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

Jump to bottom

[NLP] Use helpers from utils_funcs.py in Llama conversion #7979

Merged
merged 1 commit into from
Dec 7, 2023
Merged

[NLP] Use helpers from utils_funcs.py in Llama conversion #7979

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
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
8 changes: 7 additions & 1 deletion nemo/collections/nlp/parts/utils_funcs.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,7 +12,13 @@
# See the License for the specific language governing permissions and
# limitations under the License.

__all__ = ['list2str', 'tensor2list', 'plot_confusion_matrix', 'get_classification_report']
__all__ = [
'torch_dtype_from_precision',
'list2str',
'tensor2list',
'plot_confusion_matrix',
'get_classification_report',
]

import os
import time
Expand Down
11 changes: 2 additions & 9 deletions scripts/nlp_language_modeling/convert_hf_llama_to_nemo.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -41,6 +41,7 @@
NLPSaveRestoreConnector,
PipelineMixedPrecisionPlugin,
)
from nemo.collections.nlp.parts.utils_funcs import torch_dtype_from_precision
from nemo.utils import logging


Expand Down Expand Up @@ -170,15 +171,6 @@ def convert(args):
else:
plugins.append(PipelineMixedPrecisionPlugin(precision=plugin_precision, device='cuda', scaler=scaler))

if precision == 32:
dtype = torch.float32
elif precision in [16, "16", "16-mixed"]:
dtype = torch.float16
elif precision in ["bf16", "bf16-mixed"]:
dtype = torch.bfloat16
else:
dtype = torch.float32 # fallback

nemo_config.precision = precision
print(f"nemo_config: {nemo_config}")

Expand Down Expand Up @@ -315,6 +307,7 @@ def convert(args):
model._save_restore_connector = NLPSaveRestoreConnector()

# cast to target precision and disable cpu init
dtype = torch_dtype_from_precision(precision)
model = model.to(dtype=dtype)
model.cfg.use_cpu_initialization = False

Expand Down
Loading