casinca 9c4be478f8 Optional weight tying for Qwen3 and Llama3.2 pretraining (#949) ...

* optional weight tying for Qwen3 and Llama3.2

* typo

2026-01-14 09:07:04 -06:00

..

__init__.py

Llama 3 KV Cache (#685)

2025-06-21 10:55:20 -05:00

generate.py

Add defensive context trimming for multiturn (#815)

2025-09-09 20:19:00 -05:00

gpt2.py

remove redundant next_cache (#817)

2025-09-11 15:16:08 -05:00

llama3.py

Optional weight tying for Qwen3 and Llama3.2 pretraining (#949)

2026-01-14 09:07:04 -06:00

qwen3.py

Improve MoE implementation (#841)

2025-09-22 15:21:06 -05:00

utils.py

Improve KV cache code for torch.compile (#705)

2025-06-23 18:08:49 -05:00