Use CircleCI version 2.1 at the top of your .circleci/config.yml file.
1
version: 2.1
Add the orbs
stanza below your version, invoking the orb:
1
2
orbs:
aws-sagemaker: circleci/aws-sagemaker@1.0.5
Use aws-sagemaker
elements in your existing workflows and jobs.
This example defines a CircleCI workflow that uses the aws-sagemaker orb to: - Create a SageMaker model - Create an endpoint configuration for the model - Deploy the model to an endpoint It shows a typical machine learning workflow using SageMaker with model releases being tracked on CircleCI.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
version: '2.1'
orbs:
aws-sagemaker: circleci/aws-sagemaker@1.0
workflows:
ai-workflow:
jobs:
- aws-sagemaker/create_model:
bucket: replace_with_s3_bucket_name
model_name: replace_with_model_name
name: create-model
pipeline_id: << pipeline.id >>
region_name: replace_with_region_name
- aws-sagemaker/create_endpoint_configuration:
bucket: replace_with_s3_bucket_name
model_name: replace_with_model_name
name: create-endpoint-configuration
pipeline_id: << pipeline.id >>
region_name: replace_with_region_name
requires:
- create-model
- aws-sagemaker/deploy_endpoint:
bucket: replace_with_s3_bucket_name
model_desc: replace_with_description
model_name: replace_with_model_name
name: deploy-endpoint
pipeline_id: <<pipeline.id>>
project_id: replace_with_project_id
region_name: replace_with_region_name
requires:
- create-endpoint-configuration
Deploys a model to SageMaker with an existing endpoint configuration.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
version: '2.1'
orbs:
aws-sagemaker: circleci/aws-sagemaker@1.0
workflows:
ai-workflow:
jobs:
- aws-sagemaker/deploy_endpoint:
bucket: replace_with_s3_bucket_name
endpoint_config_name: replace_with_existing_endpoint_config_name
endpoint_name: replace_with_endpoint_name
model_desc: replace_with_description
model_name: replace_with_model_name
name: deploy-endpoint
pipeline_id: <<pipeline.id>>
project_id: replace_with_project_id
region_name: replace_with_region_name
This job creates an create an endpoint configuration.
PARAMETER | DESCRIPTION | REQUIRED | DEFAULT | TYPE |
---|---|---|---|---|
bucket | S3 bucket for the model | Yes | - | string |
circle_pipeline_id | CircleCI Pipeline ID | Yes | - | string |
circle_project_id | CircleCI project ID | Yes | - | string |
endpoint_instance_count | Number of instances to run endpoint inference | No | 1 | integer |
endpoint_instance_type | EC2 instance type to run endpoint inference | No | ml.t2.medium | string |
model_name | Model name | Yes | - | string |
region_name | AWS region name | Yes | - | string |
This job creates sagemaker model given a model package available in registry.
PARAMETER | DESCRIPTION | REQUIRED | DEFAULT | TYPE |
---|---|---|---|---|
bucket | S3 bucket for the model | Yes | - | string |
circle_pipeline_id | CircleCI Pipeline ID | Yes | - | string |
model_name | Model name | Yes | - | string |
region_name | AWS region name | Yes | - | string |
This job deploys inference endpoint by create or update
PARAMETER | DESCRIPTION | REQUIRED | DEFAULT | TYPE |
---|---|---|---|---|
bucket | S3 bucket for the model | Yes | - | string |
circle_pipeline_id | CircleCI Pipeline ID | Yes | - | string |
circle_project_id | CircleCI project ID | Yes | - | string |
enable_restore_version | Enable restoring releases on circleci | No | 'true' | string |
enable_scaling_component | Enable scaling release components on circleci | No | 'true' | string |
endpoint_config_name | Existing endpoint config name | No | '' | string |
endpoint_name | Existing endpoint name | No | '' | string |
model_desc | sagemaker model desription | Yes | - | string |
model_name | Model name | Yes | - | string |
region_name | AWS region name | Yes | - | string |
This command creates an create an endpoint configuration.
PARAMETER | DESCRIPTION | REQUIRED | DEFAULT | TYPE |
---|---|---|---|---|
bucket | S3 bucket for the model | Yes | - | string |
circle_pipeline_id | CircleCI pipeline ID | Yes | - | string |
circle_project_id | CircleCI project ID | Yes | - | string |
endpoint_instance_count | Number of instances to run endpoint inference | No | 1 | integer |
endpoint_instance_type | EC2 instance type to run endpoint inference | No | ml.t2.medium | string |
model_name | Model name | Yes | - | string |
region_name | AWS region name | No | us-east-1 | string |
This command creates a sagemaker model with latest model package.
PARAMETER | DESCRIPTION | REQUIRED | DEFAULT | TYPE |
---|---|---|---|---|
bucket | S3 bucket for the model | Yes | - | string |
circle_pipeline_id | CircleCI pipeline ID | Yes | - | string |
model_name | Model name | Yes | - | string |
region_name | AWS region name | No | us-east-1 | string |
This command deploys inference endpoint by create or update
PARAMETER | DESCRIPTION | REQUIRED | DEFAULT | TYPE |
---|---|---|---|---|
bucket | S3 bucket for the model | Yes | - | string |
circle_pipeline_id | CircleCI pipeline ID | Yes | - | string |
circle_project_id | CircleCI project ID | Yes | - | string |
enable_restore_version | Enable restoring releases on circleci | No | 'true' | string |
enable_scaling_component | Enable scaling release components on circleci | No | 'true' | string |
endpoint_config_name | Existing endpoint config name | No | '' | string |
endpoint_name | Existing endpoint name | No | '' | string |
model_desc | sagemaker model desription | Yes | - | string |
model_name | Model name | Yes | - | string |
region_name | AWS region name | Yes | - | string |
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
# This code is licensed from CircleCI to the user under the MIT license.
# See here for details: https://circleci.com/developer/ja/orbs/licensing
version: 2.1
description: |
Orb for deploying models to Amazon SageMaker and tracking releases in CircleCI. (OPEN PREVIEW - send feedback at sagemaker-integration-feedback@circleci.com)
display:
home_url: https://github.com/CircleCI-Public/aws-sagemaker-orb
source_url: https://github.com/CircleCI-Public/aws-sagemaker-orb
commands:
create_endpoint_configuration:
description: |
This command creates an create an endpoint configuration.
parameters:
bucket:
description: S3 bucket for the model
type: string
circle_pipeline_id:
description: CircleCI pipeline ID
type: string
circle_project_id:
description: CircleCI project ID
type: string
endpoint_instance_count:
default: 1
description: Number of instances to run endpoint inference
type: integer
endpoint_instance_type:
default: ml.t2.medium
description: EC2 instance type to run endpoint inference
type: string
model_name:
description: Model name
type: string
region_name:
default: us-east-1
description: AWS region name
type: string
steps:
- run:
command: |
cci-sagemaker sagemaker create-endpoint-configuration
description: create an endpoint configuration
environment:
CIRCLE_PIPELINE_ID: << parameters.circle_pipeline_id >>
CIRCLE_PROJECT_ID: << parameters.circle_project_id >>
ENDPOINT_INSTANCE_COUNT: << parameters.endpoint_instance_count >>
ENDPOINT_INSTANCE_TYPE: << parameters.endpoint_instance_type >>
MODEL_NAME: << parameters.model_name >>
REGION_NAME: << parameters.region_name >>
S3_BUCKET_NAME: << parameters.bucket >>
name: Create Endpoint Configuration
create_model:
description: |
This command creates a sagemaker model with latest model package.
parameters:
bucket:
description: S3 bucket for the model
type: string
circle_pipeline_id:
description: CircleCI pipeline ID
type: string
model_name:
description: Model name
type: string
region_name:
default: us-east-1
description: AWS region name
type: string
steps:
- run:
command: |
cci-sagemaker sagemaker create-model
description: create a model for inference
environment:
CIRCLE_PIPELINE_ID: << parameters.circle_pipeline_id >>
MODEL_NAME: << parameters.model_name >>
REGION_NAME: << parameters.region_name >>
S3_BUCKET_NAME: << parameters.bucket >>
name: Create a model
deploy_endpoint:
description: |
This command deploys inference endpoint by create or update
parameters:
bucket:
description: S3 bucket for the model
type: string
circle_pipeline_id:
description: CircleCI pipeline ID
type: string
circle_project_id:
description: CircleCI project ID
type: string
enable_restore_version:
default: "true"
description: Enable restoring releases on circleci
type: string
enable_scaling_component:
default: "true"
description: Enable scaling release components on circleci
type: string
endpoint_config_name:
default: ""
description: Existing endpoint config name
type: string
endpoint_name:
default: ""
description: Existing endpoint name
type: string
model_desc:
description: sagemaker model desription
type: string
model_name:
description: Model name
type: string
region_name:
description: AWS region name
type: string
steps:
- run:
command: |
cci-sagemaker sagemaker upsert-endpoint
description: create or update endpoint with new model
environment:
CIRCLE_PIPELINE_ID: << parameters.circle_pipeline_id >>
CIRCLE_PROJECT_ID: << parameters.circle_project_id >>
ENDPOINT_CONFIGURATION_NAME: << parameters.endpoint_config_name >>
ENDPOINT_NAME: << parameters.endpoint_name >>
MODEL_DESC: << parameters.model_desc >>
MODEL_NAME: << parameters.model_name >>
REGION_NAME: << parameters.region_name >>
RESTORE_VERSION_ENABLED: << parameters.enable_restore_version >>
S3_BUCKET_NAME: << parameters.bucket >>
SCALE_COMPONENT_ENABLED: << parameters.enable_scaling_component >>
name: Create or Update endpoint
setup:
description: |
This command downloads the sagemaker cli binary.
steps:
- run:
command: |
#!/bin/sh
# Hi! Based off the slack-orb-go main.sh. If you plan to have osx & windows
# support go look at that file and bring it back in.
# Determine the http client to use
# Returns 1 if no HTTP client is found
determine_http_client() {
if command -v curl >/dev/null 2>&1; then
HTTP_CLIENT=curl
elif command -v wget >/dev/null 2>&1; then
HTTP_CLIENT=wget
else
return 1
fi
}
# Download a binary file
# $1: The path to save the file to
# $2: The URL to download the file from
# $3: The HTTP client to use (curl or wget)
download_binary() {
if [ "$3" = "curl" ]; then
set -x
curl --fail --retry 3 -L -o "$1" "$2"
set +x
elif [ "$3" = "wget" ]; then
set -x
wget --tries=3 --timeout=10 --quiet -O "$1" "$2"
set +x
else
return 1
fi
}
detect_os() {
detected_platform="$(uname -s | tr '[:upper:]' '[:lower:]')"
case "$detected_platform" in
linux*) PLATFORM=linux ;;
# darwin*) PLATFORM=darwin ;;
# msys* | cygwin*) PLATFORM=windows ;;
*) return 1 ;;
esac
}
detect_arch() {
detected_arch="$(uname -m)"
case "$detected_arch" in
x86_64 | amd64) ARCH=amd64 ;;
i386 | i486 | i586 | i686) ARCH=386 ;;
arm64 | aarch64) ARCH=arm64 ;;
arm*) ARCH=arm ;;
*) return 1 ;;
esac
}
# Confirm we have unzip available
# Returns 1 if unzip not found
detect_unzip() {
if command -v unzip >/dev/null 2>&1; then
return 0
fi
return 1
}
# Print a warning message
# $1: The warning message to print
print_warn() {
yellow="\033[1;33m"
normal="\033[0m"
printf "${yellow}%s${normal}\n" "$1"
}
# Print a success message
# $1: The success message to print
print_success() {
green="\033[0;32m"
normal="\033[0m"
printf "${green}%s${normal}\n" "$1"
}
# Print an error message
# $1: The error message to print
print_error() {
red="\033[0;31m"
normal="\033[0m"
printf "${red}%s${normal}\n" "$1"
}
print_warn "This is an experimental version of the Sagemaker Orb."
print_warn "Thank you for trying it out and please provide feedback to us at https://github.com/CircleCI-Public/sagemaker-orb-go/issues"
if ! detect_os; then
print_error "Unsupported operating system: $(uname -s)."
exit 1
fi
printf '%s\n' "Operating system: $PLATFORM."
if ! detect_arch; then
print_error "Unsupported architecture: $(uname -m)."
exit 1
fi
printf '%s\n' "Architecture: $ARCH."
if ! detect_unzip; then
print_error "Unzip is required to download the Sagemaker Orb binary."
exit 1
fi
base_dir="$(printf "%s" "$CIRCLE_WORKING_DIRECTORY" | sed "s|~|$HOME|")"
orb_bin_dir="${base_dir}/.circleci/orbs/circleci/sagemaker/${PLATFORM}/${ARCH}"
org="circleci"
repo_name="cci-sagemaker"
# binary="${orb_bin_dir}/${repo_name}"
# TODO: Make the version configurable via parameter
# Don't forget the v!
binary_version="v0.0.14"
basic_name="cci-sagemaker"
binary_name="${basic_name}-${binary_version}-${PLATFORM}-${ARCH}"
binary_zip="${orb_bin_dir}/${binary_name}.zip"
# Where to move the binary
path_destination="$HOME/bin"
# Slack orb seems to put this outside the script as a parameter
# https://github.com/CircleCI-Public/slack-orb-go/blob/8c4e86c9a787c240138244610aada066059b5b46/src/commands/notify.yml#L80
# TODO: keep support for this? Or will people not bother? They would have to do to the packagecloud URL and find it
# when we parametize the version, we should support this as well
input_sha256=""
if [ ! -f "$binary_zip" ]; then
mkdir -p "$orb_bin_dir"
if ! determine_http_client; then
printf '%s\n' "cURL or wget is required to download the Sagemaker Orb binary."
printf '%s\n' "Please install cURL or wget and try again."
exit 1
fi
printf '%s\n' "HTTP client: $HTTP_CLIENT."
binary_url="https://packagecloud.io/${org}/${repo_name}/packages/anyfile/${binary_name}.zip/download?distro_version_id=230"
printf '%s\n' "Release URL: $binary_url."
if ! download_binary "$binary_zip" "$binary_url" "$HTTP_CLIENT"; then
printf '%s\n' "Failed to download $repo_name binary from Packagecloud."
exit 1
fi
printf '%s\n' "Downloaded $repo_name zip to $orb_bin_dir"
else
printf '%s\n' "Skipping zip download since it already exists at $binary_zip."
fi
# Validate binary
## This validates, even if the binary already existed before.
## This can help with cache integrity but was also a convenience for testing where the binary will never be downloaded.
if [ -n "$input_sha256" ]; then
actual_sha256=""
actual_sha256=$(sha256sum "$binary_zip" | cut -d' ' -f1)
if [ "$actual_sha256" != "$input_sha256" ]; then
print_error "SHA256 checksum does not match. Expected $input_sha256 but got $actual_sha256"
exit 1
else
print_success "SHA256 checksum matches. Binary is valid."
fi
else
print_warn "SHA256 checksum not provided. Skipping validation."
fi
# Unzip binary
# Please someone explain this to me - only works if I UNZIP to the $orb_bin_dir
# Can't unzip to any other dest. then it fails. bad. wtf. If you can explain it, please let me know! Really.
printf '%s\n' "Unzip ${binary_zip}..."
if ! unzip -q "${binary_zip}" -d "$orb_bin_dir"; then
print_error "Failed to unzip $binary_zip."
exit 1
fi
printf '%s\n' "Making ${orb_bin_dir}/${binary_name} binary executable..."
if ! chmod +x "${orb_bin_dir}/${binary_name}"; then
print_error "Failed to make ${orb_bin_dir}/${binary_name} binary executable."
exit 1
fi
printf '%s\n' "Moving $binary_name to PATH and renaming to ${basic_name}..."
mkdir -p "$path_destination"
# shellcheck disable=SC2016
# shellcheck disable=SC2086
echo 'export PATH='${path_destination}':$PATH' >> "$BASH_ENV"
# shellcheck disable=SC1090
# shellcheck disable=SC3046
source "$BASH_ENV"
if ! mv "$orb_bin_dir/$binary_name" "${path_destination}/${basic_name}"; then
print_error "Failed to move $orb_bin_dir/$binary_name binary executable."
exit 1
fi
print_success "Successfully installed $basic_name."
exit 0
name: Download cci-sagemaker binary
jobs:
create_endpoint_configuration:
description: |
This job creates an create an endpoint configuration.
docker:
- image: cimg/base:current-22.04
parameters:
bucket:
description: S3 bucket for the model
type: string
circle_pipeline_id:
description: CircleCI Pipeline ID
type: string
circle_project_id:
description: CircleCI project ID
type: string
endpoint_instance_count:
default: 1
description: Number of instances to run endpoint inference
type: integer
endpoint_instance_type:
default: ml.t2.medium
description: EC2 instance type to run endpoint inference
type: string
model_name:
description: Model name
type: string
region_name:
description: AWS region name
type: string
steps:
- setup
- create_endpoint_configuration:
bucket: << parameters.bucket >>
circle_pipeline_id: << parameters.circle_pipeline_id >>
circle_project_id: << parameters.circle_project_id >>
endpoint_instance_count: << parameters.endpoint_instance_count >>
endpoint_instance_type: << parameters.endpoint_instance_type >>
model_name: << parameters.model_name >>
region_name: << parameters.region_name >>
create_model:
description: |
This job creates sagemaker model given a model package available in registry.
docker:
- image: cimg/base:current-22.04
parameters:
bucket:
description: S3 bucket for the model
type: string
circle_pipeline_id:
description: CircleCI Pipeline ID
type: string
model_name:
description: Model name
type: string
region_name:
description: AWS region name
type: string
steps:
- setup
- create_model:
bucket: << parameters.bucket >>
circle_pipeline_id: << parameters.circle_pipeline_id >>
model_name: << parameters.model_name >>
region_name: << parameters.region_name >>
deploy_endpoint:
description: |
This job deploys inference endpoint by create or update
docker:
- image: cimg/base:current-22.04
parameters:
bucket:
description: S3 bucket for the model
type: string
circle_pipeline_id:
description: CircleCI Pipeline ID
type: string
circle_project_id:
description: CircleCI project ID
type: string
enable_restore_version:
default: "true"
description: Enable restoring releases on circleci
type: string
enable_scaling_component:
default: "true"
description: Enable scaling release components on circleci
type: string
endpoint_config_name:
default: ""
description: Existing endpoint config name
type: string
endpoint_name:
default: ""
description: Existing endpoint name
type: string
model_desc:
description: sagemaker model desription
type: string
model_name:
description: Model name
type: string
region_name:
description: AWS region name
type: string
steps:
- setup
- deploy_endpoint:
bucket: << parameters.bucket >>
circle_pipeline_id: << parameters.circle_pipeline_id >>
circle_project_id: << parameters.circle_project_id >>
enable_restore_version: << parameters.enable_restore_version>>
enable_scaling_component: << parameters.enable_scaling_component>>
endpoint_config_name: << parameters.endpoint_config_name >>
endpoint_name: << parameters.endpoint_name >>
model_desc: << parameters.model_desc >>
model_name: << parameters.model_name >>
region_name: << parameters.region_name >>
examples:
deploy_model:
description: |
This example defines a CircleCI workflow that uses the aws-sagemaker orb to:
- Create a SageMaker model
- Create an endpoint configuration for the model
- Deploy the model to an endpoint
It shows a typical machine learning workflow using SageMaker with model releases being tracked on CircleCI.
usage:
version: "2.1"
orbs:
aws-sagemaker: circleci/aws-sagemaker@1.0
workflows:
ai-workflow:
jobs:
- aws-sagemaker/create_model:
bucket: replace_with_s3_bucket_name
model_name: replace_with_model_name
name: create-model
pipeline_id: << pipeline.id >>
region_name: replace_with_region_name
- aws-sagemaker/create_endpoint_configuration:
bucket: replace_with_s3_bucket_name
model_name: replace_with_model_name
name: create-endpoint-configuration
pipeline_id: << pipeline.id >>
region_name: replace_with_region_name
requires:
- create-model
- aws-sagemaker/deploy_endpoint:
bucket: replace_with_s3_bucket_name
model_desc: replace_with_description
model_name: replace_with_model_name
name: deploy-endpoint
pipeline_id: <<pipeline.id>>
project_id: replace_with_project_id
region_name: replace_with_region_name
requires:
- create-endpoint-configuration
deploy_model_with_endpoint_config_name:
description: |
Deploys a model to SageMaker with an existing endpoint configuration.
usage:
version: "2.1"
orbs:
aws-sagemaker: circleci/aws-sagemaker@1.0
workflows:
ai-workflow:
jobs:
- aws-sagemaker/deploy_endpoint:
bucket: replace_with_s3_bucket_name
endpoint_config_name: replace_with_existing_endpoint_config_name
endpoint_name: replace_with_endpoint_name
model_desc: replace_with_description
model_name: replace_with_model_name
name: deploy-endpoint
pipeline_id: <<pipeline.id>>
project_id: replace_with_project_id
region_name: replace_with_region_name