KEMBAR78
Comparing v0.2.6...v0.2.7 · turboderp-org/exllamav2 · GitHub
Skip to content
Permalink

Comparing changes

Choose two branches to see what’s changed or to start a new pull request. If you need to, you can also or learn more about diff comparisons.

Open a pull request

Create a new pull request by comparing changes across two branches. If you need to, you can also . Learn more about diff comparisons here.
base repository: turboderp-org/exllamav2
Failed to load repositories. Confirm that selected base ref is valid, then try again.
Loading
base: v0.2.6
Choose a base ref
...
head repository: turboderp-org/exllamav2
Failed to load repositories. Confirm that selected head ref is valid, then try again.
Loading
compare: v0.2.7
Choose a head ref
  • 20 commits
  • 46 files changed
  • 3 contributors

Commits on Dec 8, 2024

  1. Prevent UnboundLocalError when loading with yarn/su with short ctx len ( 

    #694)

* scaling_factor is left unbound when the requested max_seq_len < the model's original unscaled max_seq_len
    @DocShotgun
    DocShotgun authored Dec 8, 2024
    Configuration menu
    Copy the full SHA
    af69ce9 View commit details
    Browse the repository at this point in the history

Commits on Dec 9, 2024

  1. Fix background filter eval when draft model used

    @turboderp
    turboderp committed Dec 9, 2024
    Configuration menu
    Copy the full SHA
    f4119ae View commit details
    Browse the repository at this point in the history

  2. Cleanup build actions

    @turboderp
    turboderp committed Dec 9, 2024
    Configuration menu
    Copy the full SHA
    8bb283d View commit details
    Browse the repository at this point in the history

  3. Merge remote-tracking branch 'origin/dev' into dev

    @turboderp
    turboderp committed Dec 9, 2024
    Configuration menu
    Copy the full SHA
    254e76b View commit details
    Browse the repository at this point in the history

  4. Util function to get byte size of MM embeddings object

    @turboderp
    turboderp committed Dec 9, 2024
    Configuration menu
    Copy the full SHA
    edf1a35 View commit details
    Browse the repository at this point in the history

Commits on Dec 15, 2024

  1. Refactoring

    @turboderp
    turboderp committed Dec 15, 2024
    Configuration menu
    Copy the full SHA
    9934f06 View commit details
    Browse the repository at this point in the history

  2. Fix ChatML template in multimodal example

    @turboderp
    turboderp committed Dec 15, 2024
    Configuration menu
    Copy the full SHA
    c78d902 View commit details
    Browse the repository at this point in the history

  3. Qwen2-VL: Basic video support

    @turboderp
    turboderp committed Dec 15, 2024
    Configuration menu
    Copy the full SHA
    4061c24 View commit details
    Browse the repository at this point in the history

Commits on Dec 18, 2024

  1. Read number of vision tower layers from config for Pixtral (fix Pixtr… 

    …al-Large)
    @turboderp
    turboderp committed Dec 18, 2024
    Configuration menu
    Copy the full SHA
    f76bc85 View commit details
    Browse the repository at this point in the history

  2. Fix chat example system prompt

    @turboderp
    turboderp committed Dec 18, 2024
    Configuration menu
    Copy the full SHA
    cf7fcd1 View commit details
    Browse the repository at this point in the history

Commits on Dec 21, 2024

  1. Modify handling for Pixtral Large model params (#701) 

    * Modify handling for Pixtral Large model params.

* Fix multimodal_projector_bias to default to True if not in model config.json
    @nintwentydo
    nintwentydo authored Dec 21, 2024
    Configuration menu
    Copy the full SHA
    b2dd5a7 View commit details
    Browse the repository at this point in the history

Commits on Dec 25, 2024

  1. Fix RoPE alpha after refactor in #4d25874

    @turboderp
    turboderp committed Dec 25, 2024
    Configuration menu
    Copy the full SHA
    d815f5f View commit details
    Browse the repository at this point in the history

  2. Support Cohere2 architecture

    @turboderp
    turboderp committed Dec 25, 2024
    Configuration menu
    Copy the full SHA
    ebfefc4 View commit details
    Browse the repository at this point in the history

Commits on Dec 27, 2024

  1. Chat example Granite3 template

    @turboderp
    turboderp committed Dec 27, 2024
    Configuration menu
    Copy the full SHA
    ab4d9e1 View commit details
    Browse the repository at this point in the history

  2. Support Granite 3.x arch

    @turboderp
    turboderp committed Dec 27, 2024
    Configuration menu
    Copy the full SHA
    bee449d View commit details
    Browse the repository at this point in the history

  3. Fix Granite3 logit scaling

    @turboderp
    turboderp committed Dec 27, 2024
    Configuration menu
    Copy the full SHA
    82bb648 View commit details
    Browse the repository at this point in the history

Commits on Dec 29, 2024

  1. Don't compile AVX2 functions when building without AVX2 support

    @turboderp
    turboderp committed Dec 29, 2024
    Configuration menu
    Copy the full SHA
    fb5000a View commit details
    Browse the repository at this point in the history

  2. Fix compilation errors on aarch64

    @turboderp
    turboderp committed Dec 29, 2024
    Configuration menu
    Copy the full SHA
    b010cb9 View commit details
    Browse the repository at this point in the history

Commits on Dec 30, 2024

  1. Bump to v0.2.7

    @turboderp
    turboderp committed Dec 30, 2024
    Configuration menu
    Copy the full SHA
    1ef6183 View commit details
    Browse the repository at this point in the history

  2. Fix video example

    @turboderp
    turboderp committed Dec 30, 2024
    Configuration menu
    Copy the full SHA
    ae241a9 View commit details
    Browse the repository at this point in the history
Loading