8 Releases published by 1 person

• v4.2.1

  published Jan 6, 2025

• v4.2.2

  published Jan 7, 2025

• v4.3.0

  published Jan 14, 2025

• v4.3.1

  published Jan 15, 2025

• v4.4.0

  published Jan 20, 2025

• v4.4.1

  published Jan 27, 2025

• v4.4.2

  published Jan 29, 2025

• v4.4.3

  published Jan 30, 2025

19 Pull requests merged by 3 people

• fix: bad chat template diffing

  #98 merged Jan 30, 2025

• fix: chat template repetition bug

  #95 merged Jan 29, 2025

• fix: invoke min_p when min_p is selected

  #92 merged Jan 29, 2025

• feat: Showcase section in readme

  #91 merged Jan 29, 2025

• Demo video for README

  #90 merged Jan 29, 2025

• Community standards

  #86 merged Jan 29, 2025

• Draft: Address NoKvCacheSlot

  #83 merged Jan 24, 2025

• More README improvements

  #85 merged Jan 21, 2025

• Sampler variants

  #64 merged Jan 20, 2025

• fix: tell people about the ignore asset root

  #84 merged Jan 16, 2025

• Multiple contexts, one model

  #82 merged Jan 15, 2025

• faster ci

  #80 merged Jan 15, 2025

• Add in-editor documentation

  #76 merged Jan 9, 2025

• Include debug file in windows builds

  #77 merged Jan 9, 2025

• Implement context shifting condition from upstream llama.cpp

  #79 merged Jan 9, 2025

• Automatic integration test

  #71 merged Jan 9, 2025

• feat: update readme, add funding link

  #78 merged Jan 8, 2025

• Fix mistral template

  #73 merged Jan 7, 2025

• Replace wgpu dependency with GGML API calls

  #70 merged Jan 6, 2025

1 Pull request opened by 1 person

• Draft: Run test in docker container, avoid runner env inconsistencies

  #89 opened Jan 27, 2025

10 Issues closed by 2 people

• Crash on Arch Linux

  #88 closed Jan 28, 2025

• Running out of context bug: "Decode Error 1: NoKvCacheSlot"

  #74 closed Jan 24, 2025

• Inferencing the same model from two contexts at the same time crashes Godot.

  #81 closed Jan 15, 2025

• Hard crash in `ggml_backend_dev_count`

  #75 closed Jan 9, 2025

• Include windows debug files in distributable

  #65 closed Jan 9, 2025

• Add docstrings to Nobody nodes.

  #26 closed Jan 9, 2025

• Integration test

  #50 closed Jan 9, 2025

• Mistral chat template causes crash

  #72 closed Jan 8, 2025

• Crashes after run()

  #43 closed Jan 7, 2025

• Hung with Rust function panic

  #69 closed Jan 6, 2025

5 Issues opened by 3 people

• [BUG] It seems some tokens crash the model worker

  #99 opened Jan 30, 2025

• [FEATURE] Please document Samplers, the DB, and Context

  #97 opened Jan 30, 2025

• [BUG] Llama2 and some Mistral based ggufs seem to not work.

  #96 opened Jan 30, 2025

• Support Deepseek R1 models' `<think>` token stuff

  #94 opened Jan 29, 2025

• Always valid JSON output

  #87 opened Jan 26, 2025

5 Unresolved conversations

Sometimes conversations happen on old items that aren’t yet closed. Here is a list of all the Issues and Pull Requests with unresolved conversations.

• Allow implementation in physics_process

  #21 commented on Jan 8, 2025 • 0 new comments

• Control vectors for each chat completion

  #40 commented on Jan 8, 2025 • 0 new comments

• Explicit actor modelling proposal

  #57 commented on Jan 8, 2025 • 0 new comments

• Can't export for Android

  #66 commented on Jan 8, 2025 • 0 new comments

• Draft: Compile for android

  #67 commented on Jan 8, 2025 • 0 new comments