Code

Merge dev branch#6118

Code

Merged

oobabooga merged 26 commits intomainoobabooga/text-generation-webui:mainfrom devoobabooga/text-generation-webui:devCopy head branch name to clipboardJun 13, 2024

+497-711

Commits

Commits on May 19, 2024

• llama.cpp: increase the max threads from 32 to 256 (#5889)

  chraac

  authored
  6b546a2

  Copy full SHA for 6b546a2

• Update accelerate requirement from ==0.27.* to ==0.30.* (#5989)

  dependabot[bot]

  authored
  2de586f

  Copy full SHA for 2de586f

• UI: Warn user if they are trying to load a model from no path (#6006)

  poshul

  authored
  b63dc4e

  Copy full SHA for b63dc4e

• [docs] small docker changes (#5917)

  jvanmelckebeke

  authored
  8456d13

  Copy full SHA for 8456d13

• fix: grammar not support utf-8 (#5900)

  A0nameless0man

  authored
  5cb5970

  Copy full SHA for 5cb5970

• Add Llama 3 instruction template (#5891)

  Touch-Night

  authored
  d7bd3da

  Copy full SHA for d7bd3da

• Fix gguf multipart file loading (#5857)

  Tisjwlf

  authored
  907702c

  Copy full SHA for 907702c

• Let grammar escape backslashes (#5865)

  altoiddealer

  authored
  818b4e0

  Copy full SHA for 818b4e0

Commits on May 20, 2024

• --idle-timeout flag to unload the model if unused for N minutes (#6026)

  oobabooga

  authored
  9f77ed1

  Copy full SHA for 9f77ed1

• DRY: A modern repetition penalty that reliably prevents looping (#5677)

  p-e-w

  authored
  852c943

  Copy full SHA for 852c943

• README: update command-line flags with raw --help output

  Show description for 6a1682a
  oobabooga

  committed
  6a1682a

  Copy full SHA for 6a1682a

Commits on May 21, 2024

• Backend cleanup (#6025)

  oobabooga

  authored
  bd7cc42

  Copy full SHA for bd7cc42

• Fix getting Phi-3-small-128k-instruct logits

  oobabooga

  committed
  ae86292

  Copy full SHA for ae86292

• Minor fix after bd7cc42 (thanks @belladoreai)

  oobabooga

  committed
  9e18994

  Copy full SHA for 9e18994

• Fixed minor typo in docs - Training Tab.md (#6038)

  iamrohitanshu

  authored
  8aaa0a6

  Copy full SHA for 8aaa0a6

Commits on May 22, 2024

• Fix stopping strings for llama-3 and phi (#6043)

  oobabooga

  authored
  5499bc9

  Copy full SHA for 5499bc9

Commits on May 23, 2024

• Revert "Fix stopping strings for llama-3 and phi (#6043)"

  Show description for ad54d52
  oobabooga

  committed
  ad54d52

  Copy full SHA for ad54d52

Commits on May 24, 2024

• Downloader: Add --model-dir argument, respect --model-dir in the UI

  oobabooga

  committed
  4f1e96b

  Copy full SHA for 4f1e96b

Commits on May 27, 2024

• Remove MinPLogitsWarper (it's now a transformers built-in)

  oobabooga

  committed
  8df68b0

  Copy full SHA for 8df68b0

• Fix missing bos token for some models (including Llama-3) (#6050)

  belladoreai

  authored
  a363cdf

  Copy full SHA for a363cdf

Commits on Jun 13, 2024

• Fix error when bos_token_id is None. (#6061)

  belladoreai

  authored
  46174a2

  Copy full SHA for 46174a2

• Remove obsolete pre_layer parameter

  oobabooga

  committed
  2d196ed

  Copy full SHA for 2d196ed

• Lint

  oobabooga

  committed
  a36fa73

  Copy full SHA for a36fa73

• Use reentrant generation lock (#6107)

  theo77186

  authored
  b675151

  Copy full SHA for b675151

• DRY sampler improvements (#6053)

  belladoreai

  authored
  3abafee

  Copy full SHA for 3abafee

• Fix flash-attn UI parameter to actually store true. (#6076)

  Ph0rk0z

  authored
  1d79aa6

  Copy full SHA for 1d79aa6