Skip to content

change: TGI NeuronX #4375

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 4 commits into from
Jan 19, 2024
Merged

change: TGI NeuronX #4375

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
4 commits
Select commit Hold shift + click to select a range
a7c6722
TGI NeuronX
amzn-choeric Jan 16, 2024
d19af36
Update
amzn-choeric Jan 17, 2024
231b4e7
Update
amzn-choeric Jan 17, 2024
4e45ad7
Merge branch 'master' into tgi-neuronx
amzn-choeric Jan 19, 2024
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 8 additions & 0 deletions src/sagemaker/huggingface/llm_utils.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -57,6 +57,14 @@ def get_huggingface_llm_image_uri(
version=version,
image_scope="inference",
)
if backend == "huggingface-neuronx":
return image_uris.retrieve(
"huggingface-llm-neuronx",
region=region,
version=version,
image_scope="inference",
inference_tool="neuronx",
)
if backend == "lmi":
version = version or "0.24.0"
return image_uris.retrieve(framework="djl-deepspeed", region=region, version=version)
Expand Down
41 changes: 41 additions & 0 deletions src/sagemaker/image_uri_config/huggingface-llm-neuronx.json
View file
Open in desktop
Original file line number Diff line number Diff line change
@@ -0,0 +1,41 @@
{
"inference": {
"processors": [
"inf2"
],
"version_aliases": {
"0.0": "0.0.16"
},
"versions": {
"0.0.16": {
"py_versions": [
"py310"
],
"registries": {
"ap-northeast-1": "763104351884",
"ap-south-1": "763104351884",
"ap-south-2": "772153158452",
"ap-southeast-1": "763104351884",
"ap-southeast-2": "763104351884",
"ap-southeast-4": "457447274322",
"eu-central-1": "763104351884",
"eu-central-2": "380420809688",
"eu-south-2": "503227376785",
"eu-west-1": "763104351884",
"eu-west-3": "763104351884",
"il-central-1": "780543022126",
"sa-east-1": "763104351884",
"us-east-1": "763104351884",
"us-east-2": "763104351884",
"us-west-2": "763104351884",
"ca-west-1": "204538143572"
},
"tag_prefix": "1.13.1-optimum0.0.16",
"repository": "huggingface-pytorch-tgi-inference",
"container_version": {
"inf2": "ubuntu22.04"
}
}
}
}
}
2 changes: 2 additions & 0 deletions src/sagemaker/image_uris.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -37,6 +37,7 @@
ECR_URI_TEMPLATE = "{registry}.dkr.{hostname}/{repository}"
HUGGING_FACE_FRAMEWORK = "huggingface"
HUGGING_FACE_LLM_FRAMEWORK = "huggingface-llm"
HUGGING_FACE_LLM_NEURONX_FRAMEWORK = "huggingface-llm-neuronx"
XGBOOST_FRAMEWORK = "xgboost"
SKLEARN_FRAMEWORK = "sklearn"
TRAINIUM_ALLOWED_FRAMEWORKS = "pytorch"
Expand Down Expand Up @@ -470,6 +471,7 @@ def _validate_version_and_set_if_needed(version, config, framework):
if version is None and framework in [
DATA_WRANGLER_FRAMEWORK,
HUGGING_FACE_LLM_FRAMEWORK,
HUGGING_FACE_LLM_NEURONX_FRAMEWORK,
STABILITYAI_FRAMEWORK,
]:
version = _get_latest_versions(available_versions)
Expand Down
31 changes: 20 additions & 11 deletions tests/unit/sagemaker/image_uris/test_huggingface_llm.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -19,29 +19,38 @@

LMI_VERSIONS = ["0.24.0"]
HF_VERSIONS_MAPPING = {
"0.6.0": "2.0.0-tgi0.6.0-gpu-py39-cu118-ubuntu20.04",
"0.8.2": "2.0.0-tgi0.8.2-gpu-py39-cu118-ubuntu20.04",
"0.9.3": "2.0.1-tgi0.9.3-gpu-py39-cu118-ubuntu20.04",
"1.0.3": "2.0.1-tgi1.0.3-gpu-py39-cu118-ubuntu20.04",
"1.1.0": "2.0.1-tgi1.1.0-gpu-py39-cu118-ubuntu20.04",
"1.2.0": "2.1.1-tgi1.2.0-gpu-py310-cu121-ubuntu20.04",
"1.3.1": "2.1.1-tgi1.3.1-gpu-py310-cu121-ubuntu20.04",
"1.3.3": "2.1.1-tgi1.3.3-gpu-py310-cu121-ubuntu20.04",
"gpu": {
"0.6.0": "2.0.0-tgi0.6.0-gpu-py39-cu118-ubuntu20.04",
"0.8.2": "2.0.0-tgi0.8.2-gpu-py39-cu118-ubuntu20.04",
"0.9.3": "2.0.1-tgi0.9.3-gpu-py39-cu118-ubuntu20.04",
"1.0.3": "2.0.1-tgi1.0.3-gpu-py39-cu118-ubuntu20.04",
"1.1.0": "2.0.1-tgi1.1.0-gpu-py39-cu118-ubuntu20.04",
"1.2.0": "2.1.1-tgi1.2.0-gpu-py310-cu121-ubuntu20.04",
"1.3.1": "2.1.1-tgi1.3.1-gpu-py310-cu121-ubuntu20.04",
"1.3.3": "2.1.1-tgi1.3.3-gpu-py310-cu121-ubuntu20.04",
},
"inf2": {
"0.0.16": "1.13.1-optimum0.0.16-neuronx-py310-ubuntu22.04",
},
}


@pytest.mark.parametrize("load_config", ["huggingface-llm.json"], indirect=True)
@pytest.mark.parametrize(
"load_config", ["huggingface-llm.json", "huggingface-llm-neuronx.json"], indirect=True
)
def test_huggingface_uris(load_config):
VERSIONS = load_config["inference"]["versions"]
device = load_config["inference"]["processors"][0]
backend = "huggingface-neuronx" if device == "inf2" else "huggingface"
for version in VERSIONS:
ACCOUNTS = load_config["inference"]["versions"][version]["registries"]
for region in ACCOUNTS.keys():
uri = get_huggingface_llm_image_uri("huggingface", region=region, version=version)
uri = get_huggingface_llm_image_uri(backend, region=region, version=version)
expected = expected_uris.huggingface_llm_framework_uri(
"huggingface-pytorch-tgi-inference",
ACCOUNTS[region],
version,
HF_VERSIONS_MAPPING[version],
HF_VERSIONS_MAPPING[device][version],
region=region,
)
assert expected == uri
Expand Down