Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

stage_1_and_2.py: do gradient scale only for fp16 #3166

Merged
merged 10 commits into from
Apr 26, 2023
Merged

stage_1_and_2.py: do gradient scale only for fp16 #3166

Changes from all commits
Commits
Show all changes
10 commits
Select commit Hold shift + click to select a range
4070c4d
stage_1_and_2.py: do gradient scale only for fp16
guoyejun Apr 9, 2023
72b83ec
Merge branch 'master' into fp16_gscale
guoyejun Apr 11, 2023
6fa91aa
Merge branch 'master' into fp16_gscale
guoyejun Apr 16, 2023
996381c
Merge branch 'master' into fp16_gscale
tjruwase Apr 17, 2023
d081c97
Merge branch 'master' into fp16_gscale
tjruwase Apr 19, 2023
9cc24d9
Merge branch 'master' into fp16_gscale
guoyejun Apr 24, 2023
7d0d90e
Merge branch 'master' into fp16_gscale
tjruwase Apr 25, 2023
dbf1ab4
Merge branch 'master' into fp16_gscale
tjruwase Apr 25, 2023
acb5578
Merge branch 'master' into fp16_gscale
tjruwase Apr 25, 2023
cf56ae6
Merge branch 'master' into fp16_gscale
tjruwase Apr 26, 2023
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 11 additions & 6 deletions deepspeed/runtime/zero/stage_1_and_2.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1665,18 +1665,21 @@ def step(self, closure=None):
return

# Step 1:- Calculate gradient norm using fp-16 grads
see_memory_usage('Before norm calculation')
scaled_global_grad_norm = self.scaled_global_norm()
self._global_grad_norm = scaled_global_grad_norm / prev_scale
if self.dtype == torch.float16:
see_memory_usage('Before norm calculation')
scaled_global_grad_norm = self.scaled_global_norm()
self._global_grad_norm = scaled_global_grad_norm / prev_scale
see_memory_usage('After norm before optimizer')

see_memory_usage('After norm before optimizer')
# Step 2:- run optimizer and upscaling simultaneously
for i, group in enumerate(self.bit16_groups):
self.start_timers([OPTIMIZER_GRADIENTS])
partition_id = dist.get_rank(group=self.real_dp_process_group[i])
if self.cpu_offload:
single_grad_partition = self.single_partition_of_fp32_groups[i].grad
self.unscale_and_clip_grads([single_grad_partition], scaled_global_grad_norm)
if self.dtype == torch.float16:
self.unscale_and_clip_grads([single_grad_partition], scaled_global_grad_norm)

self.stop_timers([OPTIMIZER_GRADIENTS])
self.start_timers([OPTIMIZER_STEP])
self._optimizer_step(i)
Expand Down Expand Up @@ -1715,7 +1718,9 @@ def step(self, closure=None):

self.averaged_gradients[i] = None

self.unscale_and_clip_grads([single_grad_partition], scaled_global_grad_norm)
if self.dtype == torch.float16:
self.unscale_and_clip_grads([single_grad_partition], scaled_global_grad_norm)

self.stop_timers([OPTIMIZER_GRADIENTS])

# Step 3:- run the optimizer if no offloading
Expand Down