-
Notifications
You must be signed in to change notification settings - Fork 6
Update llm-jp-tokenizer tag to the latest for cloning #12
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Changes from all commits
File filter
Filter by extension
Conversations
Jump to
Diff view
Diff view
There are no files selected for viewing
| Original file line number | Diff line number | Diff line change | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
|
|
@@ -15,7 +15,9 @@ export PRETRAIN_TORCHVISION_VERSION=0.18.1 | |||||||||
| export PRETRAIN_APEX_VERSION=24.04.01 | ||||||||||
| export PRETRAIN_TRANSFORMER_ENGINE_VERSION=1.4 | ||||||||||
| export PRETRAIN_MEGATRON_TAG=nii-geniac | ||||||||||
| export PRETRAIN_TOKENIZER_TAG=Release-ver3.0b1 | ||||||||||
| # Ensure the appropriate Huggingface tokenizer is included | ||||||||||
| # https://github.com/llm-jp/scripts/pull/12#discussion_r1708415209 | ||||||||||
| export PRETRAIN_TOKENIZER_TAG=v3.0b2 | ||||||||||
ytivy marked this conversation as resolved.
Show resolved
Hide resolved
|
||||||||||
|
|
||||||||||
|
Comment on lines
+20
to
21
Contributor
Author
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more.
Suggested change
Contributor
Author
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. 今対応してしまった方が良さそうなので書いてみました
Member
There was a problem hiding this comment. Choose a reason for hiding this commentThe reason will be displayed to describe this comment to others. Learn more. @YumaTsuta ちょうど蔦さんのコメントへのリンクが取れるので、これをコメントに追加しておくのが親切だと思いました。 |
||||||||||
| module load cuda/${PRETRAIN_CUDA_VERSION} | ||||||||||
| module load /data/cudnn-tmp-install/modulefiles/${PRETRAIN_CUDNN_VERSION} | ||||||||||
|
|
||||||||||
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ablationシート
の18-19行目では学習に3.0b1を使用し、評価に3.0b2を使用することになっています。こちらの設定は学習用なので、変更不要に見えます。
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
コメントが足りていませんでした。
Megatron -> HFのconvert時にHFのtokenizerを指定する必要があり、Release-ver3.0b1では修正版のv3.0b2 (hf tokenizer) が含まれていません。
convert スクリプトではMagatron-LMのスクリプトをベースとするため、事前学習モデルの環境構築スクリプトを利用しており、こちらの修正PRを出しました
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
なるほど、了解です。
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
approveしましたが、上記はスクリプトにコメントが書いてあった方がよいかもしれません。