Commits

Branch selector

 v0.18.dev0

User selector

All users

Datepicker

All time

Commit History

Commits on Sep 9, 2024

• Added hermes 3 support (#2886)

  Show description for 77b20e1
  AMKCode

  authored
  77b20e1

  Copy full SHA for 77b20e1

Commits on Sep 7, 2024

• [Serving] Tree drafting (#2879)

  Show description for 20e7c4c
  vinx13

  authored
  20e7c4c

  Copy full SHA for 20e7c4c

Commits on Sep 4, 2024

• [Bench] Truncate to model max length when tokenizing (#2881)

  Show description for de6e916
  MasterJH5574

  authored
  de6e916

  Copy full SHA for de6e916

Commits on Sep 2, 2024

• [Engine] Preparation for switching between spec-decode mode and normal mode (#2867)

  Show description for cbaa2d0
  MasterJH5574

  authored
  cbaa2d0

  Copy full SHA for cbaa2d0

• [Fix] Fix RWKV v6 weights loading for 7B/14B models (#2874)

  Show description for 91d861d
  MollySophia

  authored
  91d861d

  Copy full SHA for 91d861d

Commits on Sep 1, 2024

• [Conv] Fix Qwen2 conv template (#2872)

  Show description for 55c2d9a
  MasterJH5574

  authored
  55c2d9a

  Copy full SHA for 55c2d9a

Commits on Aug 29, 2024

• Auto updated submodule references

  

  Git bot

  committed
  23094e7

  Copy full SHA for 23094e7

Commits on Aug 28, 2024

• [Fix] Attribute mismatch in Phi-3-Vision Model (#2862)

  Show description for 52b37e8
  mengshyu

  authored
  52b37e8

  Copy full SHA for 52b37e8

• [Fix] Update seq len info after prefix cache operation (#2860)

  Show description for ab096ce
  cyx-6

  authored
  ab096ce

  Copy full SHA for ab096ce

• [Python] Reuse KVCache classes from TVM (#2857)

  Show description for 5a1bd3c
  MasterJH5574

  authored
  5a1bd3c

  Copy full SHA for 5a1bd3c

Commits on Aug 27, 2024

• [Fix] Improve prefill policy for prefix cache reuse (#2859)

  Show description for 66730dc
  cyx-6

  authored
  66730dc

  Copy full SHA for 66730dc

Commits on Aug 26, 2024

• [Model] Use attention kernel implemented by tir for CLIP vision encoder (#2827)

  Show description for ae0853e
  

  3 peopleauthored
  ae0853e

  Copy full SHA for ae0853e

• [MultiGPU] Fix pipeline parallelism prompt message (#2855)

  Show description for 7264faa
  MasterJH5574

  authored
  7264faa

  Copy full SHA for 7264faa

Commits on Aug 24, 2024

• [Preset] Add Phi-3.5-mini to preset (#2845)

  Show description for 83d0fe3
  CharlieFRuan

  authored
  83d0fe3

  Copy full SHA for 83d0fe3

• [Model] Remove hidden size assertion in LLaMA model (#2848)

  Show description for 9336aab
  YiyanZhai

  authored
  9336aab

  Copy full SHA for 9336aab

• [Docs] Update installation instructions for ROCm (#2850)

  Show description for 42cf2ad
  MasterJH5574

  authored
  42cf2ad

  Copy full SHA for 42cf2ad

Commits on Aug 23, 2024

• [ROCm] hipBLAS integration (#2830)

  Show description for 3869fa1
  

  MasterJH5574

  and

  LeshengJin

  authored
  3869fa1

  Copy full SHA for 3869fa1

• Auto updated submodule references

  

  Git bot

  committed
  2cbf393

  Copy full SHA for 2cbf393

• [Android] Update Andorid APK (#2842)

  Show description for 935be14
  mengshyu

  authored
  935be14

  Copy full SHA for 935be14

Commits on Aug 22, 2024

• [SLM] MiniCPM Multi-GPU support (#2826)

  Show description for 9f87508
  tlopex

  authored
  9f87508

  Copy full SHA for 9f87508

• [Docs] Clarifying the number of additional model (#2841)

  Show description for c5dbae2
  MasterJH5574

  authored
  c5dbae2

  Copy full SHA for c5dbae2

• [iOS] Add Phi3.5 into `mlc-package-config.json` (#2840)

  Show description for fd06a42
  MasterJH5574

  authored
  fd06a42

  Copy full SHA for fd06a42

• [Model] Support Phi-3.5 (#2839)

  Show description for 25e3bad
  MasterJH5574

  authored
  25e3bad

  Copy full SHA for 25e3bad

• Update the length of decode token in eagle mode (#2834)

  Show description for 761583e
  sunzj

  authored
  761583e

  Copy full SHA for 761583e

Commits on Aug 20, 2024

• [Op] Improve the RoPE embedding implementation (#2816)

  Show description for 73bfbe6
  MasterJH5574

  authored
  73bfbe6

  Copy full SHA for 73bfbe6

• [Sampler] Fix CPU sampler for speculative decoding (#2824)

  Show description for 6a5f63b
  MasterJH5574

  authored
  6a5f63b

  Copy full SHA for 6a5f63b

Commits on Aug 19, 2024

• [MultiGPU] Enable pre-sharding for pipeline parallelism (#2825)

  Show description for 78494f6
  MasterJH5574

  authored
  78494f6

  Copy full SHA for 78494f6

• [MultiGPU] Fix memory estimation for pipeline parallelism (#2823)

  Show description for ee97167
  MasterJH5574

  authored
  ee97167

  Copy full SHA for ee97167

Commits on Aug 18, 2024

• [Model] Support MoE structure of MiniCPM model. (#2815)

  Show description for 1ffa0bf
  tlopex

  authored
  1ffa0bf

  Copy full SHA for 1ffa0bf

Commits on Aug 16, 2024

• [Pipeline] Switch VMBuiltinLower to LowerRuntimeBuiltin (#2817)

  Show description for 0c0c7a6
  MasterJH5574

  authored
  0c0c7a6

  Copy full SHA for 0c0c7a6

• [Android] render markdown text as default (#2814)

  Show description for 60f2b4c
  huanglizhuo

  authored
  60f2b4c

  Copy full SHA for 60f2b4c

Commits on Aug 15, 2024

• [Serving] Requiring explicit passing draft tokens to logit processor (#2810)

  Show description for 2d63553
  vinx13

  authored
  2d63553

  Copy full SHA for 2d63553

• Save the packed_weight as storage data type (#2804)

  Show description for c4fae96
  sunzj

  authored
  c4fae96

  Copy full SHA for c4fae96

Commits on Aug 14, 2024

• [iOS] Use swift-markdown-ui to display markdown text in an iOS app (#2809)

  Show description for 77eee0c
  huanglizhuo

  authored
  77eee0c

  Copy full SHA for 77eee0c

• [Model] Add tie_word_embedding option for Minicpm model (#2808)

  Show description for 38ee556
  tlopex

  authored
  38ee556

  Copy full SHA for 38ee556

Pagination

Previous

Next