Commits

Branch selector

main

User selector

All users

All time

Commit history

Commits on Mar 15, 2026

• feat(core): add verbose debug logging to longest_token_prefix fast paths

  Show description for 90d46dc
  JamePeng

  committed
  90d46dcView commit details

  Copy full SHA for 90d46dc

  Browse repository at this point

• Update MIT license copyright to collective authorship (2023-2026)

  Show description for 35957bd
  JamePeng

  committed
  35957bdView commit details

  Copy full SHA for 35957bd

  Browse repository at this point

• Update CMakeLists.txt

  JamePeng

  committed
  9c29c27View commit details

  Copy full SHA for 9c29c27

  Browse repository at this point

• Bump version to 0.3.33

  Show description for 6bbc8d2
  JamePeng

  committed
  6bbc8d2View commit details

  Copy full SHA for 6bbc8d2

  Browse repository at this point

Commits on Mar 14, 2026

• Update Submodule vendor/llama.cpp 463b6a9..d23355a

  JamePeng

  committed
  b0f00a9View commit details

  Copy full SHA for b0f00a9

  Browse repository at this point

• perf(mtmd): optimize media_id masking with bitwise AND

  Show description for bd0b069
  JamePeng

  committed
  bd0b069View commit details

  Copy full SHA for bd0b069

  Browse repository at this point

• fix(mtmd): prevent multimodal image freeze by injecting deterministic media IDs

  Show description for dd111eb
  JamePeng

  committed
  dd111ebView commit details

  Copy full SHA for dd111eb

  Browse repository at this point

• Update Submodule vendor/llama.cpp 8f974d2..463b6a9

  JamePeng

  committed
  ec2cbeaView commit details

  Copy full SHA for ec2cbea

  Browse repository at this point

Commits on Mar 13, 2026

• Update mtmd_cpp API 20260313

  JamePeng

  committed
  afd06eaView commit details

  Copy full SHA for afd06ea

  Browse repository at this point

• Update Submodule vendor/llama.cpp 557fe2d..8f974d2

  JamePeng

  committed
  a6db3f2View commit details

  Copy full SHA for a6db3f2

  Browse repository at this point

• fix(Llama.generate): add explicit fallback context reset and expand `Llama.generate` docstrings

  Show description for 6f81466
  JamePeng

  committed
  6f81466View commit details

  Copy full SHA for 6f81466

  Browse repository at this point

Commits on Mar 12, 2026

• docs(README.md): add sampling parameter guide and strategic project tips to README

  Show description for d422e82
  JamePeng

  committed
  d422e82View commit details

  Copy full SHA for d422e82

  Browse repository at this point

• fix(sampling): prevent memory drift and hallucinations in logits view

  Show description for aa38850
  JamePeng

  committed
  aa38850View commit details

  Copy full SHA for aa38850

  Browse repository at this point

• Update Submodule vendor/llama.cpp fdb1764..557fe2d

  JamePeng

  committed
  380c643View commit details

  Copy full SHA for 380c643

  Browse repository at this point

Commits on Mar 11, 2026

• feat(_ggml): extend ctypes bindings with more ggml constants, enums, and structs

  JamePeng

  committed
  2bb5cd9View commit details

  Copy full SHA for 2bb5cd9

  Browse repository at this point

• Update llama.cpp API 20260312

  JamePeng

  committed
  c420a2aView commit details

  Copy full SHA for c420a2a

  Browse repository at this point

• Update Submodule vendor/llama.cpp 182acfe..fdb1764

  JamePeng

  committed
  21c4c87View commit details

  Copy full SHA for 21c4c87

  Browse repository at this point

• fix(chat_handler): fix tools and function calling in MTMDChatHandler.

  JamePeng

  committed
  8641667View commit details

  Copy full SHA for 8641667

  Browse repository at this point

• Update Submodule vendor/llama.cpp 10e5b14..182acfe

  JamePeng

  committed
  b827932View commit details

  Copy full SHA for b827932

  Browse repository at this point

Commits on Mar 10, 2026

• perf(cache): optimize LlamaDiskCache I/O and fix LRU behavior

  Show description for a36e8d2
  JamePeng

  committed
  a36e8d2View commit details

  Copy full SHA for a36e8d2

  Browse repository at this point

• Sync examples : fix empty items in json_schema_to_grammar.py (#19968)

  JamePeng

  committed
  8e8f36fView commit details

  Copy full SHA for 8e8f36f

  Browse repository at this point

• Update Submodule vendor/llama.cpp ec947d2..10e5b14

  JamePeng

  committed
  054f04eView commit details

  Copy full SHA for 054f04e

  Browse repository at this point

• perf(cache): upgrade `LlamaRAMCache` to O(1) eviction and set `LlamaTrieCache` as default

  Show description for 5c8e056
  JamePeng

  committed
  5c8e056View commit details

  Copy full SHA for 5c8e056

  Browse repository at this point

• Update Submodule vendor/llama.cpp 59db9a3..ec947d2

  JamePeng

  committed
  4bec165View commit details

  Copy full SHA for 4bec165

  Browse repository at this point

Commits on Mar 9, 2026

• fix(core): disable swa_full for non-SWA models (sync llama.cpp upstream #20291)

  Show description for 9acc507
  JamePeng

  committed
  9acc507View commit details

  Copy full SHA for 9acc507

  Browse repository at this point

• Update llama.cpp API 20260310

  JamePeng

  committed
  955ac33View commit details

  Copy full SHA for 955ac33

  Browse repository at this point

• fix(chat_format): fix namespace and variable shadowing of llama modules

  Show description for 7efcb2b
  JamePeng

  committed
  7efcb2bView commit details

  Copy full SHA for 7efcb2b

  Browse repository at this point

• fix(cache): fix namespace shadowing to prevent AttributeError

  Show description for 5e285fe
  JamePeng

  committed
  5e285feView commit details

  Copy full SHA for 5e285fe

  Browse repository at this point

• Update Submodule vendor/llama.cpp 35bee03..59db9a3

  JamePeng

  committed
  9cc2efbView commit details

  Copy full SHA for 9cc2efb

  Browse repository at this point

Commits on Mar 8, 2026

• feat(MTMDChatHandler): support audio inputs and fix interleaved media ordering

  Show description for 964160d
  JamePeng

  committed
  964160dView commit details

  Copy full SHA for 964160d

  Browse repository at this point

• Update Submodule vendor/llama.cpp b283f6d..35bee03

  JamePeng

  committed
  5bd0870View commit details

  Copy full SHA for 5bd0870

  Browse repository at this point

• Bump version to 0.3.32

  JamePeng

  committed
  e7e1d48View commit details

  Copy full SHA for e7e1d48

  Browse repository at this point

• Update Submodule vendor/llama.cpp c5a7788..b283f6d

  JamePeng

  committed
  3e41921View commit details

  Copy full SHA for 3e41921

  Browse repository at this point

• docs(issue-template): modernize bug report for efficiency

  Show description for 8ec1763
  JamePeng

  committed
  8ec1763View commit details

  Copy full SHA for 8ec1763

  Browse repository at this point

• fix(sampling): pass seed to sampling context and remove global mutation

  Show description for ad10cfd
  JamePeng

  committed
  ad10cfdView commit details

  Copy full SHA for ad10cfd

  Browse repository at this point