Skip to content

Commit fea8d82

Browse files
authored
fix llama13b bug (#10525)
1 parent c7d44f2 commit fea8d82

File tree

1 file changed

+2
-2
lines changed

1 file changed

+2
-2
lines changed

tests/test_tipc/static/auto_parallel/llama2/pretrain_config_llama2_13b/intermediate_api_pretrain-llama2_13b.json

Lines changed: 2 additions & 2 deletions
Original file line numberDiff line numberDiff line change
@@ -18,10 +18,10 @@
1818
"virtual_pp_degree": 5,
1919
"sequence_parallel": 0,
2020
"use_flash_attention": true,
21-
"use_fused_rms_norm": false,
21+
"use_fused_rms_norm": true,
2222
"fuse_attention_ffn": true,
2323
"fuse_attention_qkv": true,
24-
"use_fused_rope": false,
24+
"use_fused_rope": true,
2525
"fused_linear_param_grad_add": true,
2626
"max_seq_length": 4096,
2727
"learning_rate": 3e-05,

0 commit comments

Comments
 (0)