Commits

Breadcrumbs

History for

onmain

User selector

All users

All time

Commit History

Commits on Apr 24, 2023

• Merge pull request #108 from eiery/main

  Show description for f374561
  abetlen

  authored
  f374561

  Copy full SHA for f374561

• Update llama.cpp

  abetlen

  committed
  02cf881

  Copy full SHA for 02cf881

• Update llama.py

  Show description for aa12d8a
  

  eiery

  authored
  aa12d8a

  Copy full SHA for aa12d8a

Commits on Apr 23, 2023

• Disable mmap when applying lora weights. Closes #107

  abetlen

  committed
  7230599

  Copy full SHA for 7230599

Commits on Apr 22, 2023

• Update llama.cpp

  abetlen

  committed
  e99caed

  Copy full SHA for e99caed

Commits on Apr 21, 2023

• Update llama.cpp

  abetlen

  committed
  1eb130a

  Copy full SHA for 1eb130a

Commits on Apr 19, 2023

• Add use_mmap flag to server

  abetlen

  committed
  e4647c7

  Copy full SHA for e4647c7

• If lora base is not set avoid re-loading the model by passing NULL

  abetlen

  committed
  0df4d69

  Copy full SHA for 0df4d69

• Update type signature to allow for null pointer to be passed.

  abetlen

  committed
  95c0dc1

  Copy full SHA for 95c0dc1

Commits on Apr 18, 2023

• Add seperate lora_base path for applying LoRA to quantized models using original unquantized model weights.

  abetlen

  committed
  453e517

  Copy full SHA for 453e517

• Add lora_path parameter to Llama model

  abetlen

  committed
  eb7f278

  Copy full SHA for eb7f278

• Add bindings for LoRA adapters. Closes #88

  abetlen

  committed
  35abf89

  Copy full SHA for 35abf89

Commits on Apr 15, 2023

• Bugfix: only eval new tokens

  abetlen

  committed
  89856ef

  Copy full SHA for 89856ef

• Add experimental cache

  abetlen

  committed
  92c0771

  Copy full SHA for 92c0771

• Update stop sequences for chat

  abetlen

  committed
  a6372a7

  Copy full SHA for a6372a7

• Update chat parameters

  abetlen

  committed
  83b2be6

  Copy full SHA for 83b2be6

• Update chat prompt

  abetlen

  committed
  6208751

  Copy full SHA for 6208751

• Bugfix

  abetlen

  committed
  02f9fb8

  Copy full SHA for 02f9fb8

• Add type annotations

  abetlen

  committed
  3cd67c7

  Copy full SHA for 3cd67c7

• Bugfix

  abetlen

  committed
  d7de0e8

  Copy full SHA for d7de0e8

• Use clear

  abetlen

  committed
  e90e122

  Copy full SHA for e90e122

• Track generated tokens internally

  abetlen

  committed
  ac7068a

  Copy full SHA for ac7068a

• Set kv cache size to f16 by default

  abetlen

  committed
  6e298d8

  Copy full SHA for 6e298d8

Commits on Apr 14, 2023

• Fix completion request

  abetlen

  committed
  6c7cec0

  Copy full SHA for 6c7cec0

• Clean up logprobs implementation

  abetlen

  committed
  6153baa

  Copy full SHA for 6153baa

• Fix signature for stop parameter

  abetlen

  committed
  26cc4ee

  Copy full SHA for 26cc4ee

Commits on Apr 13, 2023

• Add field to disable reseting between generations

  abetlen

  committed
  6595ad8

  Copy full SHA for 6595ad8

• Revert "Deprecate generate method"

  Show description for 22fa5a6
  abetlen

  committed
  22fa5a6

  Copy full SHA for 22fa5a6

• Formatting

  abetlen

  committed
  4f5f99e

  Copy full SHA for 4f5f99e

Commits on Apr 12, 2023

• Enable logprobs on completion endpoint

  abetlen

  committed
  0daf16d

  Copy full SHA for 0daf16d

• Fix threading bug. Closes #62

  abetlen

  committed
  19598ac

  Copy full SHA for 19598ac

• Update llama.cpp

  abetlen

  committed
  005c78d

  Copy full SHA for 005c78d

• Don't serialize stateful parameters

  abetlen

  committed
  c854c25

  Copy full SHA for c854c25

• Style fix

  abetlen

  committed
  2f9b649

  Copy full SHA for 2f9b649

• Deprecate generate method

  abetlen

  committed
  6cf5876

  Copy full SHA for 6cf5876