Kareem 0e2110f7b8 All attention refactor fix (#1491) ... * change initilization of n_heads, n_kv_heads, hidden_size in llama.py * do the same for cohere, mistral, gemma2, granite * do the same for flexattention,cohere, mistral, granite		2025-01-07 02:41:15 -08:00
..
__init__.py	patch_fast_lora	2024-11-20 03:36:41 -08:00
cross_entropy_loss.py	Fix Triton heuristics	2024-12-30 13:52:42 -08:00
fast_lora.py	Update fast_lora.py	2024-11-20 17:07:53 -08:00
flex_attention.py	All attention refactor fix (#1491)	2025-01-07 02:41:15 -08:00
geglu.py	Mistral Nemo 12b (#777)	2024-07-18 10:51:10 -07:00
layernorm.py	Layernorm	2024-11-03 17:05:33 -08:00
rms_layernorm.py	Fix Triton heuristics	2024-12-30 13:52:42 -08:00
rope_embedding.py	Fix Triton heuristics	2024-12-30 13:52:42 -08:00
swiglu.py	Gemma2 (#709)	2024-07-02 22:51:01 -07:00
utils.py	Update utils.py	2024-11-11 00:37:37 -08:00