Commits

History for

Commits on Dec 2, 2024

Add support for idefics3 (SmolVLM) (#1059)
xenova
authored
11db949

Commits on Nov 26, 2024

Add new models (Janus, Qwen2-VL, JinaCLIP, LLaVA-OneVision, ViTPose, MGP-STR) & refactor processors. (#1001)
xenova
authored
e848907

Commits on Nov 25, 2024

Add support for `--op_block_list` in quantization script (#1036)

pdufour
and
xenova
authored
5272b12

Commits on Nov 4, 2024

Add support for OLMo models (#1011)
xenova
authored
ed94f6b

Commits on Nov 1, 2024

Remove duplicate `gemma` value from NO_PER_CHANNEL_REDUCE_RANGE_MODELS set (#1005)

bekzod
and
xenova
authored
6bd45ac
Add support for MobileLLM (#1003)
xenova
authored
bf1412e

Commits on Oct 9, 2024

Remove in-library list of supported models
xenova
committed
8b0d330
Add whisper turbo alignment heads
xenova
committed
a32efa3

Commits on Sep 9, 2024

Merge branch 'main' into v3
xenova
committed
4f17e95
Bump `onnxruntime==1.19.2` in scripts/requirements.txt
xenova
committed
ba58ea2

Commits on Sep 2, 2024

Wrap onnxslim with try block
xenova
committed
ba5d725
Add `library_name` option to convert.py
xenova
committed
844099d

Commits on Aug 28, 2024

Fix scripts/requirements.txt
xenova
committed
bf093ae

Commits on Aug 27, 2024

Topologically sort fp16 nodes
xenova
committed
1b7f978

Commits on Aug 9, 2024

Downgrade `onnxconverter-common`
xenova
committed
83f5718
Consolidate conversion and quantization script
xenova
committed
12569b8

Commits on Jul 12, 2024

Separate IO and Quantization args
xenova
committed
974f086
Update converstion script requirements
xenova
committed
9787b75
Reload model for each quantization mode
xenova
committed
9ef3a6d

Commits on Jul 4, 2024

correct attribute
xenova
committed
cd1ea69
Fix q8 quantization for models > 2GB
xenova
committed
04a334a
Remove debug log
xenova
committed
b411e9f
formatting
xenova
committed
23440f0
Reorder functions
xenova
committed
88f3e44
Use `check_and_save_model` for saving fp16 model
xenova
committed
45cd8d4
fp16 disable_shape_infer if model is too large
xenova
committed
c835b54
save q4 quantization
xenova
committed
83cbb21

Commits on Jul 3, 2024

Fix q4 quantization
xenova
committed
9128651
update deps
xenova
committed
04af3d5
Move quantization logic to `quantize.py`
xenova
committed
dd6af93

Commits on Jul 2, 2024

Improve custom whisper ONNX config
xenova
committed
14a2990
Fix custom whisper configs
xenova
committed
0b5469b

Commits on Jul 1, 2024

Add support for decision transformer (#795)
xenova
authored
fc34517

Commits on Jun 22, 2024

Update onnxslim version
xenova
committed
af37e87
add onnxslim intergration (#811)
inisis
authored
e8cfaf6