github/ColossalAI
1
0
Fork 0
mirror of https://github.com/hpcaitech/ColossalAI.git synced 2025-09-06 11:32:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

[npu] use extension for op builder (#5172)

Browse Source 
* update extension

* update cpu adam

* update is

* add doc for cpu adam

* update kernel

* update commit

* update flash

* update memory efficient

* update flash attn

* update flash attention loader

* update api

* fix

* update doc

* update example time limit

* reverse change

* fix doc

* remove useless kernel

* fix

* not use warning

* update

* update
This commit is contained in:
Xuanlei Zhao
2024-01-08 11:39:16 +08:00
committed by GitHub
parent d6df19bae7
commit dd2c28a323
35 changed files with 1067 additions and 274 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
colossalai/shardformer/layer/utils.py
View File

@@ -6,7 +6,8 @@ import torch.distributed as dist
from torch import nn
from torch._utils import _flatten_dense_tensors, _unflatten_dense_tensors
from torch.distributed import ProcessGroup, get_world_size
from colossalai.utils.device import get_current_device, get_rng_state, set_rng_state, manual_seed
from colossalai.utils.device import get_current_device, get_rng_state, manual_seed, set_rng_state
class SeqParallelUtils:
@@ -280,21 +281,3 @@ def create_randomizer_with_offset(
Randomizer.increment_index()
return Randomizer(seed=base_seed)
def get_attention_kernel():
"""
Get the attention kernel based on the device type.
"""
from colossalai.kernel.cuda_native import AttnMaskType
if torch.cuda.is_available():
from colossalai.kernel.cuda_native import ColoAttention as AttentionKernel
else:
try:
torch.npu.is_available()
from colossalai.kernel.npu import NPUColoAttention as AttentionKernel
except:
raise Exception("No available device for attention kernel!")
return AttnMaskType, AttentionKernel