set logging_step to 5 with baichuan && qwen benchmark #8928
Conversation
|
Thanks for your contribution! |
| param+="global_batch_size=32 " | ||
| param+="model_item=qwen-qwen-14b_seqlen4096_pretrain " | ||
| param+="max_steps=100 " | ||
| param+="logging_steps=5" |
There was a problem hiding this comment.
此处需要后面加上空格，param+="logging_steps=5 "
| param+="global_batch_size=32 " | ||
| param+="model_item=qwen-qwen-7b_seqlen4096_pretrain " | ||
| param+="max_steps=100 " | ||
| param+="logging_steps=5" |
There was a problem hiding this comment.
此处需要后面加上空格，param+="logging_steps=5 "
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## develop #8928 +/- ##
===========================================
+ Coverage 55.05% 55.12% +0.07%
===========================================
Files 635 635
Lines 99410 99420 +10
===========================================
+ Hits 54729 54805 +76
+ Misses 44681 44615 -66 ☔ View full report in Codecov by Sentry. |
| "warmup_ratio": 0.01, | ||
| "max_grad_norm": 1.0, | ||
| "logging_steps": 1, | ||
| "logging_steps": 5, |
There was a problem hiding this comment.
这个max_steps是100，这里logging_steps设置成 5，打印的ips不够用
| "warmup_ratio": 0.01, | ||
| "max_grad_norm": 1.0, | ||
| "logging_steps": 1, | ||
| "logging_steps": 5, |
There was a problem hiding this comment.
这个max_steps是100，这里logging_steps设置成 5，打印的ips不够用
| param+="global_batch_size=32 " | ||
| param+="model_item=qwen-qwen-7b_seqlen4096_pretrain " | ||
| param+="max_steps=100 " | ||
| param+="logging_steps=5 " |
There was a problem hiding this comment.
这个max_steps是100，这里logging_steps设置成 5，打印的ips不够用
| param+="global_batch_size=32 " | ||
| param+="model_item=qwen-qwen-14b_seqlen4096_pretrain " | ||
| param+="max_steps=100 " | ||
| param+="logging_steps=5 " |
There was a problem hiding this comment.
这个max_steps是100，这里logging_steps设置成 5，打印的ips不够用
3 participants
PR types
Others
PR changes
Others
Description
Others