This repository has been archived by the owner on Oct 3, 2021. It is now read-only.
Releases: rish-16/gpt2client
Releases Β· rish-16/gpt2client
gpt2_client v2.1
Updates
I have added some these features into the gpt2-client
module based on feature requests on Reddit:
- Encode and decode text sequences
- Support for the 1.5B model
Changes from gpt2_client v2.0
:
- Sequences can be encoded into token vectors using the
encode_seq(...)
function and decoded using thedecode_seq(...)
function
Note: See README and code for additional details and documentation
gpt2_client v2.0
Updates:
- Allows prompts to be parsed through an array:
generate_batch_from_prompts(...)
(popular demand) - Support for the recently-released
774M
parameter model - Checks if the model assets exist before downloading (
load_model(...)
) - Allows you to specify a maximum word count for generated text
- Contains tutorials on new functionality
- Code is now more modular
Changes from gpt2_client v1.0
:
download_model(...)
is nowload_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.