Prince Canuma 2fce02acd8 Add support for Gemma3 (#1336) ... * add support for gemma3 * fix model loading * revert rmsnorm * revert is sliding pattern * revert * add tests * formatting * Update llms/mlx_lm/models/gemma3_text.py Co-authored-by: Awni Hannun <awni.hannun@gmail.com> * Update llms/mlx_lm/models/gemma3_text.py Co-authored-by: Awni Hannun <awni.hannun@gmail.com> * Update llms/mlx_lm/models/gemma3_text.py Co-authored-by: Awni Hannun <awni.hannun@gmail.com> * Update llms/mlx_lm/models/gemma3_text.py Co-authored-by: Awni Hannun <awni.hannun@gmail.com> * Update llms/mlx_lm/models/gemma3_text.py Co-authored-by: Awni Hannun <awni.hannun@gmail.com> * Update llms/mlx_lm/models/gemma3_text.py Co-authored-by: Awni Hannun <awni.hannun@gmail.com> * Update llms/mlx_lm/models/gemma3_text.py Co-authored-by: Awni Hannun <awni.hannun@gmail.com> * fix sliding window mask --------- Co-authored-by: Awni Hannun <awni.hannun@gmail.com> Co-authored-by: Awni Hannun <awni@apple.com>		2025-03-13 08:14:25 -07:00
..
__init__.py	Mlx llm package (#301)	2024-01-12 10:25:56 -08:00
base.py	use a bool mask for attention (#1319)	2025-03-04 12:47:32 -08:00
cache.py	Fix rotating kv cache size (#1093)	2024-11-05 10:24:24 -08:00
cohere2.py	Fix Cohere2: mask shape error (long context) (#1202)	2025-01-12 12:58:08 -08:00
cohere.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
dbrx.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
deepseek_v2.py	fix sharding for more even number of layers (#1276)	2025-02-11 16:26:59 -08:00
deepseek_v3.py	support kimi + more options in chat mode (#1312)	2025-02-28 11:33:18 -08:00
deepseek.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
exaone.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
gemma2.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
gemma3_text.py	Add support for Gemma3 (#1336)	2025-03-13 08:14:25 -07:00
gemma.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
gpt2.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
gpt_bigcode.py	fix gpt bigcode (#1204)	2025-01-13 10:22:32 -08:00
gpt_neox.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
granite.py	Add IBM granite model (#1265)	2025-02-08 15:46:15 -08:00
helium.py	Optimizations for mamba1 (#1213)	2025-02-03 13:36:08 -08:00
hunyuan.py	support hunyuan 7b (#1263)	2025-02-08 15:46:47 -08:00
internlm2.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
internlm3.py	add internlm3 (#1206)	2025-01-15 14:55:41 -08:00
llama.py	remove lm head if unused (#1324)	2025-03-06 15:35:47 -08:00
mamba.py	Optimizations for mamba1 (#1213)	2025-02-03 13:36:08 -08:00
minicpm.py	Optimizations for mamba1 (#1213)	2025-02-03 13:36:08 -08:00
mixtral.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
nemotron.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
olmo2.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
olmo.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
olmoe.py	adding OLMoE architecture (#1321)	2025-03-05 13:46:06 -08:00
openelm.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
phi3.py	Fixes for phi4 mini (#1305)	2025-02-26 16:21:54 -08:00
phi3small.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
phi.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
phimoe.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
phixtral.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
plamo2.py	Fix plamo2 model to use rms_norm (#1308)	2025-03-03 06:12:02 -08:00
plamo.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
qwen2_moe.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
qwen2.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
qwen.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
recurrent_gemma.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
rope_utils.py	Adds EXAONE architecture. (#1145)	2024-12-09 07:58:25 -08:00
stablelm.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
starcoder2.py	Length masking for batch inputs (#1173)	2024-12-18 19:43:52 -08:00
su_rope.py	Fixes for phi4 mini (#1305)	2025-02-26 16:21:54 -08:00
switch_layers.py	Handle longer prompt/generation (#931)	2024-08-16 15:28:39 -07:00