LayerSkip: Enabling Early Exit Inference and Self-Speculative Decoding
Paper
•
2404.16710
•
Published
•
73
LayerSkip
Models continually pretrained using LayerSkip - https://arxiv.org/abs/2404.16710
-
-
facebook/layerskip-llama2-7B
Text Generation • Updated • 207 • 10 -
facebook/layerskip-llama2-13B
Text Generation • Updated • 60 • 5 -
facebook/layerskip-llama2-70B
Text Generation • Updated • 65 • 4 -
facebook/layerskip-codellama-7B
Text Generation • Updated • 33 • 6 -
facebook/layerskip-codellama-34B
Text Generation • Updated • 48 • 4 -
facebook/layerskip-llama3-8B
Text Generation • Updated • 154 • 17 -
facebook/layerskip-llama3.2-1B
Text Generation • Updated • 446 • 15