Now training your 7B parameter #Llama2 model in float32 with 8 GPUs requires 7 * 4 * 8 = 224 GiB of host memory just to load it onto the GPUs.
Llama2 7B Model Training Memory Requirements on Multiple GPUs
By
–
Global AI News Aggregator
By
–
Now training your 7B parameter #Llama2 model in float32 with 8 GPUs requires 7 * 4 * 8 = 224 GiB of host memory just to load it onto the GPUs.
Leave a Reply