cmedina/agenx-lora-training
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix: weights_only=False for safetensors loading

Browse Source
This commit is contained in:
Christian Medina
2026-07-02 21:41:44 -04:00
parent aec7f405fe
commit 82677070ff
1 changed files with 1 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
quantize_streaming.py
View File

@@ -33,7 +33,7 @@ def streaming_quantize(model_path, output_path):
# Load shard to CPU
print(" Loading shard to CPU...")
shard_state_dict = torch.load(shard_file, map_location="cpu", weights_only=True)
shard_state_dict = torch.load(shard_file, map_location="cpu", weights_only=False)
# Quantize Linear layers in this shard using both GPUs
print(" Quantizing Linear layers (both GPUs)...")