You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
ColossalAI/colossalai
Hongxin Liu e86127925a
[plugin] support all-gather overlap for hybrid parallel (#5919)
4 months ago
..
_C Clean up 6 months ago
_analyzer [test] Fix/fix testcase (#5770) 6 months ago
accelerator [hotfix] fix typo change MoECheckpintIO to MoECheckpointIO (#5335) 9 months ago
amp [npu] change device to accelerator api (#5239) 11 months ago
auto_parallel [pre-commit.ci] pre-commit autoupdate (#5572) 5 months ago
autochunk [hotfix] Fix examples no pad token & auto parallel codegen bug; (#5606) 7 months ago
booster [plugin] support all-gather overlap for hybrid parallel (#5919) 4 months ago
checkpoint_io [plugin] support all-gather overlap for hybrid parallel (#5919) 4 months ago
cli [devops] fix extention building (#5427) 9 months ago
cluster [shardformer] DeepseekMoE support (#5871) 5 months ago
context [Fix]: implement thread-safety singleton to avoid deadlock for very large-scale training scenarios (#5625) 7 months ago
device [Feature] Distributed optimizers: Lamb, Galore, CAME and Adafactor (#5694) 7 months ago
fx [test] Fix/fix testcase (#5770) 6 months ago
inference [HotFix] CI,import,requirements-test for #5838 (#5892) 5 months ago
interface [Feature] Distributed optimizers: Lamb, Galore, CAME and Adafactor (#5694) 7 months ago
kernel [NFC] Fix code factors on inference triton kernels (#5743) 6 months ago
lazy [lazy] fix lazy cls init (#5720) 6 months ago
legacy [Hoxfix] Fix CUDA_DEVICE_MAX_CONNECTIONS for comm overlap 5 months ago
logging [misc] update pre-commit and run all files (#4752) 1 year ago
moe [MoE/ZeRO] Moe refactor with zero refactor (#5821) 5 months ago
nn [misc] fix dist logger (#5782) 6 months ago
pipeline fix object_to_tensor usage when torch>=2.3.0 (#5820) 5 months ago
quantization [quant] fix bitsandbytes version check (#5882) 5 months ago
shardformer [ShardFormer] fix qwen2 sp (#5903) 5 months ago
tensor [compatibility] support torch 2.2 (#5875) 5 months ago
testing [misc] Accelerate CI for zero and dist optim (#5758) 6 months ago
utils Merge pull request #5310 from hpcaitech/feature/npu 10 months ago
zero [zero] support all-gather overlap (#5898) 5 months ago
__init__.py [devops] remove post commit ci (#5566) 8 months ago
initialize.py [Hoxfix] Fix CUDA_DEVICE_MAX_CONNECTIONS for comm overlap 5 months ago