Skip to content
This repository has been archived by the owner on Oct 3, 2021. It is now read-only.

gpt2_client v2.0
Compare
Choose a tag to compare
@rish-16 rish-16 released this 08 Oct 05:28
· 44 commits to master since this release
2.0
f200cb5

Updates:

  1. Allows prompts to be parsed through an array: generate_batch_from_prompts(...) (popular demand)
  2. Support for the recently-released 774M parameter model
  3. Checks if the model assets exist before downloading (load_model(...))
  4. Allows you to specify a maximum word count for generated text
  5. Contains tutorials on new functionality
  6. Code is now more modular

Changes from gpt2_client v1.0:

  • download_model(...) is now load_model(...)

Note: See README and code for additional details and documentation

Patches:

  • Windows users expressed concerns about the emojis in the README that were messing with the installation process. This update no longer features emojis in any component that needs to be downloaded by a foreign system.
Assets 2
Loading