File size: 8,747 Bytes
7344bef | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 | {
"components": {
"python": {
"3.10": { "ver": "3.10.9", "label": "Python 3.10.9 (Legacy)" },
"3.11": { "ver": "3.11.14", "label": "Python 3.11.14 (Recommended)" }
},
"torch": {
"cu128": {
"label": "Torch 2.7.1 + CUDA 12.8",
"cmd": "torch==2.7.1 torchvision==0.22.1 torchaudio==2.7.1 --index-url https://download.pytorch.org/whl/cu128"
},
"cu130": {
"label": "Torch 2.10.0 + CUDA 13.0",
"cmd": "torch==2.10.0 torchvision torchaudio --index-url https://download.pytorch.org/whl/cu130"
},
"rocm65": {
"label": "ROCm 6.5 (TheRock)",
"cmd": {
"win": "https://github.com/scottt/rocm-TheRock/releases/download/v6.5.0rc-pytorch-gfx110x/torch-2.7.0a0+rocm_git3f903c3-cp311-cp311-win_amd64.whl https://github.com/scottt/rocm-TheRock/releases/download/v6.5.0rc-pytorch-gfx110x/torchvision-0.22.0+9eb57cd-cp311-cp311-win_amd64.whl https://github.com/scottt/rocm-TheRock/releases/download/v6.5.0rc-pytorch-gfx110x/torchaudio-2.7.0a0+52638ef-cp311-cp311-win_amd64.whl",
"linux": "torch torchvision torchaudio --index-url https://download.pytorch.org/whl/rocm6.2"
}
},
"mps": {
"label": "Torch (MPS / Apple Silicon)",
"cmd": "torch torchvision torchaudio"
}
},
"triton": {
"v33": {
"label": "Triton < 3.3",
"cmd": {
"win": "-U \"triton-windows<3.3\"",
"linux": "-U \"triton<3.3\""
}
},
"v34": {
"label": "Triton < 3.4",
"cmd": {
"win": "-U \"triton-windows<3.4\"",
"linux": "-U \"triton<3.4\""
}
},
"latest": {
"label": "Triton Latest",
"cmd": {
"win": "-U triton-windows",
"linux": "-U triton"
}
}
},
"sage": {
"v1": { "label": "Sage Attention 1.0.6", "cmd": "sageattention==1.0.6" },
"v211": {
"label": "Sage Attention 2.1.1",
"cmd": {
"win": "https://github.com/woct0rdho/SageAttention/releases/download/v2.1.1-windows/sageattention-2.1.1+cu126torch2.6.0-cp310-cp310-win_amd64.whl",
"linux": "pip install \"setuptools<=75.8.2\" ninja wheel --force-reinstall && git clone https://github.com/thu-ml/SageAttention && pip install --no-build-isolation -e SageAttention"
}
},
"v220": {
"label": "Sage Attention 2.2.0",
"cmd": {
"win": "https://github.com/woct0rdho/SageAttention/releases/download/v2.2.0-windows/sageattention-2.2.0+cu128torch2.7.1-cp310-cp310-win_amd64.whl",
"linux": "pip install \"setuptools<=75.8.2\" ninja wheel --force-reinstall && git clone https://github.com/thu-ml/SageAttention && pip install --no-build-isolation -e SageAttention"
}
},
"v220_cu13": {
"label": "Sage Attention 2.2.0 (CUDA 13)",
"cmd": {
"win": "https://github.com/woct0rdho/SageAttention/releases/download/v2.2.0-windows.post4/sageattention-2.2.0+cu130torch2.9.0andhigher.post4-cp39-abi3-win_amd64.whl",
"linux": "pip install \"setuptools<=75.8.2\" ninja wheel --force-reinstall && git clone https://github.com/thu-ml/SageAttention && pip install --no-build-isolation -e SageAttention"
}
}
},
"sparge": {
"v010_cu128": {
"label": "Sparge Attention 0.1.0 (CUDA 12.8 / Torch 2.7.1)",
"cmd": {
"win": "https://github.com/woct0rdho/SpargeAttn/releases/download/v0.1.0-windows.post3/spas_sage_attn-0.1.0%2Bcu128torch2.7.1.post3-cp39-abi3-win_amd64.whl",
"linux": "{pip} ninja wheel packaging && {pip} --no-build-isolation git+https://github.com/woct0rdho/SpargeAttn.git"
}
},
"v010_cu13": {
"label": "Sparge Attention 0.1.0 (CUDA 13 / Torch 2.9+ ABI)",
"cmd": {
"win": "https://github.com/woct0rdho/SpargeAttn/releases/download/v0.1.0-windows.post4/spas_sage_attn-0.1.0%2Bcu130torch2.9.0andhigher.post4-cp39-abi3-win_amd64.whl",
"linux": "{pip} ninja wheel packaging && {pip} --no-build-isolation git+https://github.com/woct0rdho/SpargeAttn.git"
}
}
},
"flash": {
"v210": {
"label": "Flash Attention 2.8.x",
"cmd": {
"win": "https://github.com/deepbeepmeep/kernels/releases/download/Flash2/flash_attn-2.8.3-cp311-cp311-win_amd64.whl",
"linux": "flash-attn"
}
}
},
"kernels": {
"nunchaku": {
"label": "Nunchaku INT4/FP4",
"cmd": {
"win": "https://github.com/deepbeepmeep/kernels/releases/download/v1.2.0_Nunchaku/nunchaku-1.2.0+torch2.7-cp310-cp310-win_amd64.whl",
"linux": "https://github.com/deepbeepmeep/kernels/releases/download/v1.2.0_Nunchaku/nunchaku-1.2.0+torch2.7-cp310-cp310-linux_x86_64.whl"
}
},
"nunchaku_cu13": {
"label": "Nunchaku INT4/FP4 (CUDA 13)",
"cmd": {
"win": "https://github.com/nunchaku-ai/nunchaku/releases/download/v1.2.1/nunchaku-1.2.1+cu13.0torch2.10-cp311-cp311-win_amd64.whl",
"linux": "https://github.com/nunchaku-ai/nunchaku/releases/download/v1.2.1/nunchaku-1.2.1+cu13.0torch2.10-cp311-cp311-linux_x86_64.whl"
}
},
"light2xv": {
"label": "Light2xv NVFP4 (RTX 50xx)",
"cmd": {
"win": "https://github.com/deepbeepmeep/kernels/releases/download/Light2xv/lightx2v_kernel-0.0.2+torch2.10.0-cp311-abi3-win_amd64.whl",
"linux": "https://github.com/deepbeepmeep/kernels/releases/download/Light2xv/lightx2v_kernel-0.0.2+torch2.10.0-cp311-abi3-linux_x86_64.whl"
}
},
"gguf": {
"label": "GGUF llama.cpp CUDA Kernels",
"cmd": {
"win": "https://github.com/deepbeepmeep/kernels/releases/download/GGUF_Kernels/llamacpp_gguf_cuda-1.0.2+torch210cu13py311-cp311-cp311-win_amd64.whl",
"linux": "https://github.com/deepbeepmeep/kernels/releases/download/GGUF_Kernels/llamacpp_gguf_cuda-1.0.2+torch210cu13py311-cp311-cp311-linux_x86_64.whl"
}
}
}
},
"gpu_profiles": {
"GTX_10": { "python": "3.10", "torch": "cu128", "triton": null, "sage": null, "sparge": null, "flash": null, "kernels": [] },
"RTX_20": { "python": "3.11", "torch": "cu130", "triton": "latest", "sage": "v1", "sparge": null, "flash": "v210", "kernels": ["nunchaku_cu13", "gguf"] },
"RTX_30": { "python": "3.11", "torch": "cu130", "triton": "latest", "sage": "v220_cu13", "sparge": "v010_cu13", "flash": "v210", "kernels": ["nunchaku_cu13", "gguf"] },
"RTX_40": { "python": "3.11", "torch": "cu130", "triton": "latest", "sage": "v220_cu13", "sparge": "v010_cu13", "flash": "v210", "kernels": ["nunchaku_cu13", "gguf"] },
"RTX_50": { "python": "3.11", "torch": "cu130", "triton": "latest", "sage": "v220_cu13", "sparge": "v010_cu13", "flash": "v210", "kernels": ["nunchaku_cu13","light2xv", "gguf"] },
"AMD_GFX110X": { "python": "3.11", "torch": "rocm65", "triton": null, "sage": null, "sparge": null, "flash": null, "kernels": [], "env": {"HSA_OVERRIDE_GFX_VERSION": "11.0.0"} },
"AMD_GFX1151": { "python": "3.11", "torch": "rocm65", "triton": null, "sage": null, "sparge": null, "flash": null, "kernels": [], "env": {"HSA_OVERRIDE_GFX_VERSION": "11.5.1"} },
"AMD_GFX1201": { "python": "3.11", "torch": "rocm65", "triton": null, "sage": null, "sparge": null, "flash": null, "kernels": [], "env": {"HSA_OVERRIDE_GFX_VERSION": "12.0.1"} },
"MPS": { "python": "3.11", "torch": "mps", "triton": null, "sage": null, "sparge": null, "flash": null, "kernels": [], "attention": "sdpa" }
}
}
|