RWKV-4 3B is a L32-D2560 causal language model trained on the Pile. See
https://github.com/BlinkDL/RWKV-LM for details.
RWKV-4-Pile-3B-20221110-ctx4096.pth (RECOMMENDED) : Fine-tuned to ctx_len 4096.
RWKV-4-Pile-3B-20221008-8023.pth : Trained on the Pile for 331B tokens.
Note I am using "Q: instruct\n\nA: result" prompt for all instructs.
RWKV-4-Pile-3B-Instruct-test2
instruct-tuned on
https://huggingface.co/datasets/Muennighoff/flan & NIv2
RWKV-4-Pile-3B-EngChn-testNovel-xxx for writing Chinese novels (trained on 200G Chinese novels.)
RWKV-4-Pile-3B-EngChn-testxxx for Chinese Q&A (trained on 10G Chinese text. only for testing purposes.)