|
1 |
| -(cli_commands_reference)= |
| 1 | +(cli_inference)= |
2 | 2 |
|
3 |
| -# CLI Commands Reference |
| 3 | +# Inference CLI |
4 | 4 |
|
5 |
| -Complete reference for all SageMaker HyperPod CLI commands, options, and parameters. |
| 5 | +Complete reference for SageMaker HyperPod PyTorch training job parameters and configuration options. |
6 | 6 |
|
7 |
| -## Table of Contents |
8 |
| - |
9 |
| -- [Create Commands](#create-commands) |
10 |
| -- [Cluster Commands](#cluster-commands) |
11 |
| -- [Training Job Management Commands](#training-job-management-commands) |
12 |
| -- [Inference Endpoint Management Commands](#inference-endpoint-management-commands) |
13 |
| - |
14 |
| -## Create Commands |
15 |
| - |
16 |
| -The `hyp create` command family is used to create various resources in your HyperPod cluster. |
17 |
| - |
18 |
| -### hyp create hyp-pytorch-job |
19 |
| - |
20 |
| -Create distributed PyTorch training jobs. |
21 |
| - |
22 |
| -#### Syntax |
23 |
| - |
24 |
| -```bash |
25 |
| -hyp create hyp-pytorch-job [OPTIONS] |
26 |
| -``` |
27 |
| - |
28 |
| -#### Required Parameters |
29 |
| - |
30 |
| -- `--job-name TEXT`: Unique name for the training job (minimum 1 character) |
31 |
| -- `--image TEXT`: Docker image URI containing your training code |
32 |
| - |
33 |
| -#### Optional Parameters |
34 |
| - |
35 |
| -- `--namespace TEXT`: Kubernetes namespace |
36 |
| -- `--command ARRAY`: Command to run in the container (array of strings) |
37 |
| -- `--args ARRAY`: Arguments for the entry script (array of strings) |
38 |
| -- `--environment OBJECT`: Environment variables as key-value pairs |
39 |
| -- `--pull-policy TEXT`: Image pull policy (Always, Never, IfNotPresent) |
40 |
| -- `--instance-type TEXT`: Instance type for training |
41 |
| -- `--node-count INTEGER`: Number of nodes (minimum: 1) |
42 |
| -- `--tasks-per-node INTEGER`: Number of tasks per node (minimum: 1) |
43 |
| -- `--label-selector OBJECT`: Node label selector as key-value pairs |
44 |
| -- `--deep-health-check-passed-nodes-only BOOLEAN`: Schedule pods only on nodes that passed deep health check |
45 |
| -- `--scheduler-type TEXT`: Scheduler type |
46 |
| -- `--queue-name TEXT`: Queue name for job scheduling |
47 |
| -- `--priority TEXT`: Priority class for job scheduling |
48 |
| -- `--max-retry INTEGER`: Maximum number of job retries (minimum: 0) |
49 |
| -- `--volumes ARRAY`: List of volumes to mount |
50 |
| -- `--persistent-volume-claims ARRAY`: List of persistent volume claims |
51 |
| -- `--service-account-name TEXT`: Service account name |
52 |
| - |
53 |
| -### hyp create hyp-jumpstart-endpoint |
| 7 | +## hyp create hyp-jumpstart-endpoint |
54 | 8 |
|
55 | 9 | Deploy pre-trained models from SageMaker JumpStart.
|
56 | 10 |
|
@@ -121,141 +75,6 @@ hyp create hyp-custom-endpoint [OPTIONS]
|
121 | 75 | - `--use-cached-metrics BOOLEAN`: Enable caching of metric values (default: true)
|
122 | 76 | - `--invocation-endpoint TEXT`: Invocation endpoint path (default: "invocations")
|
123 | 77 |
|
124 |
| -## Cluster Commands |
125 |
| - |
126 |
| -Commands for managing cluster connections and contexts. |
127 |
| - |
128 |
| -### hyp list-cluster |
129 |
| - |
130 |
| -List all available SageMaker HyperPod clusters in your account. |
131 |
| - |
132 |
| -#### Syntax |
133 |
| - |
134 |
| -```bash |
135 |
| -hyp list-cluster [OPTIONS] |
136 |
| -``` |
137 |
| - |
138 |
| -#### Optional Parameters |
139 |
| - |
140 |
| -- `--region TEXT`: AWS region to list clusters from |
141 |
| -- `--namespace TEXT`: Kubernetes namespace |
142 |
| -- `--output TEXT`: Output format (json, table) |
143 |
| - |
144 |
| -### hyp set-cluster-context |
145 |
| - |
146 |
| -Configure your local kubectl environment to interact with a specific SageMaker HyperPod cluster. |
147 |
| - |
148 |
| -#### Syntax |
149 |
| - |
150 |
| -```bash |
151 |
| -hyp set-cluster-context [OPTIONS] |
152 |
| -``` |
153 |
| - |
154 |
| -#### Required Parameters |
155 |
| - |
156 |
| -- `--cluster-name TEXT`: Name of the cluster to connect to |
157 |
| - |
158 |
| -### hyp get-cluster-context |
159 |
| - |
160 |
| -View information about the currently configured cluster context. |
161 |
| - |
162 |
| -#### Syntax |
163 |
| - |
164 |
| -```bash |
165 |
| -hyp get-cluster-context |
166 |
| -``` |
167 |
| - |
168 |
| -## Training Job Management Commands |
169 |
| - |
170 |
| -Commands for managing PyTorch training jobs. |
171 |
| - |
172 |
| -### hyp list hyp-pytorch-job |
173 |
| - |
174 |
| -List all HyperPod PyTorch jobs in a namespace. |
175 |
| - |
176 |
| -#### Syntax |
177 |
| - |
178 |
| -```bash |
179 |
| -hyp list hyp-pytorch-job [OPTIONS] |
180 |
| -``` |
181 |
| - |
182 |
| -#### Optional Parameters |
183 |
| - |
184 |
| -- `--namespace, -n TEXT`: Namespace to list jobs from (default: "default") |
185 |
| - |
186 |
| -### hyp describe hyp-pytorch-job |
187 |
| - |
188 |
| -Describe a specific HyperPod PyTorch job. |
189 |
| - |
190 |
| -#### Syntax |
191 |
| - |
192 |
| -```bash |
193 |
| -hyp describe hyp-pytorch-job [OPTIONS] |
194 |
| -``` |
195 |
| - |
196 |
| -#### Required Parameters |
197 |
| - |
198 |
| -- `--job-name TEXT`: Name of the job to describe |
199 |
| - |
200 |
| -#### Optional Parameters |
201 |
| - |
202 |
| -- `--namespace, -n TEXT`: Namespace of the job (default: "default") |
203 |
| - |
204 |
| -### hyp delete hyp-pytorch-job |
205 |
| - |
206 |
| -Delete a HyperPod PyTorch job. |
207 |
| - |
208 |
| -#### Syntax |
209 |
| - |
210 |
| -```bash |
211 |
| -hyp delete hyp-pytorch-job [OPTIONS] |
212 |
| -``` |
213 |
| - |
214 |
| -#### Required Parameters |
215 |
| - |
216 |
| -- `--job-name TEXT`: Name of the job to delete |
217 |
| - |
218 |
| -#### Optional Parameters |
219 |
| - |
220 |
| -- `--namespace, -n TEXT`: Namespace of the job (default: "default") |
221 |
| - |
222 |
| -### hyp list-pods hyp-pytorch-job |
223 |
| - |
224 |
| -List all pods associated with a PyTorch job. |
225 |
| - |
226 |
| -#### Syntax |
227 |
| - |
228 |
| -```bash |
229 |
| -hyp list-pods hyp-pytorch-job [OPTIONS] |
230 |
| -``` |
231 |
| - |
232 |
| -#### Required Parameters |
233 |
| - |
234 |
| -- `--job-name TEXT`: Name of the job to list pods for |
235 |
| - |
236 |
| -#### Optional Parameters |
237 |
| - |
238 |
| -- `--namespace, -n TEXT`: Namespace of the job (default: "default") |
239 |
| - |
240 |
| -### hyp get-logs hyp-pytorch-job |
241 |
| - |
242 |
| -Get logs from a specific pod in a PyTorch job. |
243 |
| - |
244 |
| -#### Syntax |
245 |
| - |
246 |
| -```bash |
247 |
| -hyp get-logs hyp-pytorch-job [OPTIONS] |
248 |
| -``` |
249 |
| - |
250 |
| -#### Required Parameters |
251 |
| - |
252 |
| -- `--job-name TEXT`: Name of the job |
253 |
| -- `--pod-name TEXT`: Name of the pod to get logs from |
254 |
| - |
255 |
| -#### Optional Parameters |
256 |
| - |
257 |
| -- `--namespace, -n TEXT`: Namespace of the job (default: "default") |
258 |
| - |
259 | 78 | ## Inference Endpoint Management Commands
|
260 | 79 |
|
261 | 80 | Commands for managing inference endpoints.
|
@@ -501,4 +320,4 @@ hyp get-operator-logs hyp-custom-endpoint [OPTIONS]
|
501 | 320 | | Parameter | Type | Description | Default |
|
502 | 321 | |-----------|------|-------------|---------|
|
503 | 322 | | `--namespace` | TEXT | Kubernetes namespace | Current context |
|
504 |
| -| `--help` | FLAG | Show command help | - | |
| 323 | +| `--help` | FLAG | Show command help | - | |
0 commit comments