upstage

/

SOLAR-10.7B-Instruct-v1.0

like 614

Follow

upstage 223

I deployed the model on sagemaker jupyter notebook and when I call it, I get an error.
The error log seen in cloudwatch is as follows.

[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - Prediction error	
[INFO ] W-9000-upstage__SOLAR-10.7B-Inst com.amazonaws.ml.mms.wlm.WorkerThread - Backend response time: 1	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - Traceback (most recent call last):	
[INFO ] W-9000-upstage__SOLAR-10.7B-Inst ACCESS_LOG - /169.254.178.2:48252 "POST /invocations HTTP/1.1" 400 2	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - File "/opt/conda/lib/python3.9/site-packages/sagemaker_huggingface_inference_toolkit/handler_service.py", line 219, in handle	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - self.initialize(context)	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - File "/opt/conda/lib/python3.9/site-packages/sagemaker_huggingface_inference_toolkit/handler_service.py", line 77, in initialize	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - self.model = self.load(self.model_dir)	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - File "/opt/conda/lib/python3.9/site-packages/sagemaker_huggingface_inference_toolkit/handler_service.py", line 104, in load	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - hf_pipeline = get_pipeline(task=os.environ["HF_TASK"], model_dir=model_dir, device=self.device)	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - File "/opt/conda/lib/python3.9/site-packages/sagemaker_huggingface_inference_toolkit/transformers_utils.py", line 272, in get_pipeline	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - hf_pipeline = pipeline(task=task, model=model_dir, device=device, **kwargs)	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - File "/opt/conda/lib/python3.9/site-packages/transformers/pipelines/__init__.py", line 675, in pipeline	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - config = AutoConfig.from_pretrained(model, _from_pipeline=task, **hub_kwargs, **model_kwargs)	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - File "/opt/conda/lib/python3.9/site-packages/transformers/models/auto/configuration_auto.py", line 873, in from_pretrained	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - config_class = CONFIG_MAPPING[config_dict["model_type"]]	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - File "/opt/conda/lib/python3.9/site-packages/transformers/models/auto/configuration_auto.py", line 579, in __getitem__	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - raise KeyError(key)	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - KeyError: 'llama'	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle -	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - During handling of the above exception, another exception occurred:	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle -	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - Traceback (most recent call last):	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - File "/opt/conda/lib/python3.9/site-packages/mms/service.py", line 108, in predict	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - ret = self._entry_point(input_batch, self.context)	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - File "/opt/conda/lib/python3.9/site-packages/sagemaker_huggingface_inference_toolkit/handler_service.py", line 243, in handle	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - raise PredictionException(str(e), 400)	
[INFO ] W-upstage__SOLAR-10.7B-Inst-1-stdout com.amazonaws.ml.mms.wlm.WorkerLifeCycle - mms.service.PredictionException: 'llama' : 400

The notbook code is as follows.

!pip install "sagemaker>=2.48.0" --upgrade

import sagemaker
import boto3

try:
    role = sagemaker.get_execution_role()
except ValueError:
    iam = boto3.client('iam')
    role = iam.get_role(RoleName='sagemaker_execution_role')['Role']['Arn']

print(f"sagemaker role arn: {role}")

import json
from sagemaker.huggingface import HuggingFaceModel, get_huggingface_llm_image_uri

# Hub Model configuration. https://huggingface.co/models
hub = {
    'HF_MODEL_ID':'upstage/SOLAR-10.7B-Instruct-v1.0',
    'SM_NUM_GPUS': json.dumps(1)
}

# create Hugging Face Model Class
huggingface_model = HuggingFaceModel(
    env=hub,
    role=role,
    transformers_version="4.26", # transformers version used
    pytorch_version="1.13", # pytorch version used
    py_version="py39", # python version of the DLC
)

# deploy model to SageMaker Inference
predictor = huggingface_model.deploy(
    initial_instance_count=1,
    instance_type="ml.g5.2xlarge",
    container_startup_health_check_timeout=300,
)

# send request
data = {
    "inputs": "Hey my name is Julien! How are you?",
}

predictor.predict(data)

I referred to this github: https://github.com/huggingface/notebooks/blob/main/sagemaker/11_deploy_model_from_hf_hub/deploy_transformer_model_from_hf_hub.ipynb
How can I solve the problem?
Please........ Help me ....