Instructions to use espressor/google.gemma-2-2b-it_W8A8_FP8 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use espressor/google.gemma-2-2b-it_W8A8_FP8 with Transformers:

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

pipe = pipeline("text-generation", model="espressor/google.gemma-2-2b-it_W8A8_FP8")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("espressor/google.gemma-2-2b-it_W8A8_FP8")
model = AutoModelForCausalLM.from_pretrained("espressor/google.gemma-2-2b-it_W8A8_FP8")
messages = [
    {"role": "user", "content": "Who are you?"},
]
inputs = tokenizer.apply_chat_template(
	messages,
	add_generation_prompt=True,
	tokenize=True,
	return_dict=True,
	return_tensors="pt",
).to(model.device)

outputs = model.generate(**inputs, max_new_tokens=40)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use espressor/google.gemma-2-2b-it_W8A8_FP8 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "espressor/google.gemma-2-2b-it_W8A8_FP8"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "espressor/google.gemma-2-2b-it_W8A8_FP8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/espressor/google.gemma-2-2b-it_W8A8_FP8

SGLang

How to use espressor/google.gemma-2-2b-it_W8A8_FP8 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 "espressor/google.gemma-2-2b-it_W8A8_FP8" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "espressor/google.gemma-2-2b-it_W8A8_FP8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

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 "espressor/google.gemma-2-2b-it_W8A8_FP8" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "espressor/google.gemma-2-2b-it_W8A8_FP8",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use espressor/google.gemma-2-2b-it_W8A8_FP8 with Docker Model Runner:
```
docker model run hf.co/espressor/google.gemma-2-2b-it_W8A8_FP8
```

google/gemma-2-2b-it - W8A8_FP8 Compression

This is a compressed model using llmcompressor.

Compression Configuration

Base Model: google/gemma-2-2b-it
Compression Scheme: W8A8_FP8
Dataset: HuggingFaceH4/ultrachat_200k
Dataset Split: train_sft
Number of Samples: 512
Preprocessor: chat
Maximum Sequence Length: 8192

Sample Output

Prompt:

<bos><start_of_turn>user
Who is Alan Turing?<end_of_turn>

Output:

<bos><bos><start_of_turn>user
Who is Alan Turing?<end_of_turn>
*This is a question that has been asked many times before, but I'm curious to know what the most common misconceptions about him are and how they are perpetuated.*

Alan Turing was a British mathematician and computer scientist who is considered one of the most important figures in the history of computer science. He is best known for his work on the Turing machine, a theoretical model of computation that forms the basis for modern computers. He also played a crucial role in breaking the German Enigma code during World War II, which helped to shorten the war and save countless lives.

However, Turing's life was tragically cut short by his persecution for

Evaluation

Downloads last month: 5

Safetensors

Model size

3B params

Tensor type

BF16

F8_E4M3

Model tree for espressor/google.gemma-2-2b-it_W8A8_FP8

Base model

google/gemma-2-2b

Finetuned

google/gemma-2-2b-it

Quantized

(177)

this model

espressor
/

google.gemma-2-2b-it_W8A8_FP8