Delete load_model.py with huggingface_hub
Browse files- load_model.py +0 -42
load_model.py
DELETED
|
@@ -1,42 +0,0 @@
|
|
| 1 |
-
"""Load Hydra BitNet model."""
|
| 2 |
-
import torch
|
| 3 |
-
from safetensors.torch import load_file
|
| 4 |
-
|
| 5 |
-
def load_hydra(model_path: str, device: str = "cpu"):
|
| 6 |
-
"""Load Hydra model from HuggingFace format."""
|
| 7 |
-
import sys
|
| 8 |
-
from pathlib import Path
|
| 9 |
-
|
| 10 |
-
# Add aisim to path if needed
|
| 11 |
-
aisim_path = Path(__file__).parent.parent / "aisim"
|
| 12 |
-
if aisim_path.exists():
|
| 13 |
-
sys.path.insert(0, str(aisim_path))
|
| 14 |
-
|
| 15 |
-
from bitnet_moe import M2MSentinel
|
| 16 |
-
import json
|
| 17 |
-
|
| 18 |
-
# Load config
|
| 19 |
-
with open(f"{model_path}/config.json") as f:
|
| 20 |
-
config = json.load(f)
|
| 21 |
-
|
| 22 |
-
# Create model
|
| 23 |
-
model = M2MSentinel(
|
| 24 |
-
vocab_size=config["vocab_size"],
|
| 25 |
-
dim=config["hidden_size"],
|
| 26 |
-
depth=config["num_hidden_layers"],
|
| 27 |
-
experts=config["num_experts"],
|
| 28 |
-
)
|
| 29 |
-
|
| 30 |
-
# Load weights
|
| 31 |
-
weights = load_file(f"{model_path}/model.safetensors")
|
| 32 |
-
model.load_state_dict(weights)
|
| 33 |
-
model = model.to(device)
|
| 34 |
-
model.eval()
|
| 35 |
-
|
| 36 |
-
return model, config
|
| 37 |
-
|
| 38 |
-
if __name__ == "__main__":
|
| 39 |
-
import sys
|
| 40 |
-
model_path = sys.argv[1] if len(sys.argv) > 1 else "."
|
| 41 |
-
model, config = load_hydra(model_path)
|
| 42 |
-
print(f"Loaded model: {config}")
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|