Upload model.py with huggingface_hub
Browse files
model.py
CHANGED
@@ -109,6 +109,27 @@ class MultiModalEnsembleC(nn.Module):
|
|
109 |
with open(config_path, "w") as f:
|
110 |
json.dump(config, f)
|
111 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
112 |
def forward(self, x, y, z):
|
113 |
"""
|
114 |
Forward pass of the MultiModalEnsembleC model.
|
|
|
109 |
with open(config_path, "w") as f:
|
110 |
json.dump(config, f)
|
111 |
|
112 |
+
@classmethod
|
113 |
+
def from_pretrained(cls, load_directory):
|
114 |
+
"""
|
115 |
+
Load model weights and configuration from a directory.
|
116 |
+
|
117 |
+
Args:
|
118 |
+
load_directory (str): Path to the directory containing model weights and configuration.
|
119 |
+
|
120 |
+
Returns:
|
121 |
+
MultiModalEnsembleC: A model instance with loaded weights.
|
122 |
+
"""
|
123 |
+
config_path = os.path.join(load_directory, "config.json")
|
124 |
+
with open(config_path, "r") as f:
|
125 |
+
config = json.load(f)
|
126 |
+
|
127 |
+
# Create the model instance
|
128 |
+
model = cls(num_classes=config["num_classes"])
|
129 |
+
model_path = os.path.join(load_directory, "pytorch_model.bin")
|
130 |
+
model.load_state_dict(torch.load(model_path))
|
131 |
+
return model
|
132 |
+
|
133 |
def forward(self, x, y, z):
|
134 |
"""
|
135 |
Forward pass of the MultiModalEnsembleC model.
|