rwkv7-421M-pile
This is RWKV-7 model under flash-linear attention format.
Model Details
Model Description
- Developed by: Bo Peng, Yu Zhang, Songlin Yang, Ruichong Zhang
- Funded by: RWKV Project (Under LF AI & Data Foundation)
- Model type: RWKV7
- Language(s) (NLP): English
- License: Apache-2.0
- Parameter count: 421M
- Tokenizer: GPT-NeoX 20B tokenizer
Model Sources
- Repository: https://github.com/fla-org/flash-linear-attention ; https://github.com/BlinkDL/RWKV-LM
- Paper: RWKV: Parallelizable RNN with Transformer-level LLM Performance
- Project Page: RWKV
Uses
Install flash-linear-attention and the latest version of transformers before using this model:
pip install flash-linear-attention==0.3.0
pip install 'transformers>=4.48.0'
Direct Use
You can use this model just as any other HuggingFace models:
from transformers import AutoModelForCausalLM, AutoTokenizer
model = AutoModelForCausalLM.from_pretrained('fla-hub/rwkv7-421M-pile', trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained('fla-hub/rwkv7-421M-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 7.21 acc 57.9%
piqa: acc 69.2%
FAQ
Q: safetensors metadata is none.
A: upgrade transformers to >=4.48.0: pip install 'transformers>=4.48.0'