Benchmark RWKV on Long Range Arena

Data Prepration

listops: RWKV_T_MAX=2048 CUDA_VISIBLE_DEVICES=0,5,6,7 RWKV_FLOAT_MODE=fp32 python -m train wandb=null experiment=lra/rwkv-listops trainer.devices=4
cifar:
aan: RWKV_FLOAT_MODE=fp16 python -m train trainer.devices=8 experiment=lra/rwkv-aan wandb=null

Name		Name	Last commit message	Last commit date
Latest commit History 5 Commits
assets		assets
checkpoints		checkpoints
configs		configs
cuda		cuda
extensions/kernels		extensions/kernels
models		models
notebooks		notebooks
src		src
.gitignore		.gitignore
CHANGELOG.md		CHANGELOG.md
LICENSE		LICENSE
README.md		README.md
example.py		example.py
generate.py		generate.py
requirements.txt		requirements.txt
train.py		train.py