.. |
activation_checkpointing
|
a10e4811fe
force set lf instead of crlf (https://github.com/pre-commit/pre-commit-hooks#mixed-line-ending) (#1598)
|
2 年之前 |
comm
|
67a48aaa89
1-bit LAMB optimizer (#970)
|
3 年之前 |
compression
|
68c8481bcf
1-bit Adam v2 (#817)
|
3 年之前 |
data_pipeline
|
fbea7b493f
CL for big science (#1440)
|
3 年之前 |
fp16
|
0fc11fa0e6
[squash] zero-ckpt-cpu-issue (#1673)
|
2 年之前 |
pipe
|
a8a17f234a
Several fixes for our read-the-docs build (#1579)
|
2 年之前 |
swap_tensor
|
488105ebd2
Fix zinf none swapper (#1550)
|
2 年之前 |
zero
|
dbd08236a6
formatting
|
2 年之前 |
__init__.py
|
e5bbc2e559
Sparse attn + ops/runtime refactor + v0.3.0 (#343)
|
4 年之前 |
config.py
|
fc2f378ece
Improve pre-commit hooks (#1602)
|
2 年之前 |
config_utils.py
|
a10e4811fe
force set lf instead of crlf (https://github.com/pre-commit/pre-commit-hooks#mixed-line-ending) (#1598)
|
2 年之前 |
constants.py
|
0fc11fa0e6
[squash] zero-ckpt-cpu-issue (#1673)
|
2 年之前 |
dataloader.py
|
c0b27fb019
Added drop_last to DeepSpeedDataLoader (#1321)
|
3 年之前 |
eigenvalue.py
|
a10e4811fe
force set lf instead of crlf (https://github.com/pre-commit/pre-commit-hooks#mixed-line-ending) (#1598)
|
2 年之前 |
engine.py
|
0fc11fa0e6
[squash] zero-ckpt-cpu-issue (#1673)
|
2 年之前 |
lr_schedules.py
|
76847f42cf
Add warmup_type arguments in WarmupLR and WarmupDecayLR (#1530)
|
2 年之前 |
progressive_layer_drop.py
|
a10e4811fe
force set lf instead of crlf (https://github.com/pre-commit/pre-commit-hooks#mixed-line-ending) (#1598)
|
2 年之前 |
quantize.py
|
a10e4811fe
force set lf instead of crlf (https://github.com/pre-commit/pre-commit-hooks#mixed-line-ending) (#1598)
|
2 年之前 |
sparse_tensor.py
|
fcb3ca5e66
Proposal of how we might use sparse tensors for gradients (#1484)
|
3 年之前 |
state_dict_factory.py
|
be789b1665
Fix many typos (#1423)
|
3 年之前 |
utils.py
|
fc2f378ece
Improve pre-commit hooks (#1602)
|
2 年之前 |
weight_quantizer.py
|
d2cf66a668
release inference quantized kernels (#1104)
|
3 年之前 |