Code

Optimize tp broadcast#2889

Code

Merged

lvhan028 merged 62 commits intoInternLM:mainInternLM/lmdeploy:mainfrom grimoire:optimize-tp-broadcastgrimoire/lmdeploy:optimize-tp-broadcastCopy head branch name to clipboardDec 17, 2024

+149-17

Commits

Commits on Nov 18, 2024

• refactor VL modules for internvl and qwen2-vl (InternLM#2764)

  Show description for 464d451
  lvhan028

  authored
  464d451

Commits on Nov 19, 2024

• Refactor VL modules for glm4v, deepseek-vl, llava-hf, cogvlm (InternLM#2772)

  Show description for 3ba8309
  lvhan028

  authored
  3ba8309

Commits on Nov 20, 2024

• Refactor VL modules for qwen-vl, llava and llava_next (InternLM#2773)

  Show description for accfc00
  lvhan028

  authored
  accfc00

• Refactor VL modules for qwen2-vl (InternLM#2777)

  Show description for 463b508
  lvhan028

  authored
  463b508

• Fix side-effect to internvl (InternLM#2778)

  Show description for 9e18529
  lvhan028

  authored
  9e18529

• Refactor VL modules for phi3-vision (InternLM#2779)

  Show description for b2c29be
  lvhan028

  authored
  b2c29be

• Refactor VL modules for mllama and yi-vl (InternLM#2781)

  Show description for aca0f2d
  lvhan028

  authored
  aca0f2d

Commits on Nov 22, 2024

• Refactor VLM module for minicpm and molmo (InternLM#2794)

  lvhan028

  authored
  fff7b91

• Refactor VLM modules for xcomposer series (InternLM#2796)

  lvhan028

  authored
  45cf22d

• Refactor VLM modules for internvl-llava (InternLM#2797)

  lvhan028

  authored
  38eec0d

Commits on Nov 25, 2024

• Refactor VLM modules v2 (InternLM#2806)

  Show description for c4638d5
  lvhan028

  authored
  c4638d5

• Remove vl template (InternLM#2809)

  lvhan028

  authored
  b07211b

• Resolve conflicts (InternLM#2811)

  Show description for e7b5a7e
  

  8 peopleauthored
  e7b5a7e

• resolve conflicts

  lvhan028

  committed
  70875eb

• Merge pull request InternLM#2812 from lvhan028/resolve-conflicts

  Show description for 787f765
  lvhan028

  authored
  787f765

• PytorchEngine refactor multimodal (InternLM#2742)

  Show description for 099721a
  grimoire

  authored
  099721a

• minor-fix

  lvhan028

  committed
  36a15e3

• minor-fix (InternLM#2813)

  lvhan028

  authored
  4e6760e

Commits on Dec 3, 2024

• fix

  grimoire

  committed
  f7c167e

Commits on Dec 4, 2024

• fix

  lvhan028

  committed
  e977361

• fix mono

  grimoire

  committed
  5dc967d

• fix docs

  grimoire

  committed
  ae7015a

• read norm_type

  grimoire

  committed
  c746fd3

• Merge branch 'main' into refactor-vl

  lvhan028

  committed
  dc9757c

• super().collect_images->self.collect_images

  lvhan028

  committed
  d577acb

• Merge branch 'refactor-vl' of github.com:InternLM/lmdeploy into refactor-vl

  lvhan028

  committed
  30ea075

Commits on Dec 5, 2024

• add note in supported models

  lvhan028

  committed
  bbcf9a5

• define the parameters clearly

  lvhan028

  committed
  b3a2887

• better streaming

  grimoire

  committed
  8f7a56f

Commits on Dec 6, 2024

• merge main

  grimoire

  committed
  c2a5b44

• fix molmo

  lvhan028

  committed
  29c3558

• Merge branch 'refactor-vl' of github.com:InternLM/lmdeploy into refactor-vl

  grimoire

  committed
  240f352

Commits on Dec 9, 2024

• Fix vision model batch inference (InternLM#2868)

  Show description for 74e7bf8
  lvhan028

  authored
  74e7bf8

• warn glm4v does not support multi images

  lvhan028

  committed
  fb3f8cc

• unconst

  grimoire

  committed
  a65007b

• Merge branch 'refactor-vl' of github.com:InternLM/lmdeploy into refactor-vl

  grimoire

  committed
  88f99d4

• fix deepseek-vl

  lvhan028

  committed
  ed2efb3

• merge main

  lvhan028

  committed
  d8828cf

• fix internvl

  lvhan028

  committed
  18b38e9

• Merge branch 'refactor-vl' into fix-refactor-vl

  lvhan028

  committed
  3cb7e4d

Commits on Dec 10, 2024

• fix llava

  grimoire

  committed
  92b09d0

• fix minicpm 2.6

  lvhan028

  committed
  db367f4

• fix callback

  grimoire

  committed
  4e2f1f8

• fix minicpm v2.5

  lvhan028

  committed
  715fbb3

• fix minicpm v2.6

  lvhan028

  committed
  1a6d88f

• Merge branch 'refactor-vl' into fix-refactor-vl

  lvhan028

  committed
  8ee759f

• update llava_next.py

  lvhan028

  committed
  f0a4422

• remove hardcode from xcomposer2.py

  lvhan028

  committed
  ee022ad

• Merge pull request InternLM#2879 from lvhan028/fix-refactor-vl

  Show description for 02a25eb
  lvhan028

  authored
  02a25eb

• rollback supported_models

  lvhan028

  committed
  a21abe3

• change to staticmethod

  lvhan028

  committed
  6a9342e

Commits on Dec 12, 2024

• optimize tp

  grimoire

  committed
  798298b

• solve conflict

  grimoire

  committed
  1b6ea24

• Merge branch 'refactor-vl' into optimize-tp-broadcast

  grimoire

  committed
  a9aacda

• Merge branch 'refactor-vl' of github.com:InternLM/lmdeploy into refactor-vl

  grimoire

  committed
  d005bc8

• Merge branch 'refactor-vl' into optimize-tp-broadcast

  grimoire

  committed
  e9517e1

• fix vlm quantization

  lvhan028

  committed
  4107d4f

• Merge branch 'refactor-vl' of github.com:InternLM/lmdeploy into refactor-vl

  lvhan028

  committed
  fdaa601

Commits on Dec 13, 2024

• update doc

  lvhan028

  committed
  e5a5085

• update

  lvhan028

  committed
  bc93e73

• Merge branch 'refactor-vl' into optimize-tp-broadcast

  grimoire

  committed
  67188eb

• solve conflict

  grimoire

  committed
  6cdc76a