Code
Hub
Workspaces
Following
Trending
Connect
MCP
copy
Create free account
hub
/
github.com/deepspeedai/DeepSpeed
/ empty_cache
Method
empty_cache
accelerator/hpu_accelerator.py:120–121 ·
view source on GitHub ↗
(self)
Source
from the content-addressed store, hash-verified
118
119
# Memory management
120
def
empty_cache(self):
121
return
122
123
def
memory_allocated(self, device_index=None):
124
return
self.hpu.memory_allocated()
Callers
15
test
Method · 0.45
test_nccl_backend.py
File · 0.45
test_mpi_backend.py
File · 0.45
test_compressed_backend.py
File · 0.45
cprofileme
Function · 0.45
timeme
Function · 0.45
line_profileme
Function · 0.45
cprofileme
Function · 0.45
timeme
Function · 0.45
line_profileme
Function · 0.45
quantize
Method · 0.45
init_z3
Function · 0.45
Calls
no outgoing calls
Tested by
1
test
Method · 0.36