Bloomz models having task name as textgeneration1 on JumpStart

[mrgiba]

Describe the bug I was looking for the bloomz models on jumpstart and I noticed the task name for them is oddly textgeneration1. Is that on purpose ?

To reproduce Code snippet:

from sagemaker.jumpstart.filters import And
from sagemaker.jumpstart.notebook_utils import list_jumpstart_models

filter_value = And("framework == huggingface", "training_supported == true")
model_list = [m for m in list_jumpstart_models(filter=filter_value) if "bloom" in m]

print(model_list)

Expected behavior I assume the task name should be textgeneration instead

Screenshots or logs Output:

['huggingface-textgeneration1-bloom-3b', 'huggingface-textgeneration1-bloom-3b-fp16', 'huggingface-textgeneration1-bloom-7b1', 'huggingface-textgeneration1-bloom-7b1-fp16', 'huggingface-textgeneration1-bloomz-3b-fp16', 'huggingface-textgeneration1-bloomz-7b1-fp16']

System information A description of your system. Please provide:

• SageMaker Python SDK version: 2.165.0
• Framework name (eg. PyTorch) or algorithm (eg. KMeans):
• Framework version:
• Python version:
• CPU or GPU:
• Custom Docker image (Y/N):

Additional context Add any other context about the problem here.

[Neo9061]

Hi @mrgiba sorry for the confusion. We put them in separate tasks as text generation 1 uses LMI DLC while text generation uses standard HF DLC.

We are migrating everything to llm task name which uses latest HF TGI DLC. Will keep you updated

[martinRenou]

@Neo9061 I can still reproduce this.

We are migrating everything to llm task name which uses latest HF TGI DLC. Will keep you updated

Would you be able to give an update for this?