Commits

Breadcrumbs

History for

onpatch-2

User selector

All users

All time

Commit History

Commits on May 31, 2026

• fix: avoid cleanup errors for partially initialized LlamaModel (#2173)

  Show description for fdf38b3
  

  usernames122

  and

  abetlen

  authored
  fdf38b3

  Copy full SHA for fdf38b3

Commits on May 4, 2026

• fix(_internals): use n_tokens0 offset when enabling last-token logits in add_sequence (#2205)

  Show description for 90e8df9
  Anai-Guo

  authored
  90e8df9

  Copy full SHA for 90e8df9

Commits on Apr 3, 2026

• fix(misc): replace deprecated llama.cpp references (#2170)

  Show description for 08e088c
  abetlen

  authored
  08e088c

  Copy full SHA for 08e088c

Commits on Mar 25, 2026

• fix: handle embedding models without KV memory (#2160)

  Show description for ac59e5a
  abetlen

  authored
  ac59e5a

  Copy full SHA for ac59e5a

Commits on Mar 23, 2026

• fix: Qwen 3.5 support (#2152)

  Show description for 11e7a55
  abetlen

  authored
  11e7a55

  Copy full SHA for 11e7a55

• feat: Update llama.cpp to ggerganov/llama.cpp@49bfddeca18e62fa3d39114a23e9fcbdf8a22388 (#2151)

  Show description for 18aa31e
  abetlen

  authored
  18aa31e

  Copy full SHA for 18aa31e

Commits on Mar 22, 2026

• misc: Add Ruff formatting (#2148)

  Show description for a9b4a06
  abetlen

  authored
  a9b4a06

  Copy full SHA for a9b4a06

Commits on Aug 7, 2025

• feat: Update llama.cpp

  abetlen

  committed
  4f26028

  Copy full SHA for 4f26028

Commits on Jul 1, 2025

• feat: Update llama.cpp

  abetlen

  committed
  0d475d7

  Copy full SHA for 0d475d7

Commits on Jan 29, 2025

• feat: Update llama.cpp

  abetlen

  committed
  80be68a

  Copy full SHA for 80be68a

Commits on Dec 19, 2024

• feat: Update llama.cpp

  abetlen

  committed
  2bc1d97

  Copy full SHA for 2bc1d97

Commits on Oct 31, 2024

• feat: Update llama.cpp

  abetlen

  committed
  e712cff

  Copy full SHA for e712cff

Commits on Sep 19, 2024

• misc: Format

  abetlen

  committed
  9b64bb5

  Copy full SHA for 9b64bb5

• feat: Update sampling API for llama.cpp (#1742)

  Show description for f8fcb3e
  abetlen

  authored
  f8fcb3e

  Copy full SHA for f8fcb3e

Commits on Aug 15, 2024

• feat: Update llama.cpp

  abetlen

  committed
  63d65ac

  Copy full SHA for 63d65ac

Commits on Aug 4, 2024

• fix: llama_grammar_accept_token arg order (#1649)

  Show description for 5575fed
  tc-wolf

  authored
  5575fed

  Copy full SHA for 5575fed

Commits on Jul 18, 2024

• fix: Change repeat_penalty to 1.0 to match llama.cpp defaults (#1590)

  Show description for 0700476
  

  ddh0

  and

  abetlen

  authored
  0700476

  Copy full SHA for 0700476

Commits on Jul 9, 2024

• fix(minor): Minor ruff fixes

  abetlen

  committed
  08f2bb3

  Copy full SHA for 08f2bb3

• fix(misc): Format

  abetlen

  committed
  c1ae815

  Copy full SHA for c1ae815

Commits on Jul 6, 2024

• fix: update token_to_piece

  abetlen

  committed
  157d913

  Copy full SHA for 157d913

Commits on Jun 19, 2024

• fix: Make destructor to automatically call .close() method on Llama class.

  abetlen

  committed
  4c1d74c

  Copy full SHA for 4c1d74c

Commits on Jun 13, 2024

• feat: Add `.close()` method to `Llama` class to explicitly free model from memory (#1513)

  Show description for 320a5d7
  

  jkawamoto

  and

  abetlen

  authored
  320a5d7

  Copy full SHA for 320a5d7

• feat: Support SPM infill (#1492)

  Show description for dbcf64c
  

  CISC

  and

  abetlen

  authored
  dbcf64c

  Copy full SHA for dbcf64c

Commits on Jun 7, 2024

• feat: Update llama.cpp

  abetlen

  committed
  255e1b4

  Copy full SHA for 255e1b4

Commits on Jun 4, 2024

• fix: Avoid duplicate special tokens in chat formats (#1439)

  Show description for 027f7bc
  

  CISC

  and

  abetlen

  authored
  027f7bc

  Copy full SHA for 027f7bc

Commits on Jun 1, 2024

• fix: Use numpy recarray for candidates data, fixes bug with temp < 0

  abetlen

  committed
  af3ed50

  Copy full SHA for af3ed50

Commits on May 4, 2024

• fix: detokenization case where first token does not start with a leading space (#1375)

  Show description for e0d7674
  

  noamgat

  and

  abetlen

  authored
  e0d7674

  Copy full SHA for e0d7674

Commits on Apr 30, 2024

• fix: Suppress all logs when verbose=False, use hardcoded fileno's to work in colab notebooks. Closes #796 Closes #729

  abetlen

  committed
  f116175

  Copy full SHA for f116175

Commits on Apr 26, 2024

• feat: Allow for possibly non-pooled embeddings (#1380)

  Show description for f6ed21f
  

  iamlemec

  and

  abetlen

  authored
  f6ed21f

  Copy full SHA for f6ed21f

Commits on Apr 22, 2024

• feat: Update llama.cpp

  abetlen

  committed
  159cc4e

  Copy full SHA for 159cc4e

Commits on Apr 1, 2024

• fix: last tokens passing to sample_repetition_penalties function (#1295)

  Show description for 62aad61
  

  3 peopleauthored
  62aad61

  Copy full SHA for 62aad61

Commits on Feb 28, 2024

• fix: Remove deprecated cfg sampling functions

  abetlen

  committed
  8c71725

  Copy full SHA for 8c71725

Commits on Feb 26, 2024

• feat: Update llama.cpp

  abetlen

  committed
  cbbcd88

  Copy full SHA for cbbcd88

Commits on Feb 23, 2024

• misc: use typesafe byref for internal classes

  abetlen

  committed
  b9aca61

  Copy full SHA for b9aca61

Commits on Feb 22, 2024

• fix: Raise exceptions when llama model or context fails to load

  abetlen

  committed
  dd22010

  Copy full SHA for dd22010