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

feat: add PYTORCH_CUDA_ALLOC_CONF for memory optimization

Browse Source
This commit is contained in:
Christian Medina
2026-07-01 17:01:28 -04:00
parent 19f405255c
commit 83b45c2a5c
1 changed files with 2 additions and 1 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
train-on-this-server.sh
View File

@@ -18,8 +18,9 @@ pip install --upgrade pip
echo "Installing training dependencies..." echo "Installing training dependencies..."
pip install transformers datasets trl peft accelerate bitsandbytes deepspeed pip install transformers datasets trl peft accelerate bitsandbytes deepspeed
# Run training # Run training with CUDA memory optimization
echo "Starting training with accelerate..." echo "Starting training with accelerate..."
export PYTORCH_CUDA_ALLOC_CONF=expandable_segments:True
accelerate launch --multi_gpu --num_processes 2 train.py --config training/configs/ornith-35b-lora.yaml accelerate launch --multi_gpu --num_processes 2 train.py --config training/configs/ornith-35b-lora.yaml
echo "Training completed!" echo "Training completed!"