Update modeling_gpt_refact.py
Browse files- modeling_gpt_refact.py +12 -0
modeling_gpt_refact.py
CHANGED
@@ -503,6 +503,18 @@ class GPTRefactForCausalLM(GPTRefactPreTrainedModel):
|
|
503 |
|
504 |
# Initialize weights and apply final processing
|
505 |
self.post_init()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
506 |
|
507 |
def prepare_inputs_for_generation(self, input_ids, past_key_values=None, inputs_embeds=None, **kwargs):
|
508 |
if inputs_embeds is not None and past_key_values is None:
|
|
|
503 |
|
504 |
# Initialize weights and apply final processing
|
505 |
self.post_init()
|
506 |
+
|
507 |
+
# gradient checkpointing support for lower versions of transformers
|
508 |
+
import transformers
|
509 |
+
from packaging import version
|
510 |
+
|
511 |
+
def _set_gradient_checkpointing(module, enable=False):
|
512 |
+
if isinstance(module, GPTRefactModel):
|
513 |
+
module.gradient_checkpointing = enable
|
514 |
+
|
515 |
+
v = version.parse(transformers.__version__)
|
516 |
+
if v.major <= 4 and v.minor < 35:
|
517 |
+
self._set_gradient_checkpointing = _set_gradient_checkpointing
|
518 |
|
519 |
def prepare_inputs_for_generation(self, input_ids, past_key_values=None, inputs_embeds=None, **kwargs):
|
520 |
if inputs_embeds is not None and past_key_values is None:
|