Commits

Branch selector

 main

User selector

KerfuffleV2

Datepicker

All time

Commit History

Commits on Nov 9, 2023

• Depend on llm-samplers 0.0.7 release

  KerfuffleV2

  committed
  9df5a7e

  Copy full SHA for 9df5a7e

Commits on Nov 6, 2023

• Update to llm-samplers v0.0.7

  KerfuffleV2

  committed
  5fa9bb2

  Copy full SHA for 5fa9bb2

Commits on Aug 15, 2023

• pub use the llm-samplers crate for convenience.

  Show description for 6424e4c
  

  KerfuffleV2

  authored and

  AmineDiro

  committed
  6424e4c

  Copy full SHA for 6424e4c

• Reimplement vs llm-samplers 0.0.6

  

  KerfuffleV2

  authored and

  AmineDiro

  committed
  5ac346c

  Copy full SHA for 5ac346c

• Add optional llm_samplers sampler backend

  

  KerfuffleV2

  authored and

  AmineDiro

  committed
  643b6a7

  Copy full SHA for 643b6a7

Commits on Aug 6, 2023

• pub use the llm-samplers crate for convenience.

  Show description for 59d4ab1
  KerfuffleV2

  committed
  59d4ab1

  Copy full SHA for 59d4ab1

• Reimplement vs llm-samplers 0.0.6

  KerfuffleV2

  committed
  500f068

  Copy full SHA for 500f068

Commits on Jul 28, 2023

• Add optional llm_samplers sampler backend

  KerfuffleV2

  committed
  ec9052e

  Copy full SHA for ec9052e

Commits on Jun 6, 2023

• Also default to 16bit memory for InferenceSessionConfig

  KerfuffleV2

  committed
  7106609

  Copy full SHA for 7106609

Commits on Jun 5, 2023

• Default to f16 model memory k/v in llm CLI

  Show description for 5934195
  KerfuffleV2

  committed
  5934195

  Copy full SHA for 5934195

Commits on Apr 16, 2023

• Sync to GGML version as of 20230416 08:55 AM UTC.

  KerfuffleV2

  committed
  3612dc4

  Copy full SHA for 3612dc4

Commits on Apr 14, 2023

• Sync to GGML version as of 20230414 06:00 AM UTC.

  Show description for e8ee132
  KerfuffleV2

  committed
  e8ee132

  Copy full SHA for e8ee132

Commits on Apr 10, 2023

• Sync to GGML version as of 20230410 10:00 AM UTC.

  KerfuffleV2

  committed
  44fbca2

  Copy full SHA for 44fbca2

Commits on Apr 7, 2023

• Update GGML CREDITS.txt with commit.

  KerfuffleV2

  committed
  6ea8882

  Copy full SHA for 6ea8882

• Sync to llama.cpp + GGML version as of 20230407 09:57 AM UTC.

  KerfuffleV2

  committed
  63fef50

  Copy full SHA for 63fef50

Commits on Mar 26, 2023

• Copy v_transposed like llama.cpp (#68)

  Show description for b103dcd
  KerfuffleV2

  authored
  b103dcd

  Copy full SHA for b103dcd

Commits on Mar 25, 2023

• Just use serde_bytes for serializing memory k/v in sessions.

  KerfuffleV2

  committed
  2f168b7

  Copy full SHA for 2f168b7

• Add zstd compression support for session loading/saving.

  KerfuffleV2

  committed
  700bb93

  Copy full SHA for 700bb93

Commits on Mar 24, 2023

• Strip off last \r\n in the prompt file.

  Show description for ed1f013
  KerfuffleV2

  committed
  ed1f013

  Copy full SHA for ed1f013

Commits on Mar 23, 2023

• Rename flag to --token-bias, add --ignore-eos alias.

  KerfuffleV2

  committed
  67950ec

  Copy full SHA for 67950ec

• Change TokenBias::get to return Option.

  Show description for 8e94159
  KerfuffleV2

  committed
  8e94159

  Copy full SHA for 8e94159

• Add --bias-tokens and --dump-prompt-tokens flags and associated backend changes.

  Show description for e7de8de
  KerfuffleV2

  committed
  e7de8de

  Copy full SHA for e7de8de

Commits on Mar 22, 2023

• Change flag name from --float-16 to --float16.

  KerfuffleV2

  committed
  f0707f8

  Copy full SHA for f0707f8

Commits on Mar 21, 2023

• Add --float-16 to llama-cli and associated support to llama-rs

  Show description for 7bece83
  KerfuffleV2

  committed
  7bece83

  Copy full SHA for 7bece83