Add Apertus #39381

EduardDurech
haeggee init swissai model
77810b8a
haeggee AutoModelForCausalLM
bcdaf706
haeggee AutoModelForCausalLM mapping
53a3755c
haeggee qk norm and post ln optional
7c648e73
haeggee fix wrong shape of qk norm: megatron uses head_dim
d9a923da
haeggee automodel fixes
f35ee015
haeggee minor fix in forward
e6921f7e
EduardDurech
dhia680 fix rope validation to accept llama3 scaling
46ca1ae4
EduardDurech `SwissAIForTokenClassification` support
994b1d72
EduardDurech Align `SwissAI` to v4.52.4
8b38b5a4
EduardDurech Align `SwissAI` to v4.53.1
0ffc9b9e
EduardDurech Init CUDA xIELU
7793c878
EduardDurech `SwissAI*`->`Apertus*`
590957b7
EduardDurech ci fix
353c6c0a
EduardDurech EduardDurech force pushed from 1ea63731 to 353c6c0a 1 year ago
EduardDurech check_docstring ignore ApertusConfig
833f5fe3
EduardDurech Licensing and placeholder tests
f0ec65c8
EduardDurech Placeholder doc
1f4e7158
EduardDurech EduardDurech force pushed from 1f20c58a to 1f4e7158 1 year ago
EduardDurech XIELU syntax
cf125820
EduardDurech `_xielu_python` optimization
331fc0d2
EduardDurech Fix xIELU
2728d3cb
chiffa
EduardDurech EduardDurech force pushed from 2728d3cb to b53417c9 1 year ago
ArthurZucker ArthurZucker added New model
ArthurZucker
ArthurZucker commented on 2025-07-15
EduardDurech
ArthurZucker
EduardDurech [tmp] `{beta,eps}` persistent=False
d0d42cdd
EduardDurech Modular `Apertus`
543b3430
EduardDurech CUDA xIELU logging
4d436d09
EduardDurech Merge upstream/main into model/apertus
35d6bb35
EduardDurech EduardDurech force pushed from b53417c9 to 35d6bb35 1 year ago
EduardDurech ci fix
e5ec2316
EduardDurech ci fix
1de44fd5
EduardDurech ci fix
9c0cb617
EduardDurech EduardDurech marked this pull request as ready for review 1 year ago
ArthurZucker
EduardDurech
Cyrilvallez
Cyrilvallez commented on 2025-08-18
Cyrilvallez
Cyrilvallez commented on 2025-08-18
EduardDurech Update license
8f1c0817
EduardDurech Update tests/models/apertus/test_modeling_apertus.py
dad00ca9
EduardDurech `.utils.import_utils.is_torchdynamo_compiling`
cd029ab3
EduardDurech `Apertus` class ordering
250b43af
EduardDurech `past_key_value{->s}`, `make fix-copies`
c4b6d76c
EduardDurech ci fix
98655397
EduardDurech Remove unused configuration parameters
a7abf5eb
EduardDurech `{beta,eps}` saved in checkpoint
273da51d
chiffa
Cyrilvallez
Cyrilvallez commented on 2025-08-27
Cyrilvallez
Cyrilvallez
Cyrilvallez commented on 2025-08-27
EduardDurech `{beta,eps}` Temporarily on CPU
29da453b
EduardDurech Suggestions
792b7de7
EduardDurech EduardDurech force pushed from c66e7b43 to 792b7de7 1 year ago
EduardDurech Merge branch 'main' into model/apertus
c2b3de5f
EduardDurech ci fix
a5889da2
dhia680 remove fx_compatible (deprecated)
e19d5436
dhia680 remove `rotary_embedding_layer`
69c46ed5
dhia680 fully removing `Mask4DTestHard` class
864c4ddc
dhia680 switch to `dtype` instead of `torch_dtype`
e7d03ad1
dhia680 remove unused imports
c3944468
github-actions
dhia680 remove `cache_implementation="static"`
68c6defc
dhia680 +Apertus to `docs/source/en/_toctree.yml` for the doc builder
227f026d
Cyrilvallez
Cyrilvallez approved these changes on 2025-08-28
Cyrilvallez Cyrilvallez merged d10603f7 into main 1 year ago
ArthurZucker
EduardDurech
ArthurZucker
EduardDurech

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone