Closed dongs0104 closed 2 months ago
Stay on topic with Classifier-Free Guidance CFG brings non trivial improvements on many standard benchmarks.
The response quality of LLMs using CFG averaged similarly to standard models twice the size. For example, LLaMA-7B with CFG outperformed LLaMA-13B on the Lambada text-generation benchmark. This is at the expense of inference time.
authors contribute on transformers https://github.com/huggingface/transformers/pull/24654
I'm not ready to contribution on TGI. this is a backlog.
This issue is stale because it has been open 30 days with no activity. Remove stale label or comment or this will be closed in 5 days.
Feature request
Stay on topic with Classifier-Free Guidance CFG brings non trivial improvements on many standard benchmarks.
Motivation
The response quality of LLMs using CFG averaged similarly to standard models twice the size. For example, LLaMA-7B with CFG outperformed LLaMA-13B on the Lambada text-generation benchmark. This is at the expense of inference time.
authors contribute on transformers https://github.com/huggingface/transformers/pull/24654
Your contribution
I'm not ready to contribution on TGI. this is a backlog.