metadata

license: apache-2.0
datasets:
  - EleutherAI/the_pile_deduplicated
language:
  - en
metrics:
  - accuracy
base_model:
  - BlinkDL/rwkv-7-pile
pipeline_tag: text-generation

rwkv7-168m-pile

This is RWKV-7 model under flash-linear attention format.

Model Details

Model Description

Developed by: Bo Peng, Yu Zhang, Songlin Yang, Ruochong Zhang
Funded by: Shenzhen Yuanshi Intelligent Co. Ltd.
Model type: RWKV-7
Language(s) (NLP): English
License: Apache-2.0
Parameter count: 165M
Tokenizer: GPT-NeoX 20B tokenizer

Model Sources

Repository: https://github.com/fla-org/flash-linear-attention ; https://github.com/BlinkDL/RWKV-LM
Paper: With in Progress
Weights: Converted from https://modelscope.cn/models/RWKV/rwkv-7-pile/file/view/master?fileName=RWKV-x070-Pile-168M-20241120-ctx4096.pth

Uses

Install flash-linear-attention before using this model:

git clone https://github.com/fla-org/flash-linear-attention
cd flash-linear-attention
pip install -e .

Direct Use

You can use this model just as any other HuggingFace models:

from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained('fla-hub/rwkv7-168m-pile', trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained('fla-hub/rwkv7-168m-pile', trust_remote_code=True)

Training Details

Training Data

This model is trained on the Pile with a total of 332 billion tokens.

Training Hyperparameters

Training regime: bfloat16, lr 8e-4 to 3e-5 cosine decay, wd 0.1, bsz 8x30x4096

Evaluation

Metrics

lambada_openai: ppl 14.2 acc 45.6%

piqa: acc 65.5%