Open leeaction opened 8 months ago
If you intend to perform full parameter fine-tuning, it should be carried out on the 8*A100 (80G)
. If opting for Lora, the tuning process is feasible on 4*V100 (32G)
.
To mitigate hardware demands, consider utilizing a smaller language model, like phi-2. Fine-tuning for phi-2 is achievable on 3090 GPUs. I intend to enhance the code in the upcoming weeks to facilitate fine-tuning for phi-2.
Hi,any plan to support some decent Chinese LLM models? There are now some hight quality Chinese video instruct data now.
Hi,any plan to support some decent Chinese LLM models? There are now some hight quality Chinese video instruct data now.
We will support fine-tuning of the Qwen and mixture of experts in the update.
@jpthu17 Hoping for it! Please pin me once you guys got some upates.
I'm trying this model , please provide the minimal device requirement information , such as GPU(type/memory) CPU SSD