---
language:
- en
tags:
- pytorch
- text-generation
- causal-lm
- rwkv
license: apache-2.0
datasets:
- the_pile

---

# RWKV-4 14B

## Model Description

RWKV-4 14B is a L40-D5120 causal language model trained on the Pile. See https://github.com/BlinkDL/RWKV-LM for details.

** Note: It's a BF16 model, and it may overflow if you are using FP16 (fixable by rescaling the weights). **

At this moment you have to use my Github code (https://github.com/BlinkDL/RWKV-LM) to run it.

ctx_len = 1024
n_layer = 40
n_embd = 5120

Preview checkpoint: RWKV-4-Pile-14B-20221217-3794.pth : Trained on the Pile for 156B tokens.
* Pile loss 1.8189
* LAMBADA ppl 4.04, acc 69.47%
* PIQA acc 77.20%
* SC2016 acc 74.35%
* Hellaswag acc_norm 67.76%
* WinoGrande acc 63.54%