English

RWKV Minipile

Model Specifications

  • Architecture: RWKV
  • Vocabulary Size: 65,536
  • Embedding Size: 768
  • Number of Layers: 12
  • Context Length: 512
  • Data Type: bfloat16
  • Dataset: Minipile
  • Tokens: 20,643,840 (20 Million)

The model underwent a rigorous training regimen, completing 32 epochs to optimize performance.

Wandb

Training progress

Verifying checksums

sha512sum -c sha512sums.txt

Inference

pip install torch numpy

python inference.py
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model is not currently available via any of the supported Inference Providers.
The model cannot be deployed to the HF Inference API: The model has no library tag.

Dataset used to train leliuga/rwkv-minipile