You can deploy this model and start earning money today!

Wilsonwin/
handsongpt2

Architecture: GPT-2 style transformer
Parameters: 84.6M
Vocab Size: 64,000 (Yi-1.5 tokenizer, TPU-aligned)
Max Length: 256
Layers: 6
Hidden Size: 512
Attention Heads: 8

Name: handsongpt2
Availability: InStock

Views

No views yet

handsongpt2

HandsOnGPT2 model trained on GuoFeng Webnovel Corpus using JAX/Flax on Kaggle TPU.

python

import orbax.checkpoint as ocp

checkpointer = ocp.PyTreeCheckpointer()
state = checkpointer.restore('/path/to/checkpoint')

Apache 2.0