Commits

Breadcrumbs

History for

onmain

User selector

All users

All time

Commit History

Commits on Jul 25, 2023

• Change tensor_split from array to pointer

  

  Shouyi Wang

  committed
  426dbfe

  Copy full SHA for 426dbfe

Commits on Jul 24, 2023

• Add temporary rms_norm_eps parameter

  abetlen

  committed
  11dd2bf

  Copy full SHA for 11dd2bf

• Add rms_eps_norm

  abetlen

  committed
  8cd64d4

  Copy full SHA for 8cd64d4

• add support for llama2 70b

  bretello

  committed
  0f09f10

  Copy full SHA for 0f09f10

Commits on Jul 20, 2023

• Merge pull request #481 from c0sogi/main

  Show description for 365d9a4
  abetlen

  authored
  365d9a4

  Copy full SHA for 365d9a4

• Now the last token sent when `stream=True`

  ctejada85

  committed
  0756a2d

  Copy full SHA for 0756a2d

Commits on Jul 16, 2023

• Added `RouteErrorHandler` for server

  c0sogi

  committed
  1551ba1

  Copy full SHA for 1551ba1

Commits on Jul 15, 2023

• Re-order Llama class params

  abetlen

  committed
  8ab098e

  Copy full SHA for 8ab098e

• Merge branch main into custom_rope

  abetlen

  committed
  f0797a6

  Copy full SHA for f0797a6

Commits on Jul 10, 2023

• Add bindings for custom_rope

  randoentity

  committed
  3f8f276

  Copy full SHA for 3f8f276

Commits on Jul 9, 2023

• bugfix: truncate completion max_tokens to fit context length by default

  abetlen

  committed
  a86bfdf

  Copy full SHA for a86bfdf

Commits on Jul 8, 2023

• Merge branch 'main' into add_unlimited_max_tokens

  abetlen

  authored
  5d756de

  Copy full SHA for 5d756de

• Merge pull request #453 from wu-qing-157/main

  Show description for b8e0bed
  abetlen

  authored
  b8e0bed

  Copy full SHA for b8e0bed

• bugfix: fix compatibility bug with openai api on last token

  abetlen

  committed
  d6e6aad

  Copy full SHA for d6e6aad

• Format

  abetlen

  committed
  4f2b5d0

  Copy full SHA for 4f2b5d0

• perf: convert pointer to byref

  abetlen

  committed
  34c505e

  Copy full SHA for 34c505e

Commits on Jul 7, 2023

• perf: avoid allocating new buffers during sampling

  abetlen

  committed
  11eae75

  Copy full SHA for 11eae75

• perf: assign to candidates data structure instead

  abetlen

  committed
  a14d8a9

  Copy full SHA for a14d8a9

• fix indexing token_logprobs after sorting

  wu-qing-157

  committed
  9e61661

  Copy full SHA for 9e61661

Commits on Jun 29, 2023

• Hotfix: logits_all bug

  abetlen

  committed
  e34f441

  Copy full SHA for e34f441

• Load logits directly into scores buffer

  abetlen

  committed
  a2ede37

  Copy full SHA for a2ede37

• Use pre-allocated buffers to store input_ids and scores

  abetlen

  committed
  b95b0ff

  Copy full SHA for b95b0ff

• Free model when llama is unloaded. Closes #434

  abetlen

  committed
  a5e059c

  Copy full SHA for a5e059c

Commits on Jun 26, 2023

• Merge branch 'main' of github.com:abetlen/llama_cpp_python into main

  abetlen

  committed
  3379dc4

  Copy full SHA for 3379dc4

• Update llama.cpp

  abetlen

  committed
  9522284

  Copy full SHA for 9522284

• Update type signature

  abetlen

  committed
  b4a3db3

  Copy full SHA for b4a3db3

• Merge branch 'main' into fix-state-pickle

  abetlen

  authored
  5eb4ebb

  Copy full SHA for 5eb4ebb

Commits on Jun 24, 2023

• Only concatenate after all batches are done

  samfundev

  committed
  d788fb4

  Copy full SHA for d788fb4

Commits on Jun 23, 2023

• Merge branch 'main' into fix-state-pickle

  abetlen

  authored
  877ca6d

  Copy full SHA for 877ca6d

Commits on Jun 17, 2023

• Update docs. Closes #386

  abetlen

  committed
  d410f12

  Copy full SHA for d410f12

Commits on Jun 16, 2023

• Update llama.py: Added how many input tokens in ValueError exception

  nb-programmer

  authored
  fd9f294

  Copy full SHA for fd9f294

Commits on Jun 15, 2023

• Add low_vram parameter

  abetlen

  committed
  44b83ca

  Copy full SHA for 44b83ca

• Merge pull request #351 from player1537-forks/th/add-logits-bias-parameter

  Show description for f568bae
  abetlen

  authored
  f568bae

  Copy full SHA for f568bae

Commits on Jun 13, 2023

• fix: Make LLamaState pickable for disk cache

  Show description for 10b0cb7
  Okabintaro

  committed
  10b0cb7

  Copy full SHA for 10b0cb7

Commits on Jun 10, 2023

• Re-enable cache

  abetlen

  committed
  21acd79

  Copy full SHA for 21acd79