yangjianfengo1
93fcf7e4ec
【New Feature】W4afp8 supports per group quantization (#4272)
* w4afp8 支持per group
* code style
* 精度完成
* revert append attn utils
* ffn1 动态量化
* ffn2 支持动态量化
* code style
* code style
* 修改单测
* 修改单测
* fix bug
* Implement conditional parameter creation for layers
Add parameter creation for up_gate_proj_in_scale when ep_size > 1.
* code style
* fix conflict
* code style
* code style
* 修复w4aint8 精度
* fix ci
---------
Co-authored-by: yuanxiaolan <yuanxiaolan01@baidu.com>
2025-11-05 21:00:23 +08:00
..
2025-11-01 19:13:50 +08:00
2025-07-19 23:19:27 +08:00
2025-09-29 15:56:54 +08:00
2025-09-01 17:50:17 +08:00
2025-09-01 17:50:17 +08:00
2025-10-21 22:06:06 +08:00
2025-07-19 23:19:27 +08:00
2025-09-01 17:50:17 +08:00
2025-10-23 17:57:59 +08:00
2025-09-11 10:46:09 +08:00
2025-10-21 22:06:06 +08:00
2025-11-05 21:00:23 +08:00
2025-09-24 16:39:51 +08:00
2025-08-14 22:40:44 +08:00
2025-11-03 10:08:01 +08:00
2025-11-05 21:00:23 +08:00
2025-08-19 02:54:47 -07:00
2025-10-20 14:44:58 +08:00
2025-11-05 21:00:23 +08:00
2025-09-11 10:46:09 +08:00
2025-09-03 10:54:34 +08:00
2025-07-07 20:06:28 +08:00
2025-07-22 15:03:41 +08:00
2025-10-21 22:06:06 +08:00
2025-10-21 22:06:06 +08:00
2025-08-14 22:40:44 +08:00
2025-09-19 14:07:31 +08:00
2025-11-05 10:47:40 +08:00
2025-07-22 15:03:41 +08:00
2025-10-21 14:25:45 +08:00
2025-10-24 10:14:53 +08:00
2025-09-30 17:54:19 +08:00
2025-10-31 22:00:31 +08:00
2025-10-31 22:00:31 +08:00
2025-08-01 18:03:36 +08:00
2025-07-19 23:19:27 +08:00
2025-10-30 10:28:36 +08:00
2025-09-23 18:41:33 +08:00
2025-10-31 21:25:11 +08:00
2025-10-30 10:28:26 +08:00
2025-07-19 23:19:27 +08:00
2025-07-19 23:19:27 +08:00
2025-10-20 15:18:12 +08:00
2025-09-04 17:39:59 +08:00
2025-07-19 23:19:27 +08:00
2025-07-19 23:19:27 +08:00
2025-09-01 17:50:17 +08:00
2025-09-01 17:50:17 +08:00
2025-08-29 11:04:04 +08:00
2025-07-09 18:56:27 +08:00
2025-10-20 21:09:13 +08:00
2025-07-19 23:19:27 +08:00
2025-09-01 17:50:17 +08:00
2025-07-19 23:19:27 +08:00
2025-07-07 16:53:14 +08:00
2025-10-21 22:06:06 +08:00
2025-07-19 23:19:27 +08:00
2025-10-28 18:31:23 +08:00
2025-09-10 13:11:57 +08:00
2025-07-30 09:31:29 +08:00
2025-09-01 17:50:17 +08:00
2025-07-19 23:19:27 +08:00
2025-09-28 19:42:53 +08:00
2025-11-03 10:08:01 +08:00
2025-10-27 10:50:08 +08:00
2025-10-27 10:50:08 +08:00
2025-10-27 10:50:08 +08:00
2025-09-01 17:50:17 +08:00