Issue when trying to run instruction_tuning.ipynb

[kannangce]

I'm trying to run instruction_tuning.ipynb with a custom csv. Though I have't gotten to the part of custom data yet, I'm getting issue in the below code,

model = AutoModelForCausalLM.from_pretrained(
        BASE_MODEL,
        device_map="auto",
        load_in_8bit=True,
        trust_remote_code=True,
    )

The error is,

ImportError Traceback (most recent call last)
in <cell line: 1>()
----> 1 model = AutoModelForCausalLM.from_pretrained(
2 BASE_MODEL,
3 device_map="auto",
4 load_in_8bit=True,
5 trust_remote_code=True,

1 frames
/usr/local/lib/python3.10/dist-packages/transformers/modeling_utils.py in from_pretrained(cls, pretrained_model_name_or_path, config, cache_dir, ignore_mismatched_sizes, force_download, local_files_only, token, revision, use_safetensors, *model_args, **kwargs)
2399 if load_in_8bit or load_in_4bit:
2400 if not (is_accelerate_available() and is_bitsandbytes_available()):
-> 2401 raise ImportError(
2402 "Using load_in_8bit=True requires Accelerate: pip install accelerate and the latest version of"
2403 " bitsandbytes pip install -i https://test.pypi.org/simple/ bitsandbytes or"

ImportError: Using load_in_8bit=True requires Accelerate: pip install accelerate and the latest version of bitsandbytes pip install -i https://test.pypi.org/simple/ bitsandbytes or pip install bitsandbytes`

---

NOTE: If your import is failing due to a missing package, you can
manually install dependencies using either !pip or !apt.

To view examples of installing some common dependencies, click the
"Open Examples" button below.

I have executed the below pip installs and restart the collab.

!pip install  --upgrade transformers
!pip install peft
!pip install datasets
!pip install  --upgrade accelerate
!pip install --upgrade bitsandbytes

Is it known issue? If so any idea on getting it fixed?