Reset history

.gitattributes

@@ -0,0 +1,35 @@
+*.7z filter=lfs diff=lfs merge=lfs -text
+*.arrow filter=lfs diff=lfs merge=lfs -text
+*.bin filter=lfs diff=lfs merge=lfs -text
+*.bz2 filter=lfs diff=lfs merge=lfs -text
+*.ckpt filter=lfs diff=lfs merge=lfs -text
+*.ftz filter=lfs diff=lfs merge=lfs -text
+*.gz filter=lfs diff=lfs merge=lfs -text
+*.h5 filter=lfs diff=lfs merge=lfs -text
+*.joblib filter=lfs diff=lfs merge=lfs -text
+*.lfs.* filter=lfs diff=lfs merge=lfs -text
+*.mlmodel filter=lfs diff=lfs merge=lfs -text
+*.model filter=lfs diff=lfs merge=lfs -text
+*.msgpack filter=lfs diff=lfs merge=lfs -text
+*.npy filter=lfs diff=lfs merge=lfs -text
+*.npz filter=lfs diff=lfs merge=lfs -text
+*.onnx filter=lfs diff=lfs merge=lfs -text
+*.ot filter=lfs diff=lfs merge=lfs -text
+*.parquet filter=lfs diff=lfs merge=lfs -text
+*.pb filter=lfs diff=lfs merge=lfs -text
+*.pickle filter=lfs diff=lfs merge=lfs -text
+*.pkl filter=lfs diff=lfs merge=lfs -text
+*.pt filter=lfs diff=lfs merge=lfs -text
+*.pth filter=lfs diff=lfs merge=lfs -text
+*.rar filter=lfs diff=lfs merge=lfs -text
+*.safetensors filter=lfs diff=lfs merge=lfs -text
+saved_model/**/* filter=lfs diff=lfs merge=lfs -text
+*.tar.* filter=lfs diff=lfs merge=lfs -text
+*.tar filter=lfs diff=lfs merge=lfs -text
+*.tflite filter=lfs diff=lfs merge=lfs -text
+*.tgz filter=lfs diff=lfs merge=lfs -text
+*.wasm filter=lfs diff=lfs merge=lfs -text
+*.xz filter=lfs diff=lfs merge=lfs -text
+*.zip filter=lfs diff=lfs merge=lfs -text
+*.zst filter=lfs diff=lfs merge=lfs -text
+*tfevents* filter=lfs diff=lfs merge=lfs -text

README.md

@@ -0,0 +1,12 @@
+---
+title: PLUS Lab GPUs
+emoji: ➕
+colorFrom: blue
+colorTo: purple
+sdk: gradio
+sdk_version: 5.13.1
+app_file: app.py
+pinned: false
+---
+
+Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,89 @@
+import gradio as gr
+import pandas as pd
+import datetime
+import plotly.express as px
+
+def split_multi_users(dfs):
+    df = dfs.copy()
+    df["usernames"] = df["username"].apply(lambda x: x.split(", "))
+    df["count"] = 1
+    new_df = []
+    for row in df.to_dict(orient="records"):
+        gpu_users_num = len(row["usernames"])
+        for username in row["usernames"]:
+            new_row = row.copy()
+            new_row["count"] = 1 / gpu_users_num
+            new_row["username"] = username
+            new_df.append(new_row)
+    df = pd.DataFrame(new_df)
+    return df
+
+def plot_now():
+    # dfs = pd.read_csv("hf://spaces/pluslab/PLUS_Lab_GPUs/gpus.csv")
+    dfs = pd.read_csv("hf://datasets/pluslab/PLUS_Lab_GPUs_Data/gpus.csv")
+    dfs = dfs.drop(columns=["Unnamed: 0"])
+    dfs = dfs.fillna("FREE")
+    dfs_plot = split_multi_users(dfs)
+    fig = px.bar(
+        dfs_plot, x="count", y="server", color="username",
+        title=f"Last Updated {min(dfs['timestamp'])}",
+        color_discrete_map={
+            "FREE": "black",
+        },
+        text=dfs_plot['username'].astype(str) + "<br>" + dfs_plot['device'].astype(str),
+    )
+    fig.update_layout(
+        yaxis={'categoryorder': 'array', 'categoryarray': dfs_plot["server"].unique()[::-1]},
+        barcornerradius=50,
+    )
+    fig.update_traces(textposition='inside', insidetextanchor='middle')
+    print(dfs_plot)
+    return fig, dfs
+
+def plot_history(sample=True, sampling_interval_minutes=180):
+    dfh = pd.read_pickle("hf://spaces/pluslab/PLUS_Lab_GPUs/history.pkl.gz", )
+    dfh = dfh.fillna("FREE")
+    dfh = split_multi_users(dfh)
+    dfh = dfh[["polling_timestamp", "username", "count"]]
+    dfh = dfh.groupby(["polling_timestamp", "username"]).sum()
+    dfh = dfh.reset_index()
+    dfh = dfh.sort_values(by=["polling_timestamp", "count"], ascending=False)
+    if sample:
+        unique_timestamps = dfh["polling_timestamp"].unique()
+        sampled_timestamps = [unique_timestamps[0]]
+        for i, t in enumerate(unique_timestamps[1:]):
+            diff = sampled_timestamps[-1] - t
+            if diff > datetime.timedelta(minutes=sampling_interval_minutes):
+                sampled_timestamps.append(t)
+        dfh = dfh[dfh["polling_timestamp"].isin(sampled_timestamps)]
+    fig = px.area(dfh, x="polling_timestamp", y="count", color='username', color_discrete_map={"FREE": "black",}, markers=True, line_shape='spline',)
+    return fig, dfh
+
+
+def plot_figs():
+    fig_now, dfn = plot_now()
+    try:
+        fig_history, dfh = plot_history()
+    except Exception as e:
+        print(e)
+        fig_history = None
+        dfh = None
+    return fig_now, dfn, fig_history
+
+demo = gr.Interface(
+    fn=plot_figs,
+    inputs = [
+
+    ],
+    outputs = [
+        gr.Plot(label="GPU Status", elem_classes="plotcss"),
+        gr.Dataframe(label="GPU Status Details"),
+        gr.Plot(label="History", elem_classes="plotcss"),
+    ],
+    live=True,
+    flagging_options=[],
+    css=".plotcss {max-width: 820px !important;}"
+)
+
+if __name__ == "__main__":
+    demo.launch(debug=False)

gpus.csv

@@ -0,0 +1,45 @@
+,timestamp,server,name,device,utilization.gpu [%],utilization.memory [%],temperature.gpu,power.draw [W],username
+0,2025/08/24 20:54:26.539,pluslab-a100.cs.ucla.edu,NVIDIA A100-SXM4-40GB,0,0 %,0 %,29,52.28 W,
+1,2025/08/24 20:54:26.561,pluslab-a100.cs.ucla.edu,NVIDIA A100-SXM4-40GB,1,0 %,0 %,30,52.50 W,
+2,2025/08/24 20:54:26.580,pluslab-a100.cs.ucla.edu,NVIDIA A100-SXM4-40GB,2,0 %,0 %,27,52.61 W,
+3,2025/08/24 20:54:26.596,pluslab-a100.cs.ucla.edu,NVIDIA A100-SXM4-40GB,3,0 %,0 %,28,53.10 W,
+0,2025/08/24 20:54:27.724,pluslab01-a100.cs.ucla.edu,NVIDIA A100 80GB PCIe,0,0 %,0 %,44,47.17 W,
+1,2025/08/24 20:54:27.734,pluslab01-a100.cs.ucla.edu,NVIDIA A100 80GB PCIe,1,99 %,51 %,75,305.19 W,junyizhang
+2,2025/08/24 20:54:27.742,pluslab01-a100.cs.ucla.edu,NVIDIA A100 80GB PCIe,2,0 %,0 %,43,44.71 W,
+3,2025/08/24 20:54:27.750,pluslab01-a100.cs.ucla.edu,NVIDIA A100 80GB PCIe,3,99 %,54 %,73,274.60 W,junyizhang
+4,2025/08/24 20:54:27.767,pluslab01-a100.cs.ucla.edu,NVIDIA A100 80GB PCIe,4,69 %,22 %,57,146.07 W,junyizhang
+5,2025/08/24 20:54:27.774,pluslab01-a100.cs.ucla.edu,NVIDIA A100 80GB PCIe,5,100 %,0 %,46,73.39 W,junyizhang
+6,2025/08/24 20:54:27.779,pluslab01-a100.cs.ucla.edu,NVIDIA A100 80GB PCIe,6,0 %,0 %,44,46.31 W,
+7,2025/08/24 20:54:27.790,pluslab01-a100.cs.ucla.edu,NVIDIA A100 80GB PCIe,7,0 %,0 %,41,43.86 W,
+0,2025/08/24 20:54:28.899,pluslab01-h100.cs.ucla.edu,NVIDIA H100 80GB HBM3,0,79 %,64 %,86,457.36 W,ponienkung
+1,2025/08/24 20:54:28.918,pluslab01-h100.cs.ucla.edu,NVIDIA H100 80GB HBM3,1,78 %,64 %,83,503.67 W,ponienkung
+2,2025/08/24 20:54:28.933,pluslab01-h100.cs.ucla.edu,NVIDIA H100 80GB HBM3,2,80 %,66 %,86,476.30 W,ponienkung
+3,2025/08/24 20:54:28.948,pluslab01-h100.cs.ucla.edu,NVIDIA H100 80GB HBM3,3,79 %,66 %,84,483.46 W,ponienkung
+4,2025/08/24 20:54:28.963,pluslab01-h100.cs.ucla.edu,NVIDIA H100 80GB HBM3,4,100 %,0 %,77,164.29 W,haikang
+5,2025/08/24 20:54:28.977,pluslab01-h100.cs.ucla.edu,NVIDIA H100 80GB HBM3,5,100 %,0 %,72,158.19 W,haikang
+6,2025/08/24 20:54:28.992,pluslab01-h100.cs.ucla.edu,NVIDIA H100 80GB HBM3,6,0 %,0 %,81,216.23 W,haikang
+7,2025/08/24 20:54:29.006,pluslab01-h100.cs.ucla.edu,NVIDIA H100 80GB HBM3,7,100 %,0 %,69,155.87 W,haikang
+0,2025/08/24 20:54:30.416,pluslab-a6000.cs.ucla.edu,NVIDIA RTX A6000,0,0 %,0 %,40,28.65 W,
+1,2025/08/24 20:54:30.421,pluslab-a6000.cs.ucla.edu,NVIDIA RTX A6000,1,0 %,0 %,36,33.35 W,
+2,2025/08/24 20:54:30.425,pluslab-a6000.cs.ucla.edu,NVIDIA RTX A6000,2,0 %,0 %,39,33.23 W,elvindu
+3,2025/08/24 20:54:30.429,pluslab-a6000.cs.ucla.edu,NVIDIA RTX A6000,3,0 %,0 %,35,29.11 W,
+4,2025/08/24 20:54:30.433,pluslab-a6000.cs.ucla.edu,NVIDIA RTX A6000,4,0 %,0 %,32,22.01 W,
+5,2025/08/24 20:54:30.438,pluslab-a6000.cs.ucla.edu,NVIDIA RTX A6000,5,0 %,0 %,31,28.91 W,
+6,2025/08/24 20:54:30.442,pluslab-a6000.cs.ucla.edu,NVIDIA RTX A6000,6,0 %,0 %,32,27.85 W,
+7,2025/08/24 20:54:30.445,pluslab-a6000.cs.ucla.edu,NVIDIA RTX A6000,7,0 %,0 %,29,22.26 W,
+0,2025/08/24 20:54:31.579,pluslab01-a6000.cs.ucla.edu,NVIDIA RTX A6000,0,0 %,0 %,28,22.30 W,
+1,2025/08/24 20:54:31.591,pluslab01-a6000.cs.ucla.edu,NVIDIA RTX A6000,1,0 %,0 %,27,18.46 W,
+2,2025/08/24 20:54:31.603,pluslab01-a6000.cs.ucla.edu,NVIDIA RTX A6000,2,0 %,0 %,27,15.99 W,
+3,2025/08/24 20:54:31.617,pluslab01-a6000.cs.ucla.edu,NVIDIA RTX A6000,3,0 %,0 %,28,23.15 W,
+4,2025/08/24 20:54:31.628,pluslab01-a6000.cs.ucla.edu,NVIDIA RTX A6000,4,0 %,0 %,27,21.75 W,
+5,2025/08/24 20:54:31.643,pluslab01-a6000.cs.ucla.edu,NVIDIA RTX A6000,5,0 %,0 %,26,19.21 W,
+6,2025/08/24 20:54:31.656,pluslab01-a6000.cs.ucla.edu,NVIDIA RTX A6000,6,0 %,0 %,27,20.50 W,
+7,2025/08/24 20:54:31.665,pluslab01-a6000.cs.ucla.edu,NVIDIA RTX A6000,7,0 %,0 %,26,26.94 W,
+0,2025/08/24 20:54:34.185,pluslab02-a6000.cs.ucla.edu,NVIDIA RTX A6000,0,0 %,0 %,28,26.18 W,cfyang
+1,2025/08/24 20:54:34.223,pluslab02-a6000.cs.ucla.edu,NVIDIA RTX A6000,1,0 %,0 %,26,30.27 W,cfyang
+2,2025/08/24 20:54:34.250,pluslab02-a6000.cs.ucla.edu,NVIDIA RTX A6000,2,85 %,72 %,68,201.33 W,hzhang19
+3,2025/08/24 20:54:34.279,pluslab02-a6000.cs.ucla.edu,NVIDIA RTX A6000,3,1 %,0 %,62,127.91 W,hzhang19
+4,2025/08/24 20:54:34.298,pluslab02-a6000.cs.ucla.edu,NVIDIA RTX A6000,4,75 %,63 %,70,203.11 W,hzhang19
+5,2025/08/24 20:54:34.319,pluslab02-a6000.cs.ucla.edu,NVIDIA RTX A6000,5,52 %,26 %,56,133.41 W,hzhang19
+6,2025/08/24 20:54:34.360,pluslab02-a6000.cs.ucla.edu,NVIDIA RTX A6000,6,100 %,83 %,69,215.74 W,hzhang19
+7,2025/08/24 20:54:34.393,pluslab02-a6000.cs.ucla.edu,NVIDIA RTX A6000,7,85 %,72 %,64,187.19 W,hzhang19

requirements.txt

@@ -0,0 +1,2 @@
+plotly==5.24.1
+pandas==2.2.3