Upload visual_encoder.py with huggingface_hub
Browse files- visual_encoder.py +2 -1
visual_encoder.py
CHANGED
|
@@ -383,6 +383,7 @@ class MplugOwlVisionEncoder(nn.Module):
|
|
| 383 |
|
| 384 |
class MplugOwlVisionModel(PreTrainedModel):
|
| 385 |
main_input_name = "pixel_values"
|
|
|
|
| 386 |
|
| 387 |
def __init__(self, config):
|
| 388 |
super().__init__(config)
|
|
@@ -690,7 +691,6 @@ class MplugOwlVisualAbstractorLayer(nn.Module):
|
|
| 690 |
|
| 691 |
|
| 692 |
class MplugOwlVisualAbstractorEncoder(nn.Module):
|
| 693 |
-
_no_split_modules = ["MplugOwlVisualAbstractorLayer"]
|
| 694 |
def __init__(self, config):
|
| 695 |
super().__init__()
|
| 696 |
self.config = config
|
|
@@ -755,6 +755,7 @@ class MplugOwlVisualAbstractorEncoder(nn.Module):
|
|
| 755 |
|
| 756 |
|
| 757 |
class MplugOwlVisualAbstractorModel(PreTrainedModel):
|
|
|
|
| 758 |
def __init__(self, config, language_hidden_size):
|
| 759 |
super().__init__(config)
|
| 760 |
self.config = config
|
|
|
|
| 383 |
|
| 384 |
class MplugOwlVisionModel(PreTrainedModel):
|
| 385 |
main_input_name = "pixel_values"
|
| 386 |
+
_no_split_modules = ["MplugOwlVisionEncoderLayer"]
|
| 387 |
|
| 388 |
def __init__(self, config):
|
| 389 |
super().__init__(config)
|
|
|
|
| 691 |
|
| 692 |
|
| 693 |
class MplugOwlVisualAbstractorEncoder(nn.Module):
|
|
|
|
| 694 |
def __init__(self, config):
|
| 695 |
super().__init__()
|
| 696 |
self.config = config
|
|
|
|
| 755 |
|
| 756 |
|
| 757 |
class MplugOwlVisualAbstractorModel(PreTrainedModel):
|
| 758 |
+
_no_split_modules = ["MplugOwlVisualAbstractorLayer"]
|
| 759 |
def __init__(self, config, language_hidden_size):
|
| 760 |
super().__init__(config)
|
| 761 |
self.config = config
|