transformers/docs/source
Zach Mueller e0dfd7bcaf
Speedup model init on CPU (by 10x+ for llama-3-8B as one example) (#31771)
* 1,100%!

* Clean

* Don't touch DS

* Experiment with dtype allocation

* skip test_load_save_without_tied_weights test

* A little faster

* Include proper upscaling?

* Fixup tests

* Potentially skip?

* Let's see if this fixes git history

* Maintain new dtype

* Fin

* Rm hook idea for now

* New approach, see what breaks

* stage

* Clean

* Stash

* Should be fin now, just need to mark failing models

* Clean up

* Simplify

* Deal with weird models

* Enc/Dec

* Skip w/ reason

* Adjust test

* Fix test

* one more test

* Keep experimenting

* Fix ref

* TO REMOVE: testing feedback CI

* Right push

* Update tests/utils/test_modeling_utils.py

Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>

* disable

* Add new func

* Test nits from Amy

* Update src/transformers/modeling_utils.py

Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>

* Adjust comment

* Adjust comment on skip

* make private

* Fin

* Should be a not flag

* Clarify and rename test

---------

Co-authored-by: Marc Sun <marc@huggingface.co>
Co-authored-by: Arthur <48595927+ArthurZucker@users.noreply.github.com>
Co-authored-by: amyeroberts <22614925+amyeroberts@users.noreply.github.com>
2024-07-16 09:32:01 -04:00
..
de Skip tests properly (#31308) 2024-06-26 21:59:08 +01:00
en Speedup model init on CPU (by 10x+ for llama-3-8B as one example) (#31771) 2024-07-16 09:32:01 -04:00
es [docs] Spanish translation of tokenizer_summary.md (#31154) 2024-06-03 16:52:23 -07:00
fr Add French version of run scripts tutorial (#31483) 2024-06-28 18:02:30 +02:00
hi More fixes for doctest (#30265) 2024-04-16 11:58:55 +02:00
it Docs / Quantization: Replace all occurences of load_in_8bit with bnb config (#31136) 2024-05-30 16:47:35 +02:00
ja Update CometCallback to allow reusing of the running experiment (#31366) 2024-07-05 08:13:46 +02:00
ko Skip tests properly (#31308) 2024-06-26 21:59:08 +01:00
ms Remove old TF port docs (#30426) 2024-04-23 16:06:20 +01:00
pt Use HF_HUB_OFFLINE + fix has_file in offline mode (#31016) 2024-05-29 11:55:43 +01:00
te docs: fix broken link (#31370) 2024-06-12 11:33:00 +01:00
tr Translate index.md to Turkish (#27093) 2023-11-08 08:35:20 -05:00
zh Update CometCallback to allow reusing of the running experiment (#31366) 2024-07-05 08:13:46 +02:00
_config.py [#29174] ImportError Fix: Trainer with PyTorch requires accelerate>=0.20.1 Fix (#29888) 2024-04-08 14:21:16 +01:00