Skip to content
GitLab
Menu
Projects
Groups
Snippets
Loading...
Help
Help
Support
Community forum
Keyboard shortcuts
?
Submit feedback
Contribute to GitLab
Sign in / Register
Toggle navigation
Menu
Open sidebar
chenpangpang
transformers
Commits
344e2664
Unverified
Commit
344e2664
authored
Oct 17, 2022
by
Sylvain Gugger
Committed by
GitHub
Oct 17, 2022
Browse files
Fix dtype in radnomly initialized head (#19690)
parent
07f66902
Changes
1
Hide whitespace changes
Inline
Side-by-side
Showing
1 changed file
with
2 additions
and
2 deletions
+2
-2
src/transformers/modeling_utils.py
src/transformers/modeling_utils.py
+2
-2
No files found.
src/transformers/modeling_utils.py
View file @
344e2664
...
...
@@ -2446,9 +2446,9 @@ class PreTrainedModel(nn.Module, ModuleUtilsMixin, GenerationMixin, PushToHubMix
param
=
model_state_dict
[
key
]
if
param
.
device
==
torch
.
device
(
"meta"
):
if
not
load_in_8bit
:
set_module_tensor_to_device
(
model
,
key
,
"cpu"
,
torch
.
empty
(
*
param
.
size
()))
set_module_tensor_to_device
(
model
,
key
,
"cpu"
,
torch
.
empty
(
*
param
.
size
()
,
dtype
=
dtype
))
else
:
set_module_8bit_tensor_to_device
(
model
,
key
,
"cpu"
,
torch
.
empty
(
*
param
.
size
()))
set_module_8bit_tensor_to_device
(
model
,
key
,
"cpu"
,
torch
.
empty
(
*
param
.
size
()
,
dtype
=
dtype
))
# retrieve unintialized modules and initialize before maybe overriding that with the pretrained weights.
if
_fast_init
:
...
...
Write
Preview
Markdown
is supported
0%
Try again
or
attach a new file
.
Attach a file
Cancel
You are about to add
0
people
to the discussion. Proceed with caution.
Finish editing this message first!
Cancel
Please
register
or
sign in
to comment