diff --git a/docs.json b/docs.json index f51fc74de5..4ebfc3d79b 100644 --- a/docs.json +++ b/docs.json @@ -197,15 +197,22 @@ "group": "What are runs?", "pages": [ "models/runs", - "models/runs/tags", - "models/runs/multiple-runs-per-process", - "models/runs/run-colors", - "models/runs/filter-runs", + "models/runs/run-identifiers", + "models/runs/initialize-run", + "models/runs/run-states", + "models/runs/view-logged-runs", + "models/runs/customize-run-display", "models/runs/forking", - "models/runs/grouping", - "models/runs/manage-runs", "models/runs/resuming", "models/runs/rewind", + "models/runs/grouping", + "models/runs/filter-runs", + "models/runs/search-runs", + "models/runs/stop-runs", + "models/runs/delete-runs", + "models/runs/tags", + "models/runs/manage-runs", + "models/runs/run-colors", "models/runs/color-code-runs", "models/runs/alert" ] diff --git a/images/app_ui/artifacts_tab.png b/images/app_ui/artifacts_tab.png deleted file mode 100644 index 9e723cac57..0000000000 Binary files a/images/app_ui/artifacts_tab.png and /dev/null differ diff --git a/images/app_ui/demo_move_runs.gif b/images/app_ui/demo_move_runs.gif deleted file mode 100644 index c4638da8e9..0000000000 Binary files a/images/app_ui/demo_move_runs.gif and /dev/null differ diff --git a/images/app_ui/export_to_csv.gif b/images/app_ui/export_to_csv.gif deleted file mode 100644 index da118cedb7..0000000000 Binary files a/images/app_ui/export_to_csv.gif and /dev/null differ diff --git a/images/app_ui/howto_move_runs.gif b/images/app_ui/howto_move_runs.gif deleted file mode 100644 index 37f32c7fe4..0000000000 Binary files a/images/app_ui/howto_move_runs.gif and /dev/null differ diff --git a/images/app_ui/wandb-run-page-workspace-tab.png b/images/app_ui/wandb-run-page-workspace-tab.png deleted file mode 100644 index 60b5cf86ec..0000000000 Binary files a/images/app_ui/wandb-run-page-workspace-tab.png and /dev/null differ diff --git a/images/app_ui/wandb_run_overview_page.png b/images/app_ui/wandb_run_overview_page.png deleted file mode 100644 index 9db74bd92b..0000000000 Binary files a/images/app_ui/wandb_run_overview_page.png and /dev/null differ diff --git a/images/app_ui/wandb_run_page_files_tab.png b/images/app_ui/wandb_run_page_files_tab.png deleted file mode 100644 index 2bad821f78..0000000000 Binary files a/images/app_ui/wandb_run_page_files_tab.png and /dev/null differ diff --git a/images/app_ui/wandb_run_page_log_tab.png b/images/app_ui/wandb_run_page_log_tab.png deleted file mode 100644 index 9d819808b2..0000000000 Binary files a/images/app_ui/wandb_run_page_log_tab.png and /dev/null differ diff --git a/images/data_vis/group.png b/images/data_vis/group.png deleted file mode 100644 index cd1b3d85d6..0000000000 Binary files a/images/data_vis/group.png and /dev/null differ diff --git a/images/runs/demo-project.gif b/images/runs/demo-project.gif deleted file mode 100644 index 8ce551957e..0000000000 Binary files a/images/runs/demo-project.gif and /dev/null differ diff --git a/images/runs/group_job_type.png b/images/runs/group_job_type.png new file mode 100644 index 0000000000..31d8ca9e08 Binary files /dev/null and b/images/runs/group_job_type.png differ diff --git a/images/runs/group_name_3_groups.png b/images/runs/group_name_3_groups.png new file mode 100644 index 0000000000..1ea0ac0445 Binary files /dev/null and b/images/runs/group_name_3_groups.png differ diff --git a/images/runs/run_log_example_2.png b/images/runs/run_log_example_2.png index ead79dfcc3..5a78efd225 100644 Binary files a/images/runs/run_log_example_2.png and b/images/runs/run_log_example_2.png differ diff --git a/images/runs/run_logs_info.png b/images/runs/run_logs_info.png new file mode 100644 index 0000000000..3e74dd665c Binary files /dev/null and b/images/runs/run_logs_info.png differ diff --git a/images/runs/run_table_all.png b/images/runs/run_table_all.png new file mode 100644 index 0000000000..cde533176b Binary files /dev/null and b/images/runs/run_table_all.png differ diff --git a/images/runs/runs_artifacts_info.png b/images/runs/runs_artifacts_info.png new file mode 100644 index 0000000000..1ada287af3 Binary files /dev/null and b/images/runs/runs_artifacts_info.png differ diff --git a/images/runs/runs_files_info.png b/images/runs/runs_files_info.png new file mode 100644 index 0000000000..4c5046db38 Binary files /dev/null and b/images/runs/runs_files_info.png differ diff --git a/images/runs/stop-run-manual-status.png b/images/runs/stop-run-manual-status.png deleted file mode 100644 index 191fa14a12..0000000000 Binary files a/images/runs/stop-run-manual-status.png and /dev/null differ diff --git a/images/runs/stop-run-manual.png b/images/runs/stop-run-manual.png deleted file mode 100644 index fd002410ef..0000000000 Binary files a/images/runs/stop-run-manual.png and /dev/null differ diff --git a/images/runs/stop-run-terminal.png b/images/runs/stop-run-terminal.png deleted file mode 100644 index 1ea74c85e1..0000000000 Binary files a/images/runs/stop-run-terminal.png and /dev/null differ diff --git a/images/runs/unique-run-id.png b/images/runs/unique-run-id.png index 8679046cfd..521e2a6f4d 100644 Binary files a/images/runs/unique-run-id.png and b/images/runs/unique-run-id.png differ diff --git a/images/track/demo_grouping.png b/images/track/demo_grouping.png deleted file mode 100644 index 694dd70a0f..0000000000 Binary files a/images/track/demo_grouping.png and /dev/null differ diff --git a/images/track/demo_grouping_options_for_line_plots.gif b/images/track/demo_grouping_options_for_line_plots.gif deleted file mode 100644 index df7cd49d03..0000000000 Binary files a/images/track/demo_grouping_options_for_line_plots.gif and /dev/null differ diff --git a/images/track/demo_no_grouping.png b/images/track/demo_no_grouping.png deleted file mode 100644 index 51b273d30c..0000000000 Binary files a/images/track/demo_no_grouping.png and /dev/null differ diff --git a/images/track/distributed_training_wgrouping_1.png b/images/track/distributed_training_wgrouping_1.png deleted file mode 100644 index 8dcba9d9fb..0000000000 Binary files a/images/track/distributed_training_wgrouping_1.png and /dev/null differ diff --git a/images/track/distributed_training_wgrouping_2.png b/images/track/distributed_training_wgrouping_2.png deleted file mode 100644 index 63c9c787b8..0000000000 Binary files a/images/track/distributed_training_wgrouping_2.png and /dev/null differ diff --git a/ja/models/runs.mdx b/ja/models/runs.mdx index 201209a730..0a6b03c2da 100644 --- a/ja/models/runs.mdx +++ b/ja/models/runs.mdx @@ -269,24 +269,16 @@ W&B App UI に移動して run がもはやアクティブではないことを **State** フィールドの隣で、run の状態が `running` から `Killed` に変わります。 - - - 1. run のログを記録していたプロジェクトに移動します。 2. run セレクタ内で停止したい run を選択します。 3. プロジェクトのサイドバーから **Overview** タブを選択します。 4. **State** フィールドの隣の上部ボタンを選択します。 - - - + **State** フィールドの隣で、run の状態が `running` から `Killed` に変わります。 - - - @@ -296,10 +288,6 @@ W&B App UI に移動して run がもはやアクティブではないことを run の状態、run にログ記録されたアーティファクト、run 中に記録されたログファイルなど、特定の run に関する情報を表示します。 - - - - 特定の run を表示するには： 1. [https://wandb.ai/home](https://wandb.ai/home) の W&B App UI に移動します。 @@ -344,19 +332,11 @@ W&B は概要セクションの下に次の情報を保存します： * **Config**: [`wandb.config`](/ja/models/track/config/)で保存された設定パラメータのリスト。 * **Summary**: [`wandb.log()`](/ja/models/track/log/)で保存されたサマリーパラメータのリスト。デフォルトでは、W&B はこの値を最後にログ記録した値に設定します。 - - W&B Dashboard run overview tab - - こちらでプロジェクトの概要の例を確認できます [here](https://wandb.ai/stacey/deep-drive/overview)。 ### Workspace タブ Workspace タブを使用して、生成されたプロットやカスタムプロット、システムメトリクスなどの可視化を表示、検索、グループ化、および配置してください。 - - - - こちらでワークスペースの例を確認できます [here](https://wandb.ai/stacey/deep-drive/workspace?nw=nwuserstacey) ### Runs タブ @@ -417,10 +397,6 @@ Table のある列の値で全行を並べ替えます。 ダッシュボード上の **Group by** ボタンを使用して、特定の列の値で全行をグループ化します。 - - The truth distribution shows small errors: 8s and 2s are confused for 7s and 9s for 2s. - - デフォルトでは、これにより他の数値列が、その列のグループ全体の値の分布を示すヒストグラムに変わります。グループ化は、データ内のより高水準のパターンを理解するのに役立ちます。 @@ -431,7 +407,7 @@ Table のある列の値で全行を並べ替えます。 右上の「ダウンロード」ボタンを選択してログファイルをダウンロードします。 - + ログタブの例はこちらから見ることができます [here](https://app.wandb.ai/stacey/deep-drive/runs/pr0os44x/logs). @@ -439,17 +415,13 @@ Table のある列の値で全行を並べ替えます。 ### Files タブ **Files tab** を使用して、特定の run に関連付けられたファイル（モデルチェックポイント、検証セット例など）を表示してください。 - - - - ファイルタブの例はこちらから見ることができます [here](https://app.wandb.ai/stacey/deep-drive/runs/pr0os44x/files/media/images). ### Artifacts タブ **Artifacts** タブには、指定した run の入力および出力 [Artifacts](/ja/models/artifacts/) が一覧表示されます。 - + Artifacts タブの例はこちらから見ることができます [here](https://wandb.ai/stacey/artifact_july_demo/runs/2cslp2rt/artifacts). diff --git a/ja/models/runs/filter-runs.mdx b/ja/models/runs/filter-runs.mdx index 2f08d0b042..952a2b0d75 100644 --- a/ja/models/runs/filter-runs.mdx +++ b/ja/models/runs/filter-runs.mdx @@ -63,7 +63,3 @@ description: プロジェクトページのサイドバーとテーブルの使 ## run テーブルを CSV にエクスポートする すべての run、ハイパーパラメーター、およびサマリーメトリクスのテーブルを、ダウンロード ボタンを使用して CSV にエクスポートします。 - - - - \ No newline at end of file diff --git a/ja/models/runs/grouping.mdx b/ja/models/runs/grouping.mdx index 9ed08793e2..ea19f03b86 100644 --- a/ja/models/runs/grouping.mdx +++ b/ja/models/runs/grouping.mdx @@ -29,16 +29,8 @@ description: トレーニングと評価 run をグループ化して大規模 `wandb.init()`でグループ化を設定したと仮定すると、UIではデフォルトでrunがグループ化されます。テーブルの上部にある**Group**ボタンをクリックして、これをオン・オフすることができます。こちらはグループ化を設定した[サンプルコード](http://wandb.me/grouping)から生成された[例のプロジェクト](https://wandb.ai/carey/group-demo?workspace=user-carey)です。サイドバーの各「グループ」行をクリックすると、その実験の専用グループページにアクセスできます。 - - - - 上記のプロジェクトページから、左サイドバーで**Group**をクリックすると、[このような専用ページ](https://wandb.ai/carey/group-demo/groups/exp_5?workspace=user-carey)にアクセスできます。 - - - - ## UIでの動的なグループ化 任意の列でrunをグループ化できます。例として、ハイパーパラメーターでグループ化することができます。これがどのように見えるかの例です： @@ -46,22 +38,10 @@ description: トレーニングと評価 run をグループ化して大規模 * **サイドバー**: runがエポック数でグループ化されています。 * **グラフ**: 各線はグループの平均を表し、陰影は分散を示します。この振る舞いはグラフ設定で変更できます。 - - - - ## グループ化をオフにする グループ化ボタンをクリックし、グループフィールドをいつでもクリアすることで、テーブルとグラフをグループ化されていない状態に戻します。 - - - - ## グループ化グラフの設定 -グラフの右上にある編集ボタンをクリックし、**Advanced**タブを選択して線と陰影を変更します。各グループの線には平均、最小、最大値を選択できます。陰影については無効にしたり、最小と最大、標準偏差、標準誤差を表示することができます。 - - - - \ No newline at end of file +グラフの右上にある編集ボタンをクリックし、**Advanced**タブを選択して線と陰影を変更します。各グループの線には平均、最小、最大値を選択できます。陰影については無効にしたり、最小と最大、標準偏差、標準誤差を表示することができます。 \ No newline at end of file diff --git a/ja/models/runs/manage-runs.mdx b/ja/models/runs/manage-runs.mdx index bec33bf241..1ab3cb43b2 100644 --- a/ja/models/runs/manage-runs.mdx +++ b/ja/models/runs/manage-runs.mdx @@ -20,10 +20,6 @@ run をあるプロジェクトから別のプロジェクトに移動するに 4. テーブルの上にある **Move** ボタンを選択します。 5. ドロップダウンから移動先のプロジェクトを選択します。 - - - - ## チームに run を移動する あなたがメンバーであるチームに run を移動するには: @@ -32,8 +28,4 @@ run をあるプロジェクトから別のプロジェクトに移動するに 2. プロジェクトのサイドバーから **Runs** タブを選択します。 3. 移動したい run の横にあるチェックボックスを選択します。 4. テーブルの上にある **Move** ボタンを選択します。 -5. ドロップダウンから移動先のチームとプロジェクトを選択します。 - - - - \ No newline at end of file +5. ドロップダウンから移動先のチームとプロジェクトを選択します。 \ No newline at end of file diff --git a/ja/models/track/project-page.mdx b/ja/models/track/project-page.mdx index 86caaad0a1..79ff8de5a7 100644 --- a/ja/models/track/project-page.mdx +++ b/ja/models/track/project-page.mdx @@ -160,10 +160,6 @@ Runs タブをカスタマイズすると、そのカスタマイズは[Workspac 特定の列の値で **Group by** ボタンを使い、すべての行をグループ化します。 - - The truth distribution shows small errors: 8s and 2s are confused for 7s and 9s for 2s. - - デフォルトでは、これにより他の数値列がヒストグラムに変わり、グループ全体のその列の値の分布を示します。グループ化は、データ内の高レベルのパターンを理解するのに便利です。 diff --git a/ko/models/runs.mdx b/ko/models/runs.mdx index 6cdeaef14a..d9b32e4e09 100644 --- a/ko/models/runs.mdx +++ b/ko/models/runs.mdx @@ -267,24 +267,16 @@ W&B 앱 UI로 이동하여 run이 더 이상 활성 상태가 아닌지 확인 **State** 필드 옆에 있는 run의 상태가 `running`에서 `Killed`로 변경됩니다. - - - 1. run이 기록되고 있는 프로젝트로 이동합니다. 2. run 선택기 내에서 중지할 run을 선택합니다. 3. 프로젝트 사이드바에서 **Overview** 탭을 선택합니다. 4. **State** 필드 옆에 있는 맨 위 버튼을 선택합니다. - - - + **State** 필드 옆에 있는 run의 상태가 `running`에서 `Killed`로 변경됩니다. - - - @@ -294,10 +286,6 @@ W&B 앱 UI로 이동하여 run이 더 이상 활성 상태가 아닌지 확인 Run 상태, run에 기록된 Artifacts, run 중에 기록된 로그 파일 등과 같은 특정 run에 대한 정보를 봅니다. - - - - 특정 run을 보려면 다음을 수행하십시오. 1. [https://wandb.ai/home](https://wandb.ai/home)에서 W&B 앱 UI로 이동합니다. @@ -342,19 +330,11 @@ W&B는 Overview 섹션 아래에 다음 정보를 저장합니다. * **Config**: [`wandb.config`](/ko/models/track/config/)로 저장된 구성 파라미터 목록입니다. * **Summary**: [`wandb.log()`](/ko/models/track/log/)로 저장된 요약 파라미터 목록입니다. 기본적으로 W&B는 이 값을 마지막으로 기록된 값으로 설정합니다. - - W&B 대시보드 run Overview 탭 - - [여기](https://wandb.ai/stacey/deep-drive/overview)에서 프로젝트 Overview의 예를 봅니다. ### Workspace 탭 Workspace 탭을 사용하여 자동 생성된 플롯 및 사용자 지정 플롯, 시스템 메트릭 등과 같은 시각화를 보고, 검색하고, 그룹화하고, 정렬합니다. - - - - [여기](https://wandb.ai/stacey/deep-drive/workspace?nw=nwuserstacey)에서 프로젝트 워크스페이스의 예를 봅니다. ### Runs 탭 @@ -413,10 +393,6 @@ Runs 탭을 사용자 지정하면 사용자 지정이 [Workspace 탭](#workspac 대시보드 위의 **Group by** 버튼을 사용하여 특정 열의 값을 기준으로 모든 행을 그룹화합니다. - - 진실 분포는 작은 오류를 보여줍니다. 8과 2는 7과 9로, 2는 2로 혼동됩니다. - - 기본적으로 이렇게 하면 다른 숫자 열이 해당 그룹 전체의 해당 열에 대한 값 분포를 보여주는 히스토그램으로 바뀝니다. 그룹화는 데이터에서 상위 수준 패턴을 이해하는 데 유용합니다. @@ -438,7 +414,7 @@ W&B가 추적하는 시스템 메트릭의 전체 목록은 [시스템 메트릭 로그 파일을 다운로드하려면 오른쪽 상단 모서리에 있는 **Download** 버튼을 선택합니다. - + [여기](https://app.wandb.ai/stacey/deep-drive/runs/pr0os44x/logs)에서 로그 탭의 예를 봅니다. @@ -446,17 +422,13 @@ W&B가 추적하는 시스템 메트릭의 전체 목록은 [시스템 메트릭 ### Files 탭 **Files 탭**을 사용하여 모델 체크포인트, 검증 세트 예제 등과 같은 특정 run과 연결된 파일을 봅니다. - - - - [여기](https://app.wandb.ai/stacey/deep-drive/runs/pr0os44x/files/media/images)에서 파일 탭의 예를 봅니다. ### Artifacts 탭 **Artifacts** 탭에는 지정된 run에 대한 입력 및 출력 [artifacts](/ko/models/artifacts/)가 나열됩니다. - + [여기](https://wandb.ai/stacey/artifact_july_demo/runs/2cslp2rt/artifacts)에서 artifacts 탭의 예를 봅니다. diff --git a/ko/models/runs/filter-runs.mdx b/ko/models/runs/filter-runs.mdx index 9646831a69..3134b1602b 100644 --- a/ko/models/runs/filter-runs.mdx +++ b/ko/models/runs/filter-runs.mdx @@ -63,7 +63,3 @@ W&B에 기록된 run으로부터 얻은 통찰력을 프로젝트 페이지에 ## Run 테이블을 CSV로 내보내기 다운로드 버튼을 사용하여 모든 run, 하이퍼파라미터 및 요약 메트릭 테이블을 CSV로 내보냅니다. - - - - diff --git a/ko/models/runs/grouping.mdx b/ko/models/runs/grouping.mdx index 70ee006e37..d2d2e901f1 100644 --- a/ko/models/runs/grouping.mdx +++ b/ko/models/runs/grouping.mdx @@ -29,16 +29,8 @@ description: 트레이닝 및 평가 run을 그룹화하여 더 큰 Experiments `wandb.init()` 에서 그룹화를 설정하면 UI에서 기본적으로 Runs 가 그룹화됩니다. 테이블 상단의 **Group** 버튼을 클릭하여 이를 켜거나 끌 수 있습니다. 그룹화를 설정한 [샘플 코드](http://wandb.me/grouping)에서 생성된 [예제 프로젝트](https://wandb.ai/carey/group-demo?workspace=user-carey)가 있습니다. 사이드바에서 각 "Group" 행을 클릭하여 해당 experiment 에 대한 전용 그룹 페이지로 이동할 수 있습니다. - - - - 위의 프로젝트 페이지에서 왼쪽 사이드바의 **Group** 을 클릭하여 [이 페이지](https://wandb.ai/carey/group-demo/groups/exp_5?workspace=user-carey)와 같은 전용 페이지로 이동할 수 있습니다. - - - - ## UI에서 동적으로 그룹화 예를 들어 하이퍼파라미터별로 열별로 Runs 를 그룹화할 수 있습니다. 다음은 그 모양의 예입니다. @@ -46,22 +38,10 @@ description: 트레이닝 및 평가 run을 그룹화하여 더 큰 Experiments * **사이드바**: Runs 는 에포크 수별로 그룹화됩니다. * **그래프**: 각 선은 그룹의 평균을 나타내고 음영은 분산을 나타냅니다. 이 동작은 그래프 설정에서 변경할 수 있습니다. - - - - ## 그룹화 끄기 언제든지 그룹화 버튼을 클릭하고 그룹 필드를 지우면 테이블과 그래프가 그룹 해제된 상태로 돌아갑니다. - - - - ## 그룹화 그래프 설정 그래프 오른쪽 상단 모서리에 있는 편집 버튼을 클릭하고 **Advanced** 탭을 선택하여 선과 음영을 변경합니다. 각 그룹에서 선의 평균, 최소값 또는 최대값을 선택할 수 있습니다. 음영의 경우 음영을 끄고 최소값과 최대값, 표준 편차 및 표준 오차를 표시할 수 있습니다. - - - - diff --git a/ko/models/runs/manage-runs.mdx b/ko/models/runs/manage-runs.mdx index 1abd43d429..aeb86f3aa7 100644 --- a/ko/models/runs/manage-runs.mdx +++ b/ko/models/runs/manage-runs.mdx @@ -20,9 +20,6 @@ run 을 한 프로젝트에서 다른 프로젝트로 이동하려면 다음을 4. 테이블 위의 **Move** 버튼을 선택합니다. 5. 드롭다운에서 대상 프로젝트를 선택합니다. - - - ## 팀으로 run 이동 @@ -33,7 +30,3 @@ run 을 한 프로젝트에서 다른 프로젝트로 이동하려면 다음을 3. 이동하려는 run 옆에 있는 확인란을 선택합니다. 4. 테이블 위의 **Move** 버튼을 선택합니다. 5. 드롭다운에서 대상 팀 및 프로젝트를 선택합니다. - - - - diff --git a/ko/models/track/project-page.mdx b/ko/models/track/project-page.mdx index 275e951a8a..9ee4fb1b64 100644 --- a/ko/models/track/project-page.mdx +++ b/ko/models/track/project-page.mdx @@ -159,10 +159,6 @@ Runs 탭을 사용자 지정하면 사용자 지정은 [Workspace 탭](#workspac 열 머리글에서 **그룹화 기준** 버튼을 사용하여 특정 열의 값을 기준으로 모든 행을 그룹화합니다. - - 진실 분포는 작은 오류를 보여줍니다. 8과 2는 7과 9로, 9는 2로 착각합니다. - - 기본적으로 이렇게 하면 다른 숫자 열이 해당 그룹 전체의 해당 열에 대한 값 분포를 보여주는 히스토그램으로 바뀝니다. 그룹화는 데이터에서 상위 수준 패턴을 이해하는 데 유용합니다. diff --git a/models/automations/automation-events.mdx b/models/automations/automation-events.mdx index e3d89bd034..c393f7bf1b 100644 --- a/models/automations/automation-events.mdx +++ b/models/automations/automation-events.mdx @@ -34,7 +34,7 @@ This section describes the events related to an artifact that can trigger an aut - **An artifact tag is added**: Trigger a specific step of your workflow when an artifact version in a project has a tag applied that matches the **Tag regex** you specify. For example, specify `^europe.*` to trigger a geo-specific workflow when a tag beginning with the string `europe` is added to an artifact version. Artifact tags are used for grouping and filtering, and a given tag can be assigned to multiple artifact versions simultaneously. ### Run events -An automation can be triggered by a change in a [run's status](/models/#run-states) or a change in a [metric value](/models/track/log/#what-data-is-logged-with-specific-wb-api-calls). +An automation can be triggered by a change in a [run's status](/models/(/models/runs/run-states) or a change in a [metric value](/models/track/log/#what-data-is-logged-with-specific-wb-api-calls). #### Run status change @@ -42,7 +42,7 @@ An automation can be triggered by a change in a [run's status](/models/#run-stat - A run with **Killed** status cannot trigger an automation. This status indicates that the run was stopped forcibly by an admin user. -Trigger a workflow when a run changes its [status](/models/runs/#run-states) to **Running**, **Finished**, or **Failed**. Optionally, you can further limit the runs that can trigger an automation by specifying a user or run name filter. +Trigger a workflow when a run changes its [status](/models/runs/(/models/runs/run-states) to **Running**, **Finished**, or **Failed**. Optionally, you can further limit the runs that can trigger an automation by specifying a user or run name filter. ![Screenshot showing a run status change automation](/images/automations/run_status_change.png) diff --git a/models/runs.mdx b/models/runs.mdx index 20f9d97330..9da76b327e 100644 --- a/models/runs.mdx +++ b/models/runs.mdx @@ -6,7 +6,7 @@ title: Overview A *run* is a single unit of computation logged by W&B. You can think of a W&B Run as an atomic element of your whole project. In other words, each run is a record of a specific computation, such as training a model and logging the results, hyperparameter sweeps, and so forth. -Common patterns for initiating a run include, but are not limited to: +Common reasons for initiating a run include, but are not limited to: * Training a model * Changing a hyperparameter and conducting a new experiment @@ -14,10 +14,23 @@ Common patterns for initiating a run include, but are not limited to: * Logging data or a model as a [W&B Artifact](/models/artifacts/) * [Downloading a W&B Artifact](/models/artifacts/download-and-use-an-artifact/) +To initialize a W&B run, call the [`wandb.init()`](/models/ref/python/functions/init) method from the W&B Python SDK. This starts a new run and returns a `wandb.Run` object that you can use to log metrics, artifacts, and other information to the run. For more information about initializing a run, see [Initialize runs](/models/runs/initialize-run). -W&B stores runs that you create into [*projects*](/models/track/project-page/). You can view runs and their properties within the run's project workspace on the W&B App. You can also programmatically access run properties with the [`wandb.Api.Run`](/models/ref/python/experiments/run) object. +Each run object has a [unique identifier known as a *run ID*](/models/runs/run-identifiers#unique-run-identifiers). [You can specify a unique ID](/models/runs/run-identifiers#unique-run-identifiers) or let [W&B randomly generate one for you](/models/runs/run-identifiers#autogenerated-run-ids). Each run object also has a human-readable, non-unique [run name](/models/runs/run-identifiers#run-name). You can specify a name for your run or let W&B randomly generate one for you. You can rename a run after initializing it. -Anything you log with `wandb.Run.log()` is recorded in that run. +W&B logs your run to a [*project*](/models/track/project-page/). You specify the project when you initialize the run with `wandb.init(project="")`. W&B creates a new project if the project does not exist. If the project does exist, W&B logs the run to the project you specified. + + +If you do not specify a project name, W&B stores the run in a project called `Uncategorized`. + + +`wandb.init()` returns a `wandb.Run` object that contains properties of the run, such as its ID, name, configuration, and state. Use the run object to log metrics, artifacts, and other information to the run with methods such as `wandb.Run.log()`, `wandb.Run.log_code()`, and `wandb.Run.use_artifact()`. + +Each run has a state that describes the current status of the run. See [Run states](/models/runs/run-states) for a full list of possible run states. + +[View runs and their properties](/models/runs/view-logged-runs) within the run's project workspace on the W&B App. You can also programmatically access run properties with the [`wandb.Api.Run`](/models/ref/python/experiments/run) object. + +As an example, consider the following code snippet that initializes a W&B run and logs some metrics to it: Pass your W&B entity to the `entity` variable in the code snippets below if you want to follow along. Your entity is your W&B username or team name. You can find it in the URL of your W&B App workspace. For example, if your workspace URL is `https://wandb.ai/nico/awesome-project`, then your entity is `nico`. @@ -57,7 +70,7 @@ wandb: Synced 6 W&B file(s), 0 media file(s), 0 artifact file(s) and 0 other fil wandb: Find logs at: ./wandb/run-20241105_111006-1jx1ud12/logs ``` -The URL W&B returns in the terminal to redirects you to the run's workspace in the W&B App UI. Note that the panels generated in the workspace corresponds to the single point. +W&B returns two URLs in the terminal output. The first URL directs you to the [specific run's workspace](/models/runs/view-logged-runs), and the second URL directs you to the [project](/models/track/project-page) page. Single run workspace @@ -86,7 +99,9 @@ with wandb.init(project="awesome-project", config=config) as run: run.log({"accuracy": acc, "loss": loss}) ``` -This returns the following output: +The training script calls `wandb.Run.log()` 10 times. Each time the script calls `wandb.Run.log()`, W&B logs the accuracy and loss for that epoch. + +Within your terminal, you should see output similar to the following: ```bash wandb: Syncing run jolly-haze-4 @@ -107,173 +122,14 @@ wandb: 🚀 View run jolly-haze-4 at: https://wandb.ai/nico/awesome-project/runs wandb: Find logs at: wandb/run-20241105_111816-pdo5110r/logs ``` -The training script calls `wandb.Run.log()` 10 times. Each time the script calls `wandb.Run.log()`, W&B logs the accuracy and loss for that epoch. Selecting the URL that W&B prints from the preceding output, directs you to the run's workspace in the W&B App UI. +W&B captures the simulated training loop within a single run called `jolly-haze-4`. This is because the script calls `wandb.init()` method only once. -W&B captures the simulated training loop within a single run called `jolly-haze-4`. This is because the script calls `wandb.init()` method only once. +Copy and paste the URL that W&B prints from the previous output into your browser. The URL directs you to the run's workspace in the W&B App UI. For example, the following image shows the workspace for the run `jolly-haze-4`: Training run with logged metrics -As another example, during a [sweep](/models/sweeps/), W&B explores a hyperparameter search space that you specify. W&B implements each new hyperparameter combination that the sweep creates as a unique run. - - -## Initialize a W&B Run - -Initialize a W&B Run with [`wandb.init()`](/models/ref/python/functions/init). The following code snippet shows how to import the W&B Python SDK and initialize a run. - -Ensure to replace values enclosed in angle brackets (`< >`) with your own values: - -```python -import wandb - -with wandb.init(entity="", project="") as run: - # Your code here -``` - -When you initialize a run, W&B logs your run to the project you specify for the project field (`wandb.init(project=""`). W&B creates a new project if the project does not already exist. If the project already exists, W&B stores the run in that project. - - -If you do not specify a project name, W&B stores the run in a project called `Uncategorized`. - - -Each run in W&B has a [unique identifier known as a *run ID*](#unique-run-identifiers). [You can specify a unique ID](#unique-run-identifiers) or let [W&B randomly generate one for you](#autogenerated-run-ids). - -Each run also has a human-readable, non-unique [run name](#name-your-run). You can specify a name for your run or let W&B randomly generate one for you. You can rename a run after initializing it. - -For example, consider the following code snippet: - -```python title="basic.py" -import wandb - -with wandb.init(entity="wandbee", project="awesome-project") as run: - # Your code here -``` -The code snippet produces the following output: - -```bash -🚀 View run exalted-darkness-6 at: -https://wandb.ai/nico/awesome-project/runs/pgbn9y21 -Find logs at: wandb/run-20241106_090747-pgbn9y21/logs -``` - -Since the preceding code did not specify an argument for the id parameter, W&B creates a unique run ID. Where `nico` is the entity that logged the run, `awesome-project` is the name of the project the run is logged to, `exalted-darkness-6` is the name of the run, and `pgbn9y21` is the run ID. - -If you [group runs](/models/runs/grouping/) into experiments, you can move a run into or out of a group or from one group to another. - -Each run has a state that describes the current status of the run. See [Run states](#run-states) for a full list of possible run states. - -## Run states -The proceeding table describes the possible states a run can be in: - -| State | Description | -| ----- | ----- | -| `Crashed` | Run stopped sending heartbeats in the internal process, which can happen if the machine crashes. | -| `Failed` | Run ended with a non-zero exit status. | -| `Finished`| Run ended and fully synced data, or called `wandb.Run.finish()`. | -| `Killed` | Run was forcibly stopped before it could finish. | -| `Running` | Run is still running and has recently sent a heartbeat. | -| `Pending` | Run is scheduled but not yet started (common in sweeps and Launch jobs). | - -### Run states in sweeps - -When runs are part of a [sweep](/models/sweeps/), their states behave independently from the sweep's status: - -- **Individual run states** reflect each run's execution status (Running, Finished, Failed, etc.) -- **Sweep status** controls whether new runs are created, not how existing runs execute -- Pausing or stopping a sweep doesn't affect already-running runs -- Only cancelling a sweep forcibly kills running runs (changes their state to `Killed`) - -For a detailed explanation of how sweep and run statuses interact, see [Understanding sweep and run statuses](/models/sweeps/pause-resume-and-cancel-sweeps#understanding-sweep-and-run-statuses). - -## Unique run identifiers - -Run IDs are unique identifiers for runs. By default, [W&B generates a random and unique run ID for you](#autogenerated-run-ids) when you initialize a new run. You can also [specify your own unique run ID](#custom-run-ids) when you initialize a run. - -### Autogenerated run IDs - -If you do not specify a run ID when you initialize a run, W&B generates a random run ID for you. You can find the unique ID of a run in the W&B App. - -1. Navigate to the [W&B App](https://wandb.ai/home). -2. Navigate to the W&B project you specified when you initialized the run. -3. Within your project's workspace, select the **Runs** tab. -4. Select the **Overview** tab. - -W&B displays the unique run ID in the **Run path** field. The run path consists of the name of your team, the name of the project, and the run ID. The unique ID is the last part of the run path. - -For example, in the proceeding image, the unique run ID is `9mxi1arc`: - - - Run ID location - - - -### Custom run IDs -You can specify your own run ID by passing the `id` parameter to the [`wandb.init()`](/models/ref/python/functions/init) method. - -```python -import wandb - -with wandb.init(entity="", project="", id="") as run: - # Your code here -``` - -You can use a run's unique ID to directly navigate to the run's overview page in the W&B App. The proceeding cell shows the URL path for a specific run: - -```text title="W&B App URL for a specific run" -https://wandb.ai/// -``` - -Where values enclosed in angle brackets (`< >`) are placeholders for the actual values of the entity, project, and run ID. - -## Name your run -In a project, each run has a human-readable non-unique run name. By default, W&B generates a random run name when you initialize a new run without specifying a run name. The name of a run appears within your project's workspace and at the top of the [run's overview page](#overview-tab). - -You can specify a name for your run by passing the `name` parameter to the [`wandb.init()`](/models/ref/python/functions/init) method. - -```python -import wandb - -with wandb.init(entity="", project="", name="") as run: - # Your code here -``` - -In a workspace, each run also has a run display name, which defaults to the same value as the run name. Use the run display name to override the run name displayed in that workspace without renaming the run in the project. - - -Use run names and display names as a way to quickly identify a run in your project workspace. - - - -### Rename a run - -This section shows how to rename a run from a workspace or a report. - - - -1. Navigate to your W&B project. -1. Select the **Workspace** or **Runs** tab from the project sidebar. -1. Search or scroll to the run you want to rename. - - Hover over the run name, click the three vertical dots, then click **Rename run**. -1. To change the run name, update the **Run name** field. To customize the run display name in this workspace, update the **Display name** field. -1. Click **Save**. - -A customized run display name displays in italic font, and an information icon appears next to a customized run display name. Hover over the icon for details. - - -To rename a run from a run set in a [report](/models/reports/edit-a-report/): - -1. Click the pencil icon to open the report editor. -1. In the run set, find the run to rename. Hover over the report name, click the three vertical dots, then click **Edit run name**. -1. To change the run name, update the **Run name** field. To customize the run display name in this workspace, update the **Display name** field. -1. Click **Save**. -1. Click **Publish report**. - -A customized run display name displays in italic font, and an information icon appears next to a customized run display name. Hover over the icon for details. - - - ## Add a note to a run Notes that you add to a specific run appear on the run page in the **Overview** tab and in the table of runs on the project page. @@ -282,412 +138,3 @@ Notes that you add to a specific run appear on the run page in the **Overview** 3. Select the run you want to add a note to from the run selector 4. Choose the **Overview** tab 5. Select the pencil icon next to the **Description** field and add your notes - -## Stop a run -Stop a run from the W&B App or programmatically. - - - -1. Navigate to the terminal or code editor where you initialized the run. -2. Press `Ctrl+D` to stop the run. - -For example, following the preceding instructions, your terminal might looks similar to the following: - -```bash -KeyboardInterrupt -wandb: 🚀 View run legendary-meadow-2 at: https://wandb.ai/nico/history-blaster-4/runs/o8sdbztv -wandb: Synced 5 W&B file(s), 0 media file(s), 0 artifact file(s) and 1 other file(s) -wandb: Find logs at: ./wandb/run-20241106_095857-o8sdbztv/logs -``` - -Navigate to the W&B App to confirm the run is no longer active: - -1. Navigate to the project that your run was logging to. -2. Select the name of the run. - -You can find the name of the run that you stop from the output of your terminal or code editor. For example, in the preceding example, the name of the run is `legendary-meadow-2`. - -3. Choose the **Overview** tab from the project sidebar. - -Next to the **State** field, the run's state changes from `running` to `Killed`. - - - Run stopped via terminal - - - -1. Navigate to the project that your run is logging to. -2. Select the run you want to stop within the run selector. -3. Choose the **Overview** tab from the project sidebar. -4. Select the top button next to the **State** field. - - Manual run stop button - - -Next to the **State** field, the run's state changes from `running` to `Killed`. - - - Run status after manual stop - - - - -See [State fields](#run-states) for a full list of possible run states. - -## View logged runs - -View information about a specific run such as the state of the run, artifacts logged to the run, metrics, system metrics, and more. - -You can view and monitor runs using the W&B App or the `wandb beta leet` terminal UI. - - - - -To view a specific run in the W&B App: - -1. Navigate to the [W&B App](https://wandb.ai/home). -2. Navigate to the W&B project you specified when you initialized the run. -3. Within the project sidebar, select the **Workspace** tab. -4. Within the run selector, click the run you want to view, or enter a partial run name to filter for matching runs. - - - Project navigation demo - - -The URL path of a specific run has the proceeding format: - -```text -https://wandb.ai///runs/ -``` - -Where values enclosed in angle brackets (`< >`) are placeholders for the actual values of the team name, project name, and run ID. - - - - -To view a run locally in your terminal using the `wandb beta leet` terminal UI: - -1. If you started the run locally from a script, navigate to the directory where you ran your code. It contains a `wandb/` directory with a subdirectory per run and a `latest-run/` symbolic link. Each run directory contains a transaction log named in the format `run-.wandb`. - - If you did not start the run locally but downloaded a `.wandb` transaction log file instead, make a note of its location. -2. Start `wandb beta leet` using one of these commands: - - ```bash - # View the latest run, stored in ./wandb/latest-run/ - wandb beta leet - - # Specify a run directory - wandb beta leet ./wandb/run-20250813_124246-n67z9ude - - # Specify a .wandb file - wandb beta leet ./wandb/run-20250813_124246-n67z9ude/run-n67z9ude.wandb - ``` - -![Screenshot demonstrating viewing a run in the LEET terminal UI](/images/leet/leet-tui.png) - -LEET displays a three-panel interface: - -- **Left sidebar**: Run overview with environment variables, configuration, and summary statistics -- **Center panel**: Metrics grid with Braille-style line charts showing your logged metrics -- **Right sidebar**: System metrics including GPU/CPU/RAM utilization - -Get started with these keyboard shortcuts: - -- `h` or `?` - View all keyboard shortcuts -- `/` - Filter metrics by pattern -- `[` / `]` - Toggle left/right sidebars -- `n` / `N` - Navigate between metric pages -- `q` / `CMD+C` - Quit - -See the [`wandb beta leet`](/models/ref/cli/wandb-beta/wandb-beta-leet) reference for more details. - - - - -### Customize how runs are displayed -This section shows how to customize how runs are displayed in your project's workspace and runs table. - - -A workspace is limited to displaying a maximum of 1000 runs, regardless of its configuration. - - -#### Add or remove columns - -Add or remove columns in the **Runs** tab or **Workspace** tab to customize which properties associated with your project are visible. For both cases, you can optionally search by column name using fuzzy search, an exact match, or regular expressions. - -To add a columns in the Runs table or Workspace: -1. In the project sidebar, select either the **Runs** tab or the **Workspace** tab. -1. Above the list of runs, click the **Columns** (six horizontal dashes) button. -1. Select the name of a property within the **Hidden** section of the modal. -1. Drag columns to change their order. -1. Click **Close**. - - -To remove columns in the Runs table or Workspace: -1. In the project sidebar, select either the **Runs** tab or the **Workspace** tab. -1. Above the list of runs, click the **Columns** (six horizontal dashes) button. -1. Select the name of a property within the **Visible & Pinned** section of the modal. -1. Click **Close**. - -#### Sort runs by column - -To sort the list of runs by any visible column: - -1. Hover over the column name, then click its action `...` menu. -1. Click **Sort ascending** or **Sort descending**. - -#### Pin columns - -Pinned columns are shown on the left-hand side. Unpinned columns are shown on the right-hand side of the **Runs** tab or **Workspace** tab. If you pin a column in the **Runs** tab, it is also pinned in the **Workspace** tab. Similarly, if you pin a column in the **Workspace** tab, it is also pinned in the **Runs** tab. - -To pin a column: -1. In the project sidebar, navigate to the **Workspace** or **Runs** tab. -1. Click the **Columns** (six horizontal dashes) button. -1. Within the **Visible & Pinned** section of the modal, click on the pin icon next to the column name you want to pin. - -To unpin a column: -1. In the project sidebar, navigate to the **Workspace** or **Runs** tab. -1. Either hover over the column name, then click its action `...` menu, or click the **Columns** (six horizontal dashes) button and click on the pin icon next to the column name you want to unpin. -1. Click **Unpin column**. - -#### Customize run name truncation - -By default, long run names are truncated in the middle for readability. To customize the truncation of run names: - -1. Click the action `...` menu at the top of the list of runs. -1. Set **Run name cropping** to crop the end, middle, or beginning. - -### Overview tab -Use the **Overview** tab to learn about specific run information in a project, such as: - -* **Author**: The W&B entity that creates the run. -* **Command**: The command that initializes the run. -* **Description**: A description of the run that you provided. This field is empty if you do not specify a description when you create the run. You can add a description to a run with the W&B App UI or programmatically with the Python SDK. -* **Tracked Hours**: The amount of time the run is actively computing or logging data, excluding any pauses or waiting periods. This metric helps you understand the actual computational time spent on your run. -* **Runtime**: Measures the total time from the start to the end of the run. It's the wall-clock time for the run, including any time where the run is paused or waiting for resources. This metric provides the complete elapsed time for your run. -* **Git repository**: The git repository associated with the run. You must [enable git](/platform/app/settings-page/user-settings/#personal-github-integration) to view this field. -* **Host name**: Where W&B computes the run. W&B displays the name of your machine if you initialize the run locally on your machine. -* **Name**: The name of the run. -* **OS**: Operating system that initializes the run. -* **Python executable**: The command that starts the run. -* **Python version**: Specifies the Python version that creates the run. -* **Run path**: Identifies the unique run identifier in the form `entity/project/run-ID`. -* **Start time**: The timestamp when you initialize the run. -* **State**: The [state of the run](#run-states). -* **System hardware**: The hardware W&B uses to compute the run. -* **Tags**: A list of strings. Tags are useful for organizing related runs together or applying temporary labels like `baseline` or `production`. -* **W&B CLI version**: The W&B CLI version installed on the machine that hosted the run command. -* **Git state**: The most recent git commit SHA of a repository or working directory where the run is initialized. This field is empty if you do not enable Git when you create the run or if the git information is not available. - -W&B stores the proceeding information below the overview section: - -* **Artifact Outputs**: Artifact outputs produced by the run. -* **Config**: List of config parameters saved with [`wandb.Run.config`](/models/track/config/). -* **Summary**: List of summary parameters saved with [`wandb.Run.log()`](/models/track/log/). By default, W&B sets this value to the last value logged. - - - W&B Dashboard run overview tab - - -View an example project overview [here](https://wandb.ai/stacey/deep-drive/overview). - -### Workspace tab -Use the Workspace tab to view, search, group, and arrange visualizations such as autogenerated and custom plots, system metrics, and more. - - - Run workspace tab - - -View an example project workspace [here](https://wandb.ai/stacey/deep-drive/workspace?nw=nwuserstacey) - -### Runs tab -{/* Keep this in sync with /guide/models/track/project-page.md */} -Use the Runs tab to filter, group, and sort your runs. - - - Runs table - - -{/* [Try these yourself →](https://wandb.ai/stacey/mnist-viz/artifacts/predictions/baseline/d888bc05719667811b23/files/predictions.table.json) */} - - -The proceeding tabs demonstrate some common actions you can take in the Runs tab. - - - -The Runs tab shows details about runs in the project. It shows a large number of columns by default. - -- To view all visible columns, scroll the page horizontally. -- To change the order of the columns, drag a column to the left or right. -- To pin a column, hover over the column name, click the action menu `...`. that appears, then click **Pin column**. Pinned columns appear near the left of the page, after the **Name** column. To unpin a pinned column, choose **Unpin column** -- To hide a column, hover over the column name, click the action menu `...`. that appears, then click **Hide column**. To view all columns that are currently hidden, click **Columns**. -- To show, hide, pin, and unpin multiple columns at once, click **Columns**. - - Click the name of a hidden column to unhide it. - - Click the name of a visible column to hide it. - - Click the pin icon next to a visible column to pin it. - -When you customize the Runs tab, the customization is also reflected in the **Runs** selector of the [Workspace tab](#workspace-tab). - - -Sort all rows in a Table by the value in a given column. - -1. Hover your mouse over the column title. A kebab menu will appear (three vertical docs). -2. Select on the kebab menu (three vertical dots). -3. Choose **Sort Asc** or **Sort Desc** to sort the rows in ascending or descending order, respectively. - - - Confident predictions - - -The preceding image demonstrates how to view sorting options for a Table column called `val_acc`. - - -Filter all rows by an expression with the **Filter** button above the dashboard. - - - Incorrect predictions filter - - -Select **Add filter** to add one or more filters to your rows. Three dropdown menus will appear. From left to right the filter types are based on: Column name, Operator , and Values - -| | Column name | Binary relation | Value | -| ----------- | ----------- | ----------- | ----------- | -| Accepted values | String | =, ≠, ≤, ≥, IN, NOT IN, | Integer, float, string, timestamp, null | - - -The expression editor shows a list of options for each term using autocomplete on column names and logical predicate structure. You can connect multiple logical predicates into one expression using "and" or "or" (and sometimes parentheses). - - - -Group all rows by the value in a particular column with the **Group by** button above the dashboard. - - - Error distribution analysis - - -By default, this turns other numeric columns into histograms that each show the distribution of values for that column across the group. Grouping is helpful for understanding higher-level patterns in your data. - - -The **Group by** feature is distinct from a [run's run group](/models/runs/grouping/). You can group runs by run group. To move a run to a different run group, refer to [Assign a group or job type to a run](#assign-a-group-or-job-type-to-a-run). - - - - -### Logs tab -The **Log tab** shows output printed on the command line such as the standard output (`stdout`) and standard error (`stderr`). - -Choose the **Download** button in the upper right hand corner to download the log file. - - - Run logs tab - - -View an example logs tab [here](https://app.wandb.ai/stacey/deep-drive/runs/pr0os44x/logs). - -### Files tab -Use the **Files tab** to view files associated with a specific run such as model checkpoints, validation set examples, and more - - - Run files tab - - -View an example files tab [here](https://app.wandb.ai/stacey/deep-drive/runs/pr0os44x/files/media/images). - -### Artifacts tab -The **Artifacts** tab lists the input and output [artifacts](/models/artifacts/) for the specified run. - - - Run artifacts tab - - -View [example artifact graphs](/models/artifacts/explore-and-traverse-an-artifact-graph/). - -## Delete runs - -Delete one or more runs from a project with the W&B App. - -1. Navigate to the project that contains the runs you want to delete. -2. Select the **Runs** tab from the project sidebar. -3. Select the checkbox next to the runs you want to delete. -4. Choose the **Delete** button (trash can icon) above the table. -5. From the modal that appears, choose **Delete**. - - -Once a run with a specific ID is deleted, its ID may not be used again. Trying to initiate a run with a previously deleted ID will show an error and prevent initiation. - - - -For projects that contain a large number of runs, you can use either the search bar to filter runs you want to delete using Regex or the filter button to filter runs based on their status, tags, or other properties. - - -### Run deletion workflow - -The following diagram illustrates the complete run deletion process, including the handling of associated artifacts and Model Registry links: - -```mermaid -graph TB - Start([User Initiates
Run Deletion]) --> RunSelect[Select Runs
to Delete] - - RunSelect --> DeletePrompt{Delete Associated
Artifacts?} - - DeletePrompt -->|No| DeleteRunOnly[Delete Run Only

- Run metadata removed
- Artifacts remain available
- Can still access artifacts] - - DeletePrompt -->|Yes| CheckArtifacts[Check for
Associated Artifacts] - - CheckArtifacts --> HasRegistry{Artifacts Linked to
Model Registry?} - - HasRegistry -->|Yes| RegistryWarning[⚠️ Warning

Registry models will be deleted
Production aliases affected] - HasRegistry -->|No| DirectDelete - - RegistryWarning --> ConfirmRegistry{Confirm Registry
Model Deletion?} - - ConfirmRegistry -->|No| DeleteRunOnly - ConfirmRegistry -->|Yes| DirectDelete[Delete Run + Artifacts

- Run metadata removed
- Artifacts permanently deleted
- Registry links removed
- Cannot be recovered] - - DeleteRunOnly --> PartialEnd([Run Deleted
Artifacts Preserved]) - DirectDelete --> FullEnd([Run + Artifacts
Permanently Deleted]) - - style Start fill:#e1f5fe,stroke:#333,stroke-width:2px,color:#000 - style DeletePrompt fill:#fff3e0,stroke:#333,stroke-width:2px,color:#000 - style RegistryWarning fill:#ffecb3,stroke:#333,stroke-width:2px,color:#000 - style DirectDelete fill:#ffebee,stroke:#333,stroke-width:2px,color:#000 - style DeleteRunOnly fill:#e8f5e9,stroke:#333,stroke-width:2px,color:#000 - style PartialEnd fill:#c8e6c9,stroke:#333,stroke-width:2px,color:#000 - style FullEnd fill:#ffcdd2,stroke:#333,stroke-width:2px,color:#000 -``` - - -**Important** - -When you delete a run and choose to delete associated artifacts, the artifacts are permanently removed and can't be recovered, even if the run is restored later. This includes models linked to the Model Registry. - - -## Organize runs - -This section provides instructions on how to organize runs using groups and job types. By assigning runs to groups (for example, experiment names) and specifying job types (for example, preprocessing, training, evaluation, debugging), you can streamline your workflow and improve model comparison. - -### Assign a group or job type to a run - -Each run in W&B can be categorized by **group** and a **job type**: - -- **Group**: a broad category for the experiment, used to organize and filter runs. -- **Job type**: the function of the run, such as `preprocessing`, `training`, or `evaluation`. - -The proceeding [example workspace](https://wandb.ai/stacey/model_iterz?workspace=user-stacey), trains a baseline model using increasing amounts of data from the Fashion-MNIST dataset. The workspace uses colors to represent the amount of data used: - -- **Yellow to dark green** indicate increasing amounts of data for the baseline model. -- **Light blue to violet to magenta** indicate amounts of data for a more complex "double" model with additional parameters. - -Use W&B's filtering options and search bar to compare runs based on specific conditions, such as: -- Training on the same dataset. -- Evaluating on the same test set. - -When you apply filters, the **Table** view is updated automatically. This allows you to identify performance differences between models, such as determining which classes are significantly more challenging for one model compared to another. - -{/* ### Search runs - -Search for a specific run by name in the sidebar. You can use regex to filter down your visible runs. The search box affects which runs are shown on the graph. Here's an example: - -### Filter runs - -### Organize runs */} diff --git a/models/runs/customize-run-display.mdx b/models/runs/customize-run-display.mdx new file mode 100644 index 0000000000..b43c260f99 --- /dev/null +++ b/models/runs/customize-run-display.mdx @@ -0,0 +1,96 @@ +--- +title: View all runs a project +description: This section shows how to customize how runs are displayed in your project's runs table +--- + +{/* +A workspace is limited to displaying a maximum of 1000 runs, regardless of its configuration. + */} + +View all runs logged to your W&B project in the **Runs** tab of your project sidebar. The *Runs table* shows details about all of your runs in a project. Use the Runs table to compare runs, sort runs by specific columns, and organize runs into groups. + +The following image shows the Runs table for a project named `deep-drive`: + + + Runs table + + +## Add columns + +Add columns in the Runs table to customize which properties associated with your project are visible. + +To add a columns in the Runs table: +1. In the project sidebar, select the **Runs** tab. +1. Above the list of runs, click the **Columns** (six horizontal dashes) button. +1. Select the name of a property within the **Hidden** section of the modal. +1. Drag columns to change their order. +1. Click **Close**. + +## Remove columns + +Remove columns in the Runs table to customize which properties associated with your project are visible. + +To remove columns in the Runs table: +1. In the project sidebar, select the **Runs** tab. +1. Above the list of runs, click the **Columns** (six horizontal dashes) button. +1. Select the name of a property within the **Visible & Pinned** section of the modal. +1. Click **Close**. + +## Move columns +To move columns in the Runs table: +1. In the project sidebar, select the **Runs** tab. +1. Drag a column to the left or right. + +## Pin columns + +Pinned columns are shown on the left-hand side. Unpinned columns are shown on the right-hand side of the Runs tab. If you pin a column in the **Runs** tab, it is also pinned in the **Workspace** tab. Similarly, if you pin a column in the **Workspace** tab, it is also pinned in the **Runs** tab. + +To pin a column: +1. In the project sidebar, navigate to the **Runs** tab. +1. Click the **Columns** (six horizontal dashes) button. +1. Within the **Visible & Pinned** section of the modal, click on the pin icon next to the column name you want to pin. + +To unpin a column: +1. In the project sidebar, navigate to the **Runs** tab. +1. Either hover over the column name, then click its action `...` menu, or click the **Columns** (six horizontal dashes) button and click on the pin icon next to the column name you want to unpin. +1. Click **Unpin column**. + + +W&B persists columns you pin or unpin in the Runs table in the Runs selector of the Workspace tab. + + +## Hide columns + +To hide columns in the Runs table: +1. In the project sidebar, select the **Runs** tab. +1. Hover over the column name, click the action menu `...` that appears. +1. Click **Hide column**. + +To view all columns that are currently hidden, click **Columns**. + +## Sort runs by column + +Sort runs by any visible column in the Runs table. This is particularly useful if you want to view the best (or worst) recorded value. + +To sort the list of runs by any visible column: + +1. Hover over the column name, then click its action `...` menu. +1. Optinally hover over **Show latest**. From the dropdown, select **latest**, **min**, or **max**. +1. Click **Sort ascending** or **Sort descending**. + +The following animation shows sorting runs by the maximum value of a logged metric: + + + Sort by min/max values + + + +W&B persists the sort order you select in the Runs table in the Runs selector of the Workspace tab. + + +## Export runs table to CSV + +Export the table of all your runs, hyperparameters, and summary metrics to a CSV with the download button. + +1. In the project sidebar, select the **Runs** tab. +1. Above the list of runs, click the **Download** (downward arrow) button. diff --git a/models/runs/delete-runs.mdx b/models/runs/delete-runs.mdx new file mode 100644 index 0000000000..c2c5ff4c97 --- /dev/null +++ b/models/runs/delete-runs.mdx @@ -0,0 +1,62 @@ +--- +title: Delete runs +description: Learn how to delete runs from a W&B project using the W&B App. +--- + +Delete one or more runs from a project with the W&B App. + +1. Navigate to the project that contains the runs you want to delete. +2. Select the **Runs** tab. +3. Select the checkbox next to the runs you want to delete. +4. Choose the **Delete** button (trash can icon) above the table. +5. From the modal that appears, choose **Delete**. + + +Once a run with a specific ID is deleted, its ID may not be used again. Trying to initiate a run with a previously deleted ID will show an error and prevent initiation. + + + +For projects that contain a large number of runs, you can use either the search bar to filter runs you want to delete using Regex or the filter button to filter runs based on their status, tags, or other properties. + + +### Run deletion workflow + +The following diagram illustrates the complete run deletion process, including the handling of associated artifacts and Model Registry links: + +```mermaid +graph TB + Start([User Initiates
Run Deletion]) --> RunSelect[Select Runs
to Delete] + + RunSelect --> DeletePrompt{Delete Associated
Artifacts?} + + DeletePrompt -->|No| DeleteRunOnly[Delete Run Only

- Run metadata removed
- Artifacts remain available
- Can still access artifacts] + + DeletePrompt -->|Yes| CheckArtifacts[Check for
Associated Artifacts] + + CheckArtifacts --> HasRegistry{Artifacts Linked to
Model Registry?} + + HasRegistry -->|Yes| RegistryWarning[⚠️ Warning

Registry models will be deleted
Production aliases affected] + HasRegistry -->|No| DirectDelete + + RegistryWarning --> ConfirmRegistry{Confirm Registry
Model Deletion?} + + ConfirmRegistry -->|No| DeleteRunOnly + ConfirmRegistry -->|Yes| DirectDelete[Delete Run + Artifacts

- Run metadata removed
- Artifacts permanently deleted
- Registry links removed
- Cannot be recovered] + + DeleteRunOnly --> PartialEnd([Run Deleted
Artifacts Preserved]) + DirectDelete --> FullEnd([Run + Artifacts
Permanently Deleted]) + + style Start fill:#e1f5fe,stroke:#333,stroke-width:2px,color:#000 + style DeletePrompt fill:#fff3e0,stroke:#333,stroke-width:2px,color:#000 + style RegistryWarning fill:#ffecb3,stroke:#333,stroke-width:2px,color:#000 + style DirectDelete fill:#ffebee,stroke:#333,stroke-width:2px,color:#000 + style DeleteRunOnly fill:#e8f5e9,stroke:#333,stroke-width:2px,color:#000 + style PartialEnd fill:#c8e6c9,stroke:#333,stroke-width:2px,color:#000 + style FullEnd fill:#ffcdd2,stroke:#333,stroke-width:2px,color:#000 +``` + + +**Important** + +When you delete a run and choose to delete associated artifacts, the artifacts are permanently removed and can't be recovered, even if the run is restored later. This includes models linked to the Model Registry. + \ No newline at end of file diff --git a/models/runs/filter-runs.mdx b/models/runs/filter-runs.mdx index 8d4bb9d775..1180b23c71 100644 --- a/models/runs/filter-runs.mdx +++ b/models/runs/filter-runs.mdx @@ -1,15 +1,12 @@ --- -description: How to use the sidebar and table on the project page -title: Filter and search runs +description: Filter runs based on their status, tags, regular expressions (RegEx), or other properties in the W&B App. +title: Filter runs --- -Use your project page to gain insights from runs logged to W&B. You can filter and search runs from both the **Workspace** page and the **Runs** page. - -## Filter runs Filter runs based on their status, [tags](#filter-runs-with-tags), [regular expressions (RegEx)](#filter-runs-with-regular-expressions-regex) or other properties with the filter button. -See Customize run colors for more information on how to [edit, randomize, and reset run colors](/models/runs/run-colors/). +{/* See Customize run colors for more information on how to [edit, randomize, and reset run colors](/models/runs/run-colors/). */} ### Filter runs with tags @@ -19,60 +16,37 @@ Filter runs based on their tags with the filter button. 2. Select the **Filter** button, which looks like a funnel, at the top of the runs table. 3. From left to right, select `"Tags"` from the dropdown menu, select a logic operator, and select a filter search value. + + Filter runs by tags + + ### Filter runs with regex If regex doesn't provide you the desired results, you can make use of [tags](/models/runs/filter-runs/#filter-runs-with-tags) to filter out the runs in Runs Table. Tags can be added either on run creation or after they're finished. Once the tags are added to a run, you can add a tag filter as shown in the gif below. - - Filter runs by tags - 1. Click on the **Runs** tab from the project sidebar. 2. Click on the search box at the top of the runs table. 3. Ensure that the **RegEx** toggle (.*) is enabled (the toggle should be blue). 4. Enter your regular expression in the search box. -## Search runs - -Use regular expressions (RegEx) to find runs with the regular expression you specify. When you type a query in the search box, that will filter down the visible runs in the graphs on the workspace as well as filtering the rows of the table. - -## Group runs +### Filter runs by other properties -To group runs by one or more columns (including hidden columns): - -1. Below the search box, click the **Group** button, which looks like a lined sheet of paper. -1. Select one or more columns to group results by. -1. Each set of grouped runs is collapsed by default. To expand it, click the arrow next to the group name. - -## Sort runs by minimum and maximum values -Sort the runs table by the minimum or maximum value of a logged metric. This is particularly useful if you want to view the best (or worst) recorded value. - -The following steps describe how to sort the run table by a specific metric based on the minimum or maximum recorded value: - -1. Hover your mouse over the column with the metric you want to sort with. -2. Select the kebab menu (three vertical lines). -3. From the dropdown, select either **Show min** or **Show max**. -4. From the same dropdown, select **Sort by asc** or **Sort by desc** to sort in ascending or descending order, respectively. +Filter all rows by an expression with the **Filter** button above the dashboard. - Sort by min/max values + Incorrect predictions filter -## Search End Time for runs +Select **Add filter** to add one or more filters to your rows. Three dropdown menus will appear. From left to right the filter types are based on: Column name, Operator , and Values -We provide a column named `End Time` that logs that last heartbeat from the client process. The field is hidden by default. +| | Column name | Binary relation | Value | +| ----------- | ----------- | ----------- | ----------- | +| Accepted values | String | =, ≠, ≤, ≥, IN, NOT IN, | Integer, float, string, timestamp, null | - - End Time column - - -## Export runs table to CSV -Export the table of all your runs, hyperparameters, and summary metrics to a CSV with the download button. +The expression editor shows a list of options for each term using autocomplete on column names and logical predicate structure. You can connect multiple logical predicates into one expression using "and" or "or" (and sometimes parentheses). - - Modal with preview of export to CSV - {/* ## Edit run colors diff --git a/models/runs/forking.mdx b/models/runs/forking.mdx index 2739f5dd2b..77ab52b9a7 100644 --- a/models/runs/forking.mdx +++ b/models/runs/forking.mdx @@ -1,5 +1,5 @@ --- -description: Forking a W&B run +description: Explore different parameters or models from a specific point in an experiment without impacting the original run. title: Fork a run --- @@ -7,44 +7,80 @@ title: Fork a run The ability to fork a run is in private preview. Contact W&B Support at support@wandb.com to request access to this feature. -Use `fork_from` when you initialize a run with [`wandb.init()`](/models/ref/python/functions/init) to "fork" from an existing W&B run. When you fork from a run, W&B creates a new run using the `run ID` and `step` of the source run. +Explore different parameters or models from a specific point in an experiment without impacting the original run. To do this, "fork" from an existing W&B run. When you fork from a run, W&B creates a new run based on the the source run's `run ID` and `step`. After you initialize a forked run, you can continue logging to the new run independently from the source run. -Forking a run enables you to explore different parameters or models from a specific point in an experiment without impacting the original run. +{/* Use `fork_from` when you initialize a run with [`wandb.init()`](/models/ref/python/functions/init) to "fork" from an existing W&B run. When you fork from a run, W&B creates a new run using the `run ID` and `step` of the source run. */} * Forking a run requires [`wandb`](https://pypi.org/project/wandb/) SDK version >= 0.16.5 * Forking a run requires monotonically increasing steps. You can not use non-monotonic steps defined with [`define_metric()`](/models/ref/python/experiments/run#define_metric) to set a fork point because it would disrupt the essential chronological order of run history and system metrics. +Use the `fork_from` parameter in [`wandb.init()`](/models/ref/python/functions/init) to fork from an existing run. Specify the source run's unique `run ID` and the `step` you want to start the forked run from. -## Start a forked run + +See [Unique run identifiers](/models/runs/run-identifiers#unique-run-identifiers) to learn more about run IDs and how to locate them. + + +## Fork from a previously logged run -To fork a run, use the `fork_from` argument in [`wandb.init()`](/models/ref/python/functions/init) and specify the source `run ID` and the `step` from the source run to fork from: +The following code snippet demonstrates how to fork from a run that you previously logged to W&B. This might occur if you want to fork from a run in a different script or notebook. Or if someone else logged the source run that you want to fork from. + +Replace ``, ``, and `` with your own values. ```python import wandb -# Initialize a run to be forked later +# The unique ID of the source run to fork from +source_run_id = "" + +# Specify the step to fork from +fork_step = 200 + +# Fork the run with wandb.init( - project="your_project_name", - entity="your_entity_name" + project="", + entity="", + fork_from=f"{source_run_id}?_step={fork_step}", +) as forked_run: + pass +``` + +## Fork from a run in the same script + +The following code snippet shows how to fork from a run within the same script. This might occur if you want explore different parameters or models from a specific point within the same script or notebook. + +```python +import wandb + +# Initialize a run +with wandb.init( + project="", + entity="" ) as original_run: - # ... perform training or logging ... + # ...training logic goes here ... pass -# Fork the run from a specific step +# Specify the step to fork from +fork_step = 200 + +# Use the original run's ID and specify the step to fork from with wandb.init( - project="your_project_name", - entity="your_entity_name", - fork_from=f"{original_run.id}?_step=200", + project="", + entity="", + fork_from=f"{original_run.id}?_step={fork_step}", ) as forked_run: + # ...training logic goes here ... pass ``` -## Continue from a forked run -After initializing a forked run, you can continue logging to the new run. You can log the same metrics for continuity and introduce new metrics. +Note that you use the `original_run.id` property to obtain the unique run ID of the original run. + +## Example script + +For example, the following code example shows how to first fork a run and then how to log metrics to the forked run starting from a training step of 200. -For example, the following code example shows how to first fork a run and then how to log metrics to the forked run starting from a training step of 200: +Copy and paste the following code into a Python script or notebook cell. Replace `` and `` with your own values. ```python import wandb @@ -52,8 +88,8 @@ import math # Initialize the first run and log some metrics with wandb.init( - project="your_project_name", - entity="your_entity_name" + project="", + entity="" ) as run1: for i in range(300): run1.log({"metric": i}) @@ -61,8 +97,8 @@ with wandb.init( # Fork from the first run at a specific step and log the # metric starting from step 200 with wandb.init( - project="your_project_name", - entity="your_entity_name", + project="", + entity="", fork_from=f"{run1.id}?_step=200" ) as run2: # Continue logging in the new run diff --git a/models/runs/grouping.mdx b/models/runs/grouping.mdx index 39ece10b20..b2c3917afd 100644 --- a/models/runs/grouping.mdx +++ b/models/runs/grouping.mdx @@ -1,133 +1,146 @@ --- -description: Group training and evaluation runs into larger experiments -title: Group runs into experiments +description: Organize your runs into groups and other properties. +title: Organize runs --- -Group individual jobs into experiments by passing a unique **group** name to **wandb.init()**. +Organize your runs into *groups*. A group is a collection of runs that share a common purpose, such as training runs for a specific model or evaluation runs for a specific dataset. -## Use cases +You can also organize runs by other properties such as *job type*. [Job types](/models/runs/grouping#organize-runs-by-job-type) indicate the function of a run, such as `preprocessing`, `training`, or `evaluation`. -1. **Distributed training:** Use grouping if your experiments are split up into different pieces with separate training and evaluation scripts that should be viewed as parts of a larger whole. -2. **Multiple processes**: Group multiple smaller processes together into an experiment. -3. **K-fold cross-validation**: Group together runs with different random seeds to see a larger experiment. Here's [an example](https://github.com/wandb/examples/tree/master/examples/wandb-sweeps/sweeps-cross-validation) of k-fold cross-validation with sweeps and grouping. +{/* +### Define groups using environment variables -There are several ways to set grouping: +Use `WANDB_RUN_GROUP` to specify a group for your runs as an environment variable. For more on this, check our docs for [Environment Variables](/models/track/environment-variables/). **Group** should be unique within your project and shared by all runs in the group. You can use `wandb.util.generate_id()` to generate a unique 8 character string to use in all your processes— for example, `os.environ["WANDB_RUN_GROUP"] = "experiment-" + wandb.util.generate_id()` */} -### 1. Set group in your script +## Organize runs into groups -Pass an optional group and `job_type` to `wandb.init()`. This gives you a dedicated group page for each experiment, which contains the individual runs. For example:`wandb.init(group="experiment_1", job_type="eval")` +You can add runs to a group programmatically using the W&B Python SDK or interactively in the W&B App. -### 2. Set a group environment variable + +W&B stores group names as a run [`wandb.Run.group`](/models/ref/python/experiments/run#property-run-group) property. + -Use `WANDB_RUN_GROUP` to specify a group for your runs as an environment variable. For more on this, check our docs for [Environment Variables](/models/track/environment-variables/). **Group** should be unique within your project and shared by all runs in the group. You can use `wandb.util.generate_id()` to generate a unique 8 character string to use in all your processes— for example, `os.environ["WANDB_RUN_GROUP"] = "experiment-" + wandb.util.generate_id()` -### 3. Toggle grouping by columns in the UI + + +Programmatically add one or more runs to a group with the W&B Python SDK. Pass the name of your group as an argument to the `group` parameter when you initialize a run with `wandb.init(group="")`. The name you provide is used to organize and filter runs in the W&B App. -You can dynamically group by any column, including a column that is hidden. For example, if you use `wandb.Run.config` to log batch size or learning rate, you can then group by those hyperparameters dynamically in the web app. The **Group by** feature is distinct from a [run's run group](/models/runs/grouping/). You can group runs by run group. To move a run to a different run group, refer to [Set a group in the UI](#set-a-group-in-the-ui). +The following example creates three groups named `A`, `B`, and `C`. Each group contains three runs. - -In the list of runs, the **Group** column is hidden by default. - +```python +import wandb -To group runs by one or more columns: +entity = "" +project = "" -1. Click **Group**. -1. Click the names of one or more columns. -1. If you selected more than one column, drag them to change the grouping order. -1. Click anywhere outside of the form to dismiss it. +for group in ["A", "B", "C"]: + for i in range(3): + with wandb.init(entity=entity, project=project, group=group, name=f"{group}_run_{i}") as run: + # Simulate some training + for step in range(100): + run.log({ + "acc": 0.5 + (step / 100) * 0.3 + (i * 0.05), + "loss": 1.0 - (step / 100) * 0.5 + }) +``` -### 4. Manage groups in the UI +Within the project's workspace, you can view runs organized by group. The following image shows the runs table grouped by the group name. Three groups named `A`, `B`, and `C` appear in the runs table, each containing three runs. -#### Manage a run -After a run is initialized, you can move it to a new group from your workspace or its **Runs** page. + + Runs table grouped by group name + + + 1. Navigate to your W&B project. -1. Select the **Workspace** or **Runs** tab from the project sidebar. -1. To move a single run, search or scroll to it. Hover over the run name, click the three vertical dots, then click **Move to another group**. -1. To create a new group, click **New group**. Type a group name, then submit the form. -1. To delete a run, click the three vertical dots, then click **Delete run**. -1. To move a run, select the new group from the list, then click **Move**. - -#### Manage runs in bulk -This section shows how to manage multiple runs at once, across one or more run groups. - -1. If necessary, [toggle grouping by columns](#toggle-grouping-by-column-in-the-ui), grouping by the **Group** column. -1. Expand a group to view its runs. In a single bulk operation, you can move runs from different groups to the same target group. Expand all relevant groups. -1. To select all runs in a group, click the checkbox next to the group name. -1. To select individual runs, click their checkboxes. -1. To select all runs in all groups, click the checkbox next to the **Name** column heading. -1. Above the table, choose a bulk operation: **Tag**, **Move to group**, **Move to project**, or **New sweep**. - -#### Create a group -To create a new run group: +2. Select the **Runs** tab from the project sidebar. +3. Above the list of runs, click the **Group** button. +4. Click the checkbox next to one or more runs you want to group. +5. Select **Move to group**. +6. In the modal, select an existing group or create a new group. + + + + +### View groups + +View runs organized by group in the W&B App: + +1. In your project sidebar, select the **Runs** tab. +2. Above the list of runs, click the **Group** button. +3. From the dropdown, select either **Group**. + +### Move runs between groups + +Move runs in one group to another group: + 1. Navigate to your W&B project. -1. Select the **Workspace** or **Runs** tab from the project sidebar. -1. Click **New group**. Type a group name, then submit the form. +1. Select the **Runs** tab from the project sidebar. +1. Select one or more runs by clicking their checkboxes. +1. Above the table, click **Move to group**. +1. Within the modal, select the target group or create a new group. +1. Click **Move**. + +### Remove a group -#### Delete a group +1. Navigate to your W&B project. +1. Select the **Runs** tab from the project sidebar. +1. Above the list of runs, click the **Group** button. +1. From the dropdown, select the **X** next to the name of the group you want to remove. + +### Delete a group To delete a group, remove all runs from it. This automatically deletes the group. -### Customize how runs are displayed -You can customize how runs are displayed in your project from the **Workspace** or **Runs** tabs. Both tabs use the same display configuration. +## Organize runs by job type -To customize which columns are visible: -1. Above the list of runs, click **Columns**. -1. Click the name of a hidden column to show it. Click the name of a visible column to hide it. - - You can optionally search by column name using fuzzy search, an exact match, or regular expressions. Drag columns to change their order. -1. Click **Done** to close the column browser. +Organize runs by their *job type*. A job type indicates the function of a run, such as `preprocessing`, `training`, or `evaluation`. -To sort the list of runs by any visible column: + +W&B stores job types as a run [`wandb.Run.job_type`](/models/ref/python/experiments/run#property-run-job-type) property. + -1. Hover over the column name, then click its action `...` menu. -1. Click **Sort ascending** or **Sort descending**. +Add a job type to a run by passing the `job_type` parameter to `wandb.init(job_type="")`. For example, the following code snippet creates runs with job types of either `training` or `evaluation`: -Pinned columns are shown on the right-hand side. To pin or unpin a column: -1. Hover over the column name, then click its action `...` menu. -1. Click **Pin column** or **Unpin column**. +```python +import wandb -By default, long run names are truncated in the middle for readability. To customize the truncation of run names: +entity = "" +project = "" -1. Click the action `...` menu at the top of the list of runs. -1. Set **Run name cropping** to crop the end, middle, or beginning. +for job_type in ["training", "evaluation"]: + for i in range(2): + with wandb.init(entity=entity, project=project, job_type=job_type, name=f"{job_type}_run_{i}") as run: + # Simulate some process + for step in range(50): + run.log({ + "metric1": 0.2 + (step / 50) * 0.4 + (i * 0.03), + "metric2": 0.8 - (step / 50) * 0.3 + }) -## Distributed training with grouping +``` -Suppose you set grouping in `wandb.init()`, we will group runs by default in the UI. You can toggle this on and off by clicking the **Group** button at the top of the table. Here's an [example project](https://wandb.ai/carey/group-demo?workspace=user-carey) generated from [sample code](https://wandb.me/grouping) where we set grouping. You can click on each "Group" row in the sidebar to get to a dedicated group page for that experiment. +The following image shows runs organized by the job type: - Grouped runs view + Ungrouped runs table -From the project page above, you can click a **Group** in the left sidebar to get to a dedicated page like [this one](https://wandb.ai/carey/group-demo/groups/exp_5?workspace=user-carey): +### View runs organized by job type + +View runs organized by group in the W&B App: + +1. In your project sidebar, select the **Runs** tab. +2. Above the list of runs, click the **Group** button. +3. From the dropdown, select either **Job Type**. + - - Group details page - -## Grouping dynamically in the UI -You can group runs by any column, for example by hyperparameter. Here's an example of what that looks like: -* **Sidebar**: Runs are grouped by the number of epochs. -* **Graphs**: Each line represents the group's mean, and the shading indicates the variance. This behavior can be changed in the graph settings. - - Dynamic grouping by epochs - -## Turn off grouping -Click the grouping button and clear group fields at any time, which returns the table and graphs to their ungrouped state. - - Ungrouped runs table - -## Grouping graph settings -Click the edit button in the upper right corner of a graph and select the **Advanced** tab to change the line and shading. You can select the mean, minimum, or maximum value for the line in each group. For the shading, you can turn off shading, and show the min and max, the standard deviation, and the standard error. - - Line plot grouping options - diff --git a/models/runs/multiple-runs-per-process.mdx b/models/runs/initialize-run.mdx similarity index 72% rename from models/runs/multiple-runs-per-process.mdx rename to models/runs/initialize-run.mdx index 4544f36ef0..674331f8fc 100644 --- a/models/runs/multiple-runs-per-process.mdx +++ b/models/runs/initialize-run.mdx @@ -1,15 +1,60 @@ --- -description: Manage multiple runs in a single Python process using W&B’s reinit functionality -title: Create and manage multiple runs in a single process +title: Initialize runs --- +Initialize a W&B Run with [`wandb.init()`](/models/ref/python/functions/init). + +By default, W&B assumes each Python process has only one active run at a time when you call `wandb.init()`. If you call `wandb.init()` again, W&B will either return the same run or finish the old run before starting a new one. How W&B handles multiple calls to `wandb.init()` in the same process depends on the environment (notebook vs. non-notebook) and the `reinit` configuration. + +To manage multiple active runs in the same process, see [Multiple runs in one process](/models/runs/initialize-run#multiple-runs-in-one-process). + + +W&B recommends using a `with` block when calling `wandb.init()`. This ensures that W&B properly finalizes the run and uploads all data when the block ends. + + +## Single run per process + +The following code snippet shows how to import the W&B Python SDK and initialize a run. + +{/* The following code snippet shows how to import the W&B Python SDK and initialize a run. + +Ensure to replace values enclosed in angle brackets (`< >`) with your own values: + +```python +import wandb + +with wandb.init(entity="", project="") as run: + # Your code here +``` */} + +```python title="basic.py" +import wandb + +with wandb.init(entity="nico", project="awesome-project") as run: + # Your training logice here +``` +The code snippet produces the following output: + +```bash +🚀 View run exalted-darkness-6 at: +https://wandb.ai/nico/awesome-project/runs/pgbn9y21 +Find logs at: wandb/run-20241106_090747-pgbn9y21/logs +``` + +The output shows that W&B logs the run `exalted-darkness-6` to the project `awesome-project` under the entity `nico`. `pgbn9y21` is the unique run ID that W&B generates for this run. + +{/* TEMP */} +## Multiple runs in one process + Manage multiple runs in a single Python process. This is useful for workflows where you want to keep a primary process active while creating short-lived secondary processes for sub-tasks. Some use cases include: - Keeping a single “primary” run active throughout a script while spinning up short-lived “secondary” runs for evaluations or sub-tasks. - Orchestrating sub-experiments in a single file. - Logging from one “main” process to several runs that represent different tasks or time periods. -By default, W&B assumes each Python process has only one active run at a time when you call `wandb.init()`. If you call `wandb.init()` again, W&B will either return the same run or finish the old run before starting a new one, depending on the configuration. The content in this guide explains how to use `reinit` to modify the `wandb.init()` behavior to enable multiple runs in a single Python process. +By default, W&B assumes each Python process has only one active run at a time when you call `wandb.init()`. If you call `wandb.init()` again, W&B will either return the same run or finish the old run before starting a new one, depending on the configuration. + +The content in this guide explains how to use `reinit` to modify the `wandb.init()` behavior to enable multiple runs in a single Python process. **Requirements** @@ -17,13 +62,13 @@ By default, W&B assumes each Python process has only one active run at a time wh To manage multiple runs in a single Python process, you must have W&B Python SDK version `v0.19.10` or newer. -## `reinit` options +### `reinit` options Use the `reinit` parameter to configure how W&B handles multiple calls to `wandb.init()`. The following table describes valid arguments and their effects: | | Description | Creates a run? | Example use case | |----------------|----------------|----------------| -----------------| -| `create_new` |Create a new run with `wandb.init()` without finishing existing, active runs. W&B does not automatically switch the global `wandb.Run` to new runs. You must hold onto each run object yourself. See the [multiple runs in one process example](/models/runs/multiple-runs-per-process/#example-multiple-runs-in-one-process) below for details. | Yes | Ideal for creating and managing concurrent processes. For example, a “primary” run that remains active while you start or end “secondary” runs.| +| `create_new` |Create a new run with `wandb.init()` without finishing existing, active runs. W&B does not automatically switch the global `wandb.Run` to new runs. You must hold onto each run object yourself. See the [multiple runs in one process example](/models/runs/initialize-run/#example-multiple-runs-in-one-process) below for details. | Yes | Ideal for creating and managing concurrent processes. For example, a “primary” run that remains active while you start or end “secondary” runs.| | `finish_previous` | Finish all active runs with `run.finish()` before creating a new one run with `wandb.init()`. Default behavior for non notebook environments. | Yes | Ideal when you want to break sequential sub-processes into separate individual runs. | | `return_previous` | Return the most recent, unfinished run. Default behavior for notebook environments. | No | | @@ -31,27 +76,30 @@ Use the `reinit` parameter to configure how W&B handles multiple calls to `wandb W&B does not support `create_new` mode for [W&B Integrations](/models/integrations) that assume a single global run, such as Hugging Face Trainer, Keras callbacks, and PyTorch Lightning. If you use these integrations, you should run each sub-experiment in a separate process. -## Specifying `reinit` +### Specifying `reinit` {/* There are several ways to create and manage multiple runs in a single Python process: */} - Use `wandb.init()` with the `reinit` argument directly: ```python import wandb - wandb.init(reinit="") + with wandb.init(reinit="") as run: + # Your code here ``` - Use `wandb.init()` and pass a `wandb.Settings` object to the `settings` parameter. Specify `reinit` in the `Settings` object: ```python import wandb - wandb.init(settings=wandb.Settings(reinit="")) + with wandb.init(settings=wandb.Settings(reinit="")) as run: + # Your code here ``` - Use `wandb.setup()` to set the `reinit` option globally for all runs in the current process. This is useful if you want to configure the behavior once and have it apply to all subsequent `wandb.init()` calls in that process. ```python import wandb - wandb.setup(wandb.Settings(reinit="")) + with wandb.setup(wandb.Settings(reinit="")) as run: + # Your code here ``` - Specify the desired value for `reinit` in the environment variable `WANDB_REINIT`. Defining an environment variable applies the `reinit` option to `wandb.init()` calls. @@ -82,7 +130,7 @@ with wandb.init() as experiment_results_run: experiment_results_run.log(experiment_results) ``` -## Example: Concurrent processes +### Example: Concurrent processes Suppose you want to create a primary process that remains open for the script's entire lifespan, while periodically spawning short-lived secondary processes without finishing the primary process. For example, this pattern can be useful if you want to train a model in the primary run, but compute evaluations or do other work in separate runs. diff --git a/models/runs/manage-runs.mdx b/models/runs/manage-runs.mdx index 791b182519..fbed6c1a0c 100644 --- a/models/runs/manage-runs.mdx +++ b/models/runs/manage-runs.mdx @@ -1,17 +1,14 @@ --- -title: Move runs +title: Move runs between projects and teams +description: Learn how to move runs between projects and teams using the W&B App or the W&B Python SDK. --- -This page shows how to move a run from one project to another, into or out of a team, or from one team to another. You must have access to the run at its current and new locations. +Move runs from one project to another, or between teams, using either the W&B App or the W&B Python SDK. You must have access to the run at its current and new locations. When you move a run, historical artifacts associated with it are not moved. To move an artifact manually, you can use the [`wandb artifact get`](/models/ref/cli/wandb-artifact/wandb-artifact-get/) SDK command or the [`Api.artifact` API](/models/ref/python/public-api/api/#artifact) to download the artifact, then use [`wandb artifact put`](/models/ref/cli/wandb-artifact/wandb-artifact-put/) or the `Api.artifact` API to upload it to the run's new location. -To customize the **Runs** tab, refer to [Project page](/models/track/project-page/#runs-tab). - -If you group runs into experiments, refer to [Set a group in the UI](/models/runs/grouping/#set-a-group-in-the-ui). - ## Move runs between your projects To move runs from one project to another: @@ -19,23 +16,36 @@ To move runs from one project to another: 1. Navigate to the project that contains the runs you want to move. 2. Select the **Runs** tab from the project sidebar. 3. Select the checkbox next to the runs you want to move. -4. Choose the **Move** button above the table. +4. Choose the **Move to project** button above the table. 5. Select the destination project from the dropdown. - - Demo of moving a run between projects - - -## Move runs to a team +## Move runs between your teams Move runs to a team you are a member of: 1. Navigate to the project that contains the runs you want to move. 2. Select the **Runs** tab from the project sidebar. 3. Select the checkbox next to the runs you want to move. -4. Choose the **Move** button above the table. +4. Choose the **Move to project** button above the table. 5. Select the destination team and project from the dropdown. - - Demo of moving a run to a team - \ No newline at end of file +{/* ## Move runs to a group + +You can group runs by moving them into an existing group or creating a new group. + +1. Navigate to the project that contains the runs you want to move. +2. Select the **Runs** tab from the project sidebar. +3. Select the checkbox next to the runs you want to move. +4. Choose the **Move to group** button above the table. +5. In the modal, select the target group or create a new group. +6. Click **Move**. */} + +{/* #### Manage runs in bulk +This section shows how to manage multiple runs at once, across one or more run groups. + +1. If necessary, [toggle grouping by columns](#toggle-grouping-by-column-in-the-ui), grouping by the **Group** column. +1. Expand a group to view its runs. In a single bulk operation, you can move runs from different groups to the same target group. Expand all relevant groups. +1. To select all runs in a group, click the checkbox next to the group name. +1. To select individual runs, click their checkboxes. +1. To select all runs in all groups, click the checkbox next to the **Name** column heading. +1. Above the table, choose a bulk operation: **Tag**, **Move to group**, **Move to project**, or **New sweep**. */} \ No newline at end of file diff --git a/models/runs/resuming.mdx b/models/runs/resuming.mdx index 99903db00e..5b0371a20e 100644 --- a/models/runs/resuming.mdx +++ b/models/runs/resuming.mdx @@ -6,7 +6,9 @@ title: Resume a run Specify how a run should behave in the event that run stops or crashes. To resume or enable a run to automatically resume, you will need to specify the unique run ID associated with that run for the `id` parameter: ```python +import wandb with wandb.init(entity="", project="", id="", resume="") as run: + # Your training code here ``` @@ -41,6 +43,7 @@ The following code snippet shows how to accomplish this with the W&B Python SDK: ```python with wandb.init(entity="", project="", id="", resume="must") as run: + # Your training code here ``` diff --git a/models/runs/rewind.mdx b/models/runs/rewind.mdx index 719b5f76cc..cd886ca9eb 100644 --- a/models/runs/rewind.mdx +++ b/models/runs/rewind.mdx @@ -1,9 +1,8 @@ --- -description: Rewind +description: Rewind a run to correct or modify its history without losing original data. title: Rewind a run --- -# Rewind a run The option to rewind a run is in private preview. Contact W&B Support at support@wandb.com to request access to this feature. @@ -78,7 +77,7 @@ run2.finish() ## View an archived run -After you rewind a run, you can explore archived run with the W&B App UI. Follow these steps to view archived runs: +After you rewind a run, you can explore archived run with the W&B App. Follow these steps to view archived runs: 1. **Access the Overview Tab:** Navigate to the [**Overview** tab](./#overview-tab) on the run's page. This tab provides a comprehensive view of the run's details and history. 2. **Locate the Forked From field:** Within the **Overview** tab, find the `Forked From` field. This field captures the history of the resumptions. The **Forked From** field includes a link to the source run, allowing you to trace back to the original run and understand the entire rewind history. diff --git a/models/runs/run-identifiers.mdx b/models/runs/run-identifiers.mdx new file mode 100644 index 0000000000..b9b5f9fc34 --- /dev/null +++ b/models/runs/run-identifiers.mdx @@ -0,0 +1,181 @@ +--- +title: Run IDs and names +description: Learn how to find a run's unique identifier and run name, how to create a custom run ID, and how to customize a run's name. +--- + +When you initialize a W&B Run, W&B assigns that run a [unique identifier known as a *run ID*](/models/runs/run-identifiers#run-id). Each run also has a human-readable [non-unique *run name*](/models/runs/run-identifiers#run-name) that you can customize. + +## Run ID + +Run IDs uniquely identify runs. By default, W&B generates a [random and unique run ID](#autogenerated-run-ids) for you when you initialize a new run. You can also [specify your own unique run ID](#create-a-custom-run-id) when you [initialize a run](/models/runs/initialize-run). + +### Find a run's unique ID + +W&B returns the unique run ID in the terminal when you initialize a run. For example, consider the following code snippet that initializes a W&B run: + +```python +import wandb +entity = "nico" # Replace with your W&B entity +project = "awesome-project" +with wandb.init(entity=entity, project=project) as run: + # Your code here +``` + +Within the terminal, W&B returns: + +```bash +wandb: Syncing run earnest-sunset-1 +wandb: ⭐️ View project at https://wandb.ai/nico/awesome-project +wandb: 🚀 View run at https://wandb.ai/nico/awesome-project/runs/1jx1ud12 +``` + +The last part of the run URL (`1jx1ud12`) is the unique run ID. + +You can also find a run's unique ID in the W&B App: + +1. Navigate to the [W&B App](https://wandb.ai/home). +2. Navigate to the W&B project you specified when you initialized the run. +3. Within your project's workspace, select either the **Workspace** or **Runs** tab. +4. Select the run you want to view. +5. Choose the **Overview** tab. + +W&B displays the run ID in the **Run path** field. The run path consists of the name of your team, the name of the project, and the run ID. The unique ID is the last part of the run path. + +For example, in the following image, the unique run ID is `9mxi1arc`: + + + Run ID location + + + +Use a run's unique ID to directly navigate to that run's overview page in the W&B App. The following cell shows the URL path for a specific run: + +```text title="W&B App URL for a specific run" +https://wandb.ai/// +``` + +Where values enclosed in angle brackets (`< >`) are placeholders for the actual values of the entity, project, and run ID. + + +### Autogenerated run IDs + +W&B generates a random run ID for you if you do not specify a run ID when you initialize a run. + +### Create a custom run ID + +Pass your desired run ID as a string to the `id` paramater when you initialize a run: + +```python +import wandb + +with wandb.init(entity="", project="", id="") as run: + # Your code here +``` + +## Run name + +Each run has a human-readable, non-unique run name. By default, W&B generates a random run name when you initialize a new run if you do not specify a run name for it. The name of a run appears within your project's workspace and at the top of the [run's **Overview** page](#overview-tab). + +Continuing from the previous example, the name of the run is `glwoing-shadows-8`. + + + Run ID location + + +You can name your run when [you initialize it](/models/runs/run-identifiers#create-a-custom-run-name) or [rename](/models/runs/run-identifiers#rename-a-run) it after initializing it. + +### Create a custom run name + +Specify a name for your run by passing the `name` parameter to the [`wandb.init()`](/models/ref/python/functions/init) method. + +```python +import wandb + +with wandb.init(entity="", project="", name="") as run: + # Your code here +``` + +### Rename a run + +Rename a run after initializing it programmatically with Python SDK or interactively with the W&B App. + + + + +Use [`wandb.Api.Run`](/models/ref/python/public-api/api#method-api-run) to access a run logged to W&B. This returns a [run object](/models/ref/python/public-api/run#property-run-name) that you can use to update the run name. Call `wandb.Api.Run.update()` method to persist changes. + +Ensure to replace the values enclosed in angle brackets (`< >`) with your own values: + +```python +import wandb + +api = wandb.Api() + +# Access run by its path +run = api.run(path = "//") + +# Specify a new run name +run.name = "" +run.update() +``` + + + + + +1. Navigate to your W&B project. +1. Select the **Workspace** or **Runs** tab. +1. Search or scroll to the run you want to rename. +1. Hover over the run name, click the three vertical dots, then click **Rename run**. +1. To change the run name, update the **Run name** field. +1. Click **Save**. + + + +## Run display name + +Each run also has a *run display name* that you can customize for each workspace. The display name defaults to the same value as the run name. The display name appears in the run's workspace and runs table. + + +If you change a run's display name in one workspace, it does not change the run's display name in other workspaces or projects. + + +Use the run display name to override the run name displayed in that workspace without renaming the run in the project. + +### Rename a display name + +Rename a run's display name from the W&B App: + +1. Navigate to your W&B project. +1. Select the **Workspace** or **Runs** tab. +1. Search or scroll to the run you want to rename. +1. Hover over the run name, click the three vertical dots, then click **Rename run**. +1. Specify a new value for the **Display name** field. +1. Click **Save**. + +## Customize run name truncation + +By default, long run names are truncated in the middle for readability. To customize the truncation of run names: + +1. Click the action `...` menu at the top of the list of runs. +1. Set **Run name cropping** to crop the end, middle, or beginning. + +{/* A customized run display name displays in italic font, and an information icon appears next to a customized run display name. Hover over the icon for details. */} + +{/* + + + + + +To rename a run from a run set in a [report](/models/reports/edit-a-report/): + +1. Click the pencil icon to open the report editor. +1. In the run set, find the run to rename. Hover over the report name, click the three vertical dots, then click **Edit run name**. +1. To change the run name, update the **Run name** field. To customize the run display name in this workspace, update the **Display name** field. +1. Click **Save**. +1. Click **Publish report**. + +A customized run display name displays in italic font, and an information icon appears next to a customized run display name. Hover over the icon for details. + + */} \ No newline at end of file diff --git a/models/runs/run-states.mdx b/models/runs/run-states.mdx new file mode 100644 index 0000000000..b40970eca7 --- /dev/null +++ b/models/runs/run-states.mdx @@ -0,0 +1,70 @@ +--- +title: Run states +description: Learn about the different states a W&B run can have and what they mean. +--- + +[Run states](/models/runs/run-states#run-states) indicate the current status of a W&B run. You can [view the state](/models/runs/run-states#view-the-state-of-a-run) of a run in the W&B App or programmatically using the W&B Python SDK. + +## Run states +The following table describes the possible states a run can be in: + +| State | Description | +| ----- | ----- | +| `Crashed` | Run stopped sending heartbeats in the internal process, which can happen if the machine crashes. | +| `Failed` | Run ended with a non-zero exit status. | +| `Finished`| Run ended and fully synced data, or called `wandb.Run.finish()`. | +| `Killed` | Run was forcibly stopped before it could finish. | +| `Running` | Run is still running and has recently sent a heartbeat. | +| `Pending` | Run is scheduled but not yet started (common in sweeps and Launch jobs). | + +### Run states in sweeps + +When runs are part of a [sweep](/models/sweeps/), their states behave independently from the sweep's status: + +- **Individual run states** reflect each run's execution status (Running, Finished, Failed, etc.) +- **Sweep status** controls whether new runs are created, not how existing runs execute +- Pausing or stopping a sweep doesn't affect already-running runs +- Only cancelling a sweep forcibly kills running runs (changes their state to `Killed`) + +For a detailed explanation of how sweep and run statuses interact, see [Understanding sweep and run statuses](/models/sweeps/pause-resume-and-cancel-sweeps#understanding-sweep-and-run-statuses). + + +## View the state of a run + +Programmatically or interactively view a run's state with the Python SDK or W&B App. + + + + +Use the `state` property of the [`wandb.Api.Run`](/models/ref/python/public-api/runs) object to access the current state of a run. + +Copy and paste the following code snippet into your Python environment. Replace the values enclosed in angle brackets (`< >`) with your own values: + +```python +import wandb + +api = wandb.Api() + +runs = api.runs(path="/") + +# Access run object's properties +for run in runs: + print(f"Run: {run.name}") + print(f"Run state: {run.state}") + print() +``` + + + +View the state of a run from the W&B App: + +1. Navigate to your W&B project. +1. Select the **Workspace** or **Runs** tab from the project sidebar. +1. Search or scroll to the run you want to view. +1. Select the run to open the run overview page. +1. Choose the **Overview** tab. + +Next to the **State** field, view the current state of the run. + + + \ No newline at end of file diff --git a/models/runs/search-runs.mdx b/models/runs/search-runs.mdx new file mode 100644 index 0000000000..4d18b89405 --- /dev/null +++ b/models/runs/search-runs.mdx @@ -0,0 +1,16 @@ +--- +title: "Search runs" +description: Learn how to search runs using regular expressions (RegEx) and export the runs table to CSV in the W&B App. +--- + + +Use regular expressions (RegEx) to find runs with the regular expression you specify. When you type a query in the search box, that will filter down the visible runs in the graphs on the workspace as well as filtering the rows of the table. + + +## Search End Time for runs + +We provide a column named `End Time` that logs that last heartbeat from the client process. The field is hidden by default. + + + End Time column + \ No newline at end of file diff --git a/models/runs/stop-runs.mdx b/models/runs/stop-runs.mdx new file mode 100644 index 0000000000..a3ed6ea5b3 --- /dev/null +++ b/models/runs/stop-runs.mdx @@ -0,0 +1,46 @@ +--- +title: Stop runs +description: Learn how to stop runs programmatically using the W&B Python SDK or manually from the W&B App. +--- + +Stop a run programmatically with the W&B Python SDK or manually from the W&B App. + + + +1. Navigate to the terminal or code editor where you initialized the run. +2. Press `Ctrl+D` to stop the run. + +For example, following the previous instructions, your terminal might looks similar to the following: + +```bash +KeyboardInterrupt +wandb: 🚀 View run legendary-meadow-2 at: https://wandb.ai/nico/history-blaster-4/runs/o8sdbztv +wandb: Synced 5 W&B file(s), 0 media file(s), 0 artifact file(s) and 1 other file(s) +wandb: Find logs at: ./wandb/run-20241106_095857-o8sdbztv/logs +``` + +Navigate to the W&B App to confirm the run is no longer active: + +1. Navigate to the project that your run was logging to. +2. Select the name of the run. + +You can find the name of the run that you stop from the output of your terminal or code editor. For example, in the preceding example, the name of the run is `legendary-meadow-2`. + +3. Choose the **Overview** tab. + +Next to the **State** field, the run's state changes from `running` to `Killed`. + + + + +1. Navigate to the project that your run is logging to. +2. Select the run you want to stop within the run selector. +3. Choose the **Overview** tab. +4. Select the top button next to the **State** field. + +Next to the **State** field, the run's state changes from `running` to `Killed`. + + + + +See [State fields](/models/runs/run-states) for a full list of possible run states. \ No newline at end of file diff --git a/models/runs/tags.mdx b/models/runs/tags.mdx index 0a9a141d92..60b7478354 100644 --- a/models/runs/tags.mdx +++ b/models/runs/tags.mdx @@ -72,7 +72,7 @@ This method is best suited to applying a tag or tags to a single run manually. ## Remove tags from one or more runs -Tags can also be removed from runs with the W&B App UI. +Tags can also be removed from runs with the W&B App. diff --git a/models/runs/view-logged-runs.mdx b/models/runs/view-logged-runs.mdx new file mode 100644 index 0000000000..be0ad838c5 --- /dev/null +++ b/models/runs/view-logged-runs.mdx @@ -0,0 +1,137 @@ +--- +title: View a specific run in a project +description: Learn how to view a specific logged run and its properties using the W&B App or the LEET terminal UI. +--- + +View information about a specific run such as the state of the run, artifacts logged to the run, metrics logged to a run, system metrics, and more. + + +## View a logged run + +You can view and monitor runs using the W&B App or the `wandb beta leet` terminal UI. + + + + +To view a specific run in the W&B App: + +1. Navigate to the [W&B App](https://wandb.ai/home). +2. Navigate to the W&B project you specified when you initialized the run. +3. Within the project sidebar, select the **Workspace** tab. +4. Within the run selector, click the run you want to view, or enter a partial run name to filter for matching runs. + +Alternatively, you can directly access a specific run's workspace by entering its URL in your browser. The URL path of a specific run has the following format: + +```text +https://wandb.ai///runs/ +``` + +Where values enclosed in angle brackets (`< >`) are placeholders for the actual values of the team name, project name, and [run ID](/models/runs/run-identifiers#run-id). + +After you select a run, explore its properties by navigating through the tabs: [Overview](/models/runs/view-logged-runs#overview), [Logs](/models/runs/view-logged-runs#logs), [Files](/models/runs/view-logged-runs#files), [Code](/models/runs/view-logged-runs#code), and [Artifacts](/models/runs/view-logged-runs#artifacts). + + + + +To view a run locally in your terminal using the `wandb beta leet` terminal UI: + +1. If you started the run locally from a script, navigate to the directory where you ran your code. It contains a `wandb/` directory with a subdirectory per run and a `latest-run/` symbolic link. Each run directory contains a transaction log named in the format `run-.wandb`. + + If you did not start the run locally but downloaded a `.wandb` transaction log file instead, make a note of its location. +2. Start `wandb beta leet` using one of these commands: + + ```bash + # View the latest run, stored in ./wandb/latest-run/ + wandb beta leet + + # Specify a run directory + wandb beta leet ./wandb/run-20250813_124246-n67z9ude + + # Specify a .wandb file + wandb beta leet ./wandb/run-20250813_124246-n67z9ude/run-n67z9ude.wandb + ``` + +![Screenshot demonstrating viewing a run in the LEET terminal UI](/images/leet/leet-tui.png) + +LEET displays a three-panel interface: + +- **Left sidebar**: Run overview with environment variables, configuration, and summary statistics +- **Center panel**: Metrics grid with Braille-style line charts showing your logged metrics +- **Right sidebar**: System metrics including GPU/CPU/RAM utilization + +Get started with these keyboard shortcuts: + +- `h` or `?` - View all keyboard shortcuts +- `/` - Filter metrics by pattern +- `[` / `]` - Toggle left/right sidebars +- `n` / `N` - Navigate between metric pages +- `q` / `CMD+C` - Quit + +See the [`wandb beta leet`](/models/ref/cli/wandb-beta/wandb-beta-leet) reference for more details. + + + + + +### Overview +Use the **Overview** tab to learn about specific run information in a project, such as: + +* **Author**: The W&B entity that creates the run. +* **Command**: The command that initializes the run. +* **Description**: A description of the run that you provided. This field is empty if you do not specify a description when you create the run. You can add a description to a run with the W&B App or programmatically with the Python SDK. +* **Tracked Hours**: The amount of time the run is actively computing or logging data, excluding any pauses or waiting periods. This metric helps you understand the actual computational time spent on your run. +* **Runtime**: Measures the total time from the start to the end of the run. It's the wall-clock time for the run, including any time where the run is paused or waiting for resources. This metric provides the complete elapsed time for your run. +* **Git repository**: The git repository associated with the run. You must [enable git](/platform/app/settings-page/user-settings/#personal-github-integration) to view this field. +* **Host name**: Where W&B computes the run. W&B displays the name of your machine if you initialize the run locally on your machine. +* **Name**: The name of the run. +* **OS**: Operating system that initializes the run. +* **Python executable**: The command that starts the run. +* **Python version**: Specifies the Python version that creates the run. +* **Run path**: Identifies the unique run identifier in the form `entity/project/run-ID`. +* **Start time**: The timestamp when you initialize the run. +* **State**: The [state of the run](/models/runs/run-states). +* **System hardware**: The hardware W&B uses to compute the run. +* **Tags**: A list of strings. Tags are useful for organizing related runs together or applying temporary labels like `baseline` or `production`. +* **W&B CLI version**: The W&B CLI version installed on the machine that hosted the run command. +* **Git state**: The most recent git commit SHA of a repository or working directory where the run is initialized. This field is empty if you do not enable Git when you create the run or if the git information is not available. + +W&B stores the following information below the overview section: + +* **Artifact Outputs**: Artifact outputs produced by the run. +* **Config**: List of config parameters saved with [`wandb.Run.config`](/models/track/config/). +* **Summary**: List of summary parameters saved with [`wandb.Run.log()`](/models/track/log/). By default, W&B sets this value to the last value logged. + +View an example project overview [here](https://wandb.ai/stacey/deep-drive/overview). + + +### Logs +The **Log** tab shows output printed on the command line such as the standard output (`stdout`) and standard error (`stderr`). + + + Run logs tab + + +Click the **Download** button in the upper right hand corner to download the log file. + +View an example logs tab [here](https://app.wandb.ai/stacey/deep-drive/runs/pr0os44x/logs). + +### Files +Use the **Files** tab to view files associated with a specific run such as model checkpoints, validation set examples, and more + + + Run files tab + + +View an example files tab [here](https://app.wandb.ai/stacey/deep-drive/runs/pr0os44x/files/media/images). + +### Code +The **Code** tab displays the code files associated with a specific run. This includes the main script that was executed as well as any additional code files that were part of the run's environment. + + +### Artifacts + +The **Artifacts** tab lists the input and output [artifacts](/models/artifacts/) for the specified run. + + + Run artifacts tab + \ No newline at end of file diff --git a/models/support.mdx b/models/support.mdx index 0d48647a9c..9c5ebd88a0 100644 --- a/models/support.mdx +++ b/models/support.mdx @@ -1546,7 +1546,7 @@ with wandb.init(reinit="create_new") as tracking_run: tracking_run.log({"x": x}) ``` -See [Multiple runs per process](/models/runs/multiple-runs-per-process) +See [Multiple runs per process](/models/runs/initialize-run) for more information about `reinit="create_new"`, including caveats about W&B integrations. diff --git a/models/support/launch_multiple_runs_one_script.mdx b/models/support/launch_multiple_runs_one_script.mdx index 40bf13b90a..13d6e039ec 100644 --- a/models/support/launch_multiple_runs_one_script.mdx +++ b/models/support/launch_multiple_runs_one_script.mdx @@ -55,6 +55,6 @@ with wandb.init(reinit="create_new") as tracking_run: tracking_run.log({"x": x}) ``` -See [Multiple runs per process](/models/runs/multiple-runs-per-process/) +See [Multiple runs per process](/models/runs/initialize-run/) for more information about `reinit="create_new"`, including caveats about W&B integrations. diff --git a/models/track/project-page.mdx b/models/track/project-page.mdx index 5470f6c3bd..fafcada404 100644 --- a/models/track/project-page.mdx +++ b/models/track/project-page.mdx @@ -174,11 +174,7 @@ From left to right they are: column name, operator, and the filter value. -Click the **Group** button above the table to group all rows by the value in a particular column. - - - Error distribution analysis - +Click the **Group** button above the table to group all rows by the value in a particular column. For more information on how to group runs, see [Organize runs into groups](/models/runs/grouping). @@ -327,7 +323,7 @@ Refer to the [`wandb.init()` API reference](/models/ref/python/functions/init/#e Add a star to a project to mark that project as important. Projects that you and your team mark as important with stars appear at the top of your organization's homepage. -For example, the proceeding image shows two projects that are marked as important, the `zoo_experiment` and `registry_demo`. Both projects appear within the top of the organization's homepage within the **Starred projects** section. +For example, the following image shows two projects that are marked as important, the `zoo_experiment` and `registry_demo`. Both projects appear within the top of the organization's homepage within the **Starred projects** section. Starred projects section diff --git a/support-tags.mdx b/support-tags.mdx index 1cdee7769e..5ff21424d7 100644 --- a/support-tags.mdx +++ b/support-tags.mdx @@ -1519,7 +1519,7 @@ with wandb.init(reinit="create_new") as tracking_run: tracking_run.log({"x": x}) ``` -See [Multiple runs per process](/models/runs/multiple-runs-per-process) +See [Multiple runs per process](/models/runs/initialize-run) for more information about `reinit="create_new"`, including caveats about W&B integrations.