Yiran Wang speedcell4
speedcell4
/ gather_lse.py
Created
December 21, 2025 13:26
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| from typing import Any | |
| import torch | |
| import triton | |
| import triton.language as tl | |
| from torch import Tensor | |
| from torch.nn import functional as F | |
| from torch.testing import assert_close | |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| python3 -m pip install pip setuptools pytest hypothesis --no-cache-dir | |
| python3 -m pip install torch torchvision torchaudio triton transformers datasets tokenizers liger-kernel --no-cache-dir | |
| python3 -m pip install packaging ninja aku chew torchdevice torchgather torchnyen --no-cache-dir | |
| MAX_JOBS=$(nproc) python3 -m pip install flash-attn --no-build-isolation --no-cache-dir -v | |
| python3 -m pip install git+https://github.com/speedcell4/torchlatent.git@develop | |
| python3 -m pip install git+https://github.com/speedcell4/torchglyph.git@develop | |
| python3 -m pip install git+https://github.com/speedcell4/torchrua.git@develop | |
| python3 -m pip install git+https://github.com/speedcell4/torchshya.git --no-deps |