Alex Cheema cd8efc7fbc Add support for Llama-3.1 (#907) ...

* add dynamicNTK scaling rope

* remove unused var

* fix rope base

* llama3.1 fixes

* TODO for rope eval

* vectorise llama3 base freq calculation

* removed the arbitrary 2.0 rope_scale default case

* fix slow llama3.1 generation by evaluating stateless part of DynamicNTKScalingRoPE in init

* nits + format

* use mx.pi

* fix tests and add test for 3.1

---------

Co-authored-by: Prince Canuma <prince.gdt@gmail.com>
Co-authored-by: Awni Hannun <awni@apple.com>

2024-07-23 13:21:32 -07:00

..

test_datsets.py

Configuration-based use of HF hub-hosted datasets for training (#701)

2024-06-26 10:20:50 -07:00

test_gguf.py

fix(mlx-lm): type hints in gguf.py (#621)

2024-03-26 07:56:01 -07:00

test_lora.py

Add GPT-neox model (#863)

2024-07-11 06:13:17 -07:00

test_models.py

Add support for Llama-3.1 (#907)

2024-07-23 13:21:32 -07:00

test_sample_utils.py

fix(mlx-lm): type hints in gguf.py (#621)

2024-03-26 07:56:01 -07:00

test_server.py

Server loads the model on demand from the request (#851)

2024-06-27 11:37:57 -07:00

test_tuner_utils.py

LoRA: Extract small function (#614)

2024-06-02 06:38:42 -07:00

test_utils.py

LoRA on all linear transformer block layers (#546)

2024-03-12 07:37:40 -07:00