt"
+    elif gpu_choice == "NVIDIA_CUDA13":
+        file_name = "requirements_nvidia_cuda13.txt"
     elif gpu_choice == "AMD":
         file_name = "requirements_amd.txt"
     elif gpu_choice == "APPLE":
@@ -303,8 +331,9 @@ def install_webui():
                 cmd_flags_file.write("\n--cpu\n")
 
     # Handle CUDA version display
-    elif any((is_windows(), is_linux())) and gpu_choice == "NVIDIA_CUDA128":
-        print("CUDA: 12.8")
+    elif any((is_windows(), is_linux())) and is_nvidia(gpu_choice):
+        llama_cuda = "12.4" if gpu_choice == "NVIDIA_CUDA12" else "13.1"
+        print(f"PyTorch CUDA: 12.8, llama.cpp CUDA: {llama_cuda}")
 
     # No PyTorch for AMD on Windows
     elif is_windows() and gpu_choice == "AMD":

requirements/full/requirements_nvidia_cuda13.txt

@@ -0,0 +1,51 @@
+accelerate==1.13.*
+audioop-lts<1.0; python_version >= "3.13"
+bitsandbytes==0.49.*
+datasets
+diffusers==0.37.*
+einops
+fastapi==0.112.4
+flash-linear-attention==0.4.*
+huggingface-hub==1.5.*
+jinja2==3.1.6
+markdown
+mcp==1.27.0
+numpy==2.2.*
+pandas
+peft==0.18.*
+Pillow>=9.5.0
+pydantic==2.11.0
+pymupdf==1.27.*
+python-docx==1.1.2
+pyyaml
+requests
+rich
+safetensors==0.7.*
+scipy
+sentencepiece
+tensorboard
+torchao==0.15.*
+trafilatura==2.0.0
+transformers==5.5.*
+triton-windows==3.5.1.post24; platform_system == "Windows"
+tqdm
+wandb
+
+# Gradio
+https://github.com/oobabooga/gradio/releases/download/4.37.2-custom.19/gradio-4.37.2+custom.19-py3-none-any.whl
+https://github.com/oobabooga/gradio/releases/download/4.37.2-custom.19/gradio_client-1.0.2+custom.19-py3-none-any.whl
+
+# API
+flask_cloudflared==0.0.15
+sse-starlette==1.6.5
+tiktoken
+
+# CUDA wheels
+https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.110.0/llama_cpp_binaries-0.110.0+cu131-py3-none-win_amd64.whl; platform_system == "Windows"
+https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.110.0/llama_cpp_binaries-0.110.0+cu131-py3-none-linux_x86_64.whl; platform_system == "Linux" and platform_machine == "x86_64"
+https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.110.0/ik_llama_cpp_binaries-0.110.0+cu131-py3-none-win_amd64.whl; platform_system == "Windows"
+https://github.com/oobabooga/llama-cpp-binaries/releases/download/v0.110.0/ik_llama_cpp_binaries-0.110.0+cu131-py3-none-linux_x86_64.whl; platform_system == "Linux" and platform_machine == "x86_64"
+https://github.com/turboderp-org/exllamav3/releases/download/v0.0.28/exllamav3-0.0.28+cu128.torch2.9.0-cp313-cp313-win_amd64.whl; platform_system == "Windows" and python_version == "3.13"
+https://github.com/turboderp-org/exllamav3/releases/download/v0.0.28/exllamav3-0.0.28+cu128.torch2.9.0-cp313-cp313-linux_x86_64.whl; platform_system == "Linux" and platform_machine == "x86_64" and python_version == "3.13"
+https://github.com/kingbri1/flash-attention/releases/download/v2.8.3/flash_attn-2.8.3+cu128torch2.9.0cxx11abiFALSE-cp313-cp313-win_amd64.whl; platform_system == "Windows" and python_version == "3.13"
+https://github.com/kingbri1/flash-attention/releases/download/v2.8.3/flash_attn-2.8.3+cu128torch2.9.0cxx11abiFALSE-cp313-cp313-linux_x86_64.whl; platform_system == "Linux" and platform_machine == "x86_64" and python_version == "3.13"