Skip to content

Conversation

@hkvision
Copy link

@hkvision hkvision commented Apr 8, 2024

@jason-dai @shane-huang @sgwhat Please take a look to see if this is appropriate for upstream.

https://github.com/oobabooga/text-generation-webui/wiki/Contributing-guidelines Seems no need to add tests. The style checks have been verified.

image

oobabooga and others added 11 commits February 14, 2024 11:32
* Add bigdl-llm loader

* Add BigDL-LLM if-else fork for encode

* Remove 'cpu-embedding' parameter

* Migrate source code to ipex-llm

* Add command-line flags for ipex-llm in readme
Edit description of options in GUI
* update style

* minor

* fix

* revert
@hkvision
Copy link
Author

hkvision commented Apr 9, 2024

@jason-dai @shane-huang Removed some arguments, this is a cleaner version for upstream.

Copy link

@shane-huang shane-huang left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Generally LGTM.

# bitsandbytes 4-bit
group = parser.add_argument_group('bitsandbytes 4-bit')
group.add_argument('--load-in-4bit', action='store_true', help='Load the model with 4-bit precision (using bitsandbytes).')
group.add_argument('--load-in-4bit', action='store_true', help='Load the model with 4-bit precision (using bitsandbytes or ipex-llm).')

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

"using bitsandbytes or ipex-llm" is a bit confusing. Should we tell user in which case to use bitesandbytes and in which case ipex-llm?

Copy link
Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

No which case? If one is using bitsandbytes or ipex-llm, this argument will both take effect.

@shane-huang
Copy link

shall we mention how to use IPEX-LLM in webui somewhere? Need to modify some documents in webui?

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants