How to optimize model fine-tuning if it is very slow? (NVIDIA GeForce RTX 4060 Laptop GPU)
mack007liu opened this issue · 3 comments
mack007liu commented
How to optimize model fine-tuning if it is very slow? (NVIDIA GeForce RTX 4060 Laptop GPU)
mack007liu commented
mack007liu commented
ningx-ms commented
Hi @mack007liu thanks for reporting the issue.
The slow training is mostly due to low GPU memory. Mistral-7b + QLora trained on the toy dataset consumes 17GB GPU memory. You may try a smaller model like Phi-1.5 on your laptop.
Below are training time measured a few GPUs (listed in the release notes)