Instructions to use unsloth/gemma-2-2b-bnb-4bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use unsloth/gemma-2-2b-bnb-4bit with Transformers:

# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-generation", model="unsloth/gemma-2-2b-bnb-4bit")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("unsloth/gemma-2-2b-bnb-4bit")
model = AutoModelForCausalLM.from_pretrained("unsloth/gemma-2-2b-bnb-4bit", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use unsloth/gemma-2-2b-bnb-4bit with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "unsloth/gemma-2-2b-bnb-4bit"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "unsloth/gemma-2-2b-bnb-4bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/unsloth/gemma-2-2b-bnb-4bit

SGLang

How to use unsloth/gemma-2-2b-bnb-4bit with SGLang:

Install from pip and serve model

# Install SGLang from pip:
pip install sglang
# Start the SGLang server:
python3 -m sglang.launch_server \
    --model-path "unsloth/gemma-2-2b-bnb-4bit" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "unsloth/gemma-2-2b-bnb-4bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker images

docker run --gpus all \
    --shm-size 32g \
    -p 30000:30000 \
    -v ~/.cache/huggingface:/root/.cache/huggingface \
    --env "HF_TOKEN=<secret>" \
    --ipc=host \
    lmsysorg/sglang:latest \
    python3 -m sglang.launch_server \
        --model-path "unsloth/gemma-2-2b-bnb-4bit" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "unsloth/gemma-2-2b-bnb-4bit",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Unsloth Studio

How to use unsloth/gemma-2-2b-bnb-4bit with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for unsloth/gemma-2-2b-bnb-4bit to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for unsloth/gemma-2-2b-bnb-4bit to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for unsloth/gemma-2-2b-bnb-4bit to start chatting

Load model with FastModel

pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="unsloth/gemma-2-2b-bnb-4bit",
    max_seq_length=2048,
)

Docker Model Runner
How to use unsloth/gemma-2-2b-bnb-4bit with Docker Model Runner:
```
docker model run hf.co/unsloth/gemma-2-2b-bnb-4bit
```

OOM when finetuning with lora.

by zokica - opened Aug 1, 2024

Discussion

zokica

Aug 1, 2024

Not sure if this applies to just this model or also official version of gemma2, but when doing Peft finetuning, I always get OOM error at the time when the model gets saved.

There is enough memory for sure, 48GB card, uses less than 20GB and then a huge spike when saving the model. Does not happen even with 9B model.

#############
OutOfMemoryError: CUDA out of memory. Tried to allocate 15.26 GiB. GPU 0 has a total capacty of 47.40 GiB of which 11.35 GiB is free. Process 2397559 has 35.67 GiB memory in use. Of the allocated memory 34.28 GiB is allocated by PyTorch, and 903.30 MiB is reserved by PyTorch but unallocated. If reserved but unallocated memory is large try setting max_split_size_mb to avoid fragmentation. See documentation for Memory Management and PYTORCH_CUDA_ALLOC_CONF
###################################

zokica

Aug 1, 2024

Yes, the original model does the same.

zokica

Aug 1, 2024

Works with changed params:

optim="paged_adamw_8bit",
evaluation_strategy="no"
do_eval=False,

shimmyshimmer

Unsloth AI org Aug 2, 2024

Works with changed params:

optim="paged_adamw_8bit",
evaluation_strategy="no"
do_eval=False,

oh so that was the problem im guessing? glad you got it solved

jvelja

Aug 8, 2024

Same problem here, old Gemma 2b wouldnt go OOM for LoRA though. Questioning why normal adamw should have these kinda issues.

shimmyshimmer

Unsloth AI org Aug 11, 2024

It's mostly because of Flash Attention is not installed for Gemma models. Please update Unsloth and it will tell you to install Flash Attention

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment