the problem about mosaicml-streaming

[sysusicily]

when I follow pip install -e ".[gpu]"，I find the error about mosaicml-streaming
#-------------------------------------------------------------------------------------------
root@7730f5bd29fa:/home/mosaicml/llm-foundry# pip list|grep stream
mosaicml-streaming 0.2.1
root@7730f5bd29fa:/home/mosaicml/llm-foundry#
root@7730f5bd29fa:/home/mosaicml/llm-foundry# python scripts/data_prep/convert_dataset_hf.py --dataset c4 --data_subset en --out_root ./my-copy-c4 --splits train_small val_small --concat_tokens 2048 --tokenizer EleutherAI/gpt-neox-20b --eos_text '<|endoftext|>'
Traceback (most recent call last):
File "/home/mosaicml/llm-foundry/llmfoundry/init.py", line 8, in
from llmfoundry.data import (ConcatTokensDataset,
File "/home/mosaicml/llm-foundry/llmfoundry/data/init.py", line 5, in
from llmfoundry.data.denoising import (MixtureOfDenoisersCollator,
File "/home/mosaicml/llm-foundry/llmfoundry/data/denoising.py", line 19, in
from llmfoundry.data.text_data import StreamingTextDataset
File "/home/mosaicml/llm-foundry/llmfoundry/data/text_data.py", line 15, in
from streaming import Stream, StreamingDataset
ImportError: cannot import name 'Stream' from 'streaming' (/usr/lib/python3/dist-packages/streaming/init.py)

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
File "/home/mosaicml/llm-foundry/scripts/data_prep/convert_dataset_hf.py", line 19, in
from llmfoundry.data.datasets import ConcatTokensDataset, NoConcatDataset
File "/home/mosaicml/llm-foundry/llmfoundry/init.py", line 32, in
raise ImportError(
ImportError: Please make sure to pip install . to get the requirements for the LLM example.
root@7730f5bd29fa:/home/mosaicml/llm-foundry#

[hanlint]

Hello @sysusicily , the Stream functionality was introduced in v0.4.0 (Release Notes), but you are on v0.2.1. Can you try:

pip install mosaicml-streaming==0.4.1

We will update our install requirements to have this minimum dependency, thank you!