muellerzr/llama-3-8B-self-align

Weights from the Llama-3-8B Self-Align Experiments

[WEIGHTS TO BE UPLOADED ONCE DONE]

Training Config

The config.yaml should be used during accelerate launch, and run.sh was used to launch the training using the StarCoder2 Self-Align training script. Some tweaks were performed to get this working on 48GB vRAM:

• FSDP was used
• per_device_batch_size is 2
• A learning rate of 3e-6 was used

Environment:

• Trained with 2x4090 GPUs
• 128GB RAM