Skip to content

Commit 63d6a98

Browse files
oindrillacMichaelClifford
authored andcommitted
Added some description to mlflow example
1 parent 23a5404 commit 63d6a98

File tree

1 file changed

+24
-3
lines changed

1 file changed

+24
-3
lines changed

MLflow/MLflow_example.ipynb

Lines changed: 24 additions & 3 deletions
Original file line numberDiff line numberDiff line change
@@ -11,7 +11,7 @@
1111
},
1212
{
1313
"cell_type": "code",
14-
"execution_count": 1,
14+
"execution_count": 3,
1515
"metadata": {},
1616
"outputs": [],
1717
"source": [
@@ -30,14 +30,35 @@
3030
"## Set up connection "
3131
]
3232
},
33+
{
34+
"cell_type": "markdown",
35+
"metadata": {},
36+
"source": [
37+
"ML Flow helps track experiments and runs within a particular data science project. As per MLFlow tracking conventions :\n",
38+
"\n",
39+
"- An **experiment** corresponds to a particular version of code or model. Each experiment will be tracked as different commits within git.\n",
40+
"\n",
41+
"- A **run** corresponds to a different hyper parameter setting or different feature engineering within same experiment. Runs are tracked within each ML Flow experiment\n",
42+
"\n",
43+
"1) Set up the ML Flow client and point it to our current MLFlow deployment : http://mlflow-server-route-aiops-prod-prometheus-scrape.cloud.paas.psi.redhat.com/#/\n",
44+
"\n",
45+
"2) Start a new MLflow run, setting it as the active run under which metrics and parameters will be logged\n",
46+
"\n",
47+
"3) Set an experiment for tracking a particular project \n",
48+
"\n",
49+
"4) Each run corresponds to a training cycle. Associate run name to version of code (experiment)\n",
50+
"\n",
51+
"5) Start a run"
52+
]
53+
},
3354
{
3455
"cell_type": "markdown",
3556
"metadata": {},
3657
"source": [
3758
"### Connection configurations set by the user:\n",
3859
"\n",
3960
"* **MLFLOW_URI**: route to deployment\n",
40-
"* **EXPERIMENT_NAME**: Name for entire set of experiments you want to compare. Think of this as the folder this run will live in. \n",
61+
"* **EXPERIMENT_NAME**: Name for entire set of experiments you want to compare. Think of this as the folder this run will live in or a particular version of the code or model.\n",
4162
"* **RUN_NAME**: custom name for this specific run"
4263
]
4364
},
@@ -560,7 +581,7 @@
560581
"name": "python",
561582
"nbconvert_exporter": "python",
562583
"pygments_lexer": "ipython3",
563-
"version": "3.6.3"
584+
"version": "3.7.4"
564585
}
565586
},
566587
"nbformat": 4,

0 commit comments

Comments
 (0)