Skip to content

Activity

Showing most recent first

fix stability issue with request already being closed

keldenlpushed 1 commit to master • 6b2934d…1c8b1c1 • 
on Jun 12, 2023

add wizardlm template style

keldenlpushed 3 commits to master • cb46d55…6b2934d • 
on Jun 10, 2023

revert inference engine changes and restore sibling folder structure …

keldenlpushed 3 commits to master • 6a72280…cb46d55 • 
on Jun 9, 2023

remove ggml inference engine

keldenlpushed 8 commits to master • 334ef17…6a72280 • 
on Jun 9, 2023

add readme to models folder

keldenlpushed 1 commit to master • 547d909…334ef17 • 
on May 13, 2023

bump ggml (support for redpajama via gpt-neox & starcoder)

keldenlpushed 3 commits to master • 15236f7…547d909 • 
on May 13, 2023

ggml routes now use inference engine. added scripts for inference eng…

keldenlpushed 2 commits to master • 42f9316…15236f7 • 
on May 13, 2023

fix bug by renaming chatEngine to ChatEngine

keldenlpushed 1 commit to master • 4781f29…42f9316 • 
on May 12, 2023

remove requirement to pass in ggml, smart detection via model path

keldenlpushed 2 commits to master • 7a38f42…4781f29 • 
on May 11, 2023

add ggml as a submodule

keldenlpushed 5 commits to master • 751eba5…7a38f42 • 
on May 11, 2023

clean up completions ggml route

keldenlpushed 3 commits to master • a954958…751eba5 • 
on May 11, 2023

fix chat routes to use modeltype

keldenlpushed 2 commits to master • c2fd630…a954958 • 
on May 11, 2023

fix ggml route in chatcompletios and trim request to allow for better…

keldenlpushed 1 commit to master • eb03776…c2fd630 • 
on May 10, 2023

polish chatengine for redpajama

keldenlpushed 4 commits to master • 45a43ce…eb03776 • 
on May 10, 2023

give more leeway for chat interactions

keldenlpushed 1 commit to master • 0ea6dc7…45a43ce • 
on May 9, 2023

oops push util changes

keldenlpushed 1 commit to master • c3910fa…0ea6dc7 • 
on May 8, 2023

improve filtering initial prompt, actually use model paths from api k…

keldenlpushed 1 commit to master • cbb4f49…c3910fa • 
on May 8, 2023

init ggml support (PLS DO NOT USE, VERY EXPERIMENTAL AND ROUGH AND NO…

keldenlpushed 1 commit to master • 86f0f0e…cbb4f49 • 
on May 8, 2023

update readme

keldenlpushed 1 commit to master • 3d2562b…86f0f0e • 
on May 7, 2023

fix completion bug where streaming always cuts short

keldenlpushed 1 commit to master • 0ae14cc…3d2562b • 
on May 6, 2023

add back context default param. llama has a default of 2048 max conte…

keldenlpushed 2 commits to master • f615c1e…0ae14cc • 
on May 6, 2023

update stderr to only print when readstream closes unexpectely

keldenlpushed 1 commit to master • ad51f20…f615c1e • 
on May 6, 2023

fix little merge oopsie

keldenlpushed 1 commit to master • 1c60707…ad51f20 • 
on May 6, 2023

Merge pull request #25 from swg/stderr-newline-n_predict

Pull request merge
keldenlpushed 4 commits to master • 0f7193e…1c60707 • 
on May 6, 2023

Merge pull request #39 from eiriklv/fix-various-issues

Pull request merge
keldenlpushed 7 commits to master • 81e5319…0f7193e • 
on May 6, 2023

update chatroutes to also use prompt length to caluclate response sta…

keldenlpushed 2 commits to master • c6ee4da…81e5319 • 
on May 6, 2023

default 1000 tokens

keldenlpushed 1 commit to master • b758952…c6ee4da • 
on May 5, 2023

FIX BUG WHERE NULL MAX_TOKENS WILL NOT GENERATE A PROPER RESPONSE!

keldenlpushed 1 commit to master • 7284d04…b758952 • 
on May 5, 2023

add mirostat as a parameter

keldenlpushed 1 commit to master • 4ff550f…7284d04 • 
on May 4, 2023

ADD SCRIPT

keldenlpushed 1 commit to master • 84f7f93…4ff550f • 
on May 4, 2023