Instructions to use jsgreenawalt/gemma-advanced-v1 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use jsgreenawalt/gemma-advanced-v1 with Transformers:

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

pipe = pipeline("text-generation", model="jsgreenawalt/gemma-advanced-v1")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("jsgreenawalt/gemma-advanced-v1")
model = AutoModelForCausalLM.from_pretrained("jsgreenawalt/gemma-advanced-v1")
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 jsgreenawalt/gemma-advanced-v1 with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/jsgreenawalt/gemma-advanced-v1

SGLang

How to use jsgreenawalt/gemma-advanced-v1 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 "jsgreenawalt/gemma-advanced-v1" \
    --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": "jsgreenawalt/gemma-advanced-v1",
		"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 "jsgreenawalt/gemma-advanced-v1" \
        --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": "jsgreenawalt/gemma-advanced-v1",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use jsgreenawalt/gemma-advanced-v1 with Docker Model Runner:
```
docker model run hf.co/jsgreenawalt/gemma-advanced-v1
```

Gemma Advanced V1 (obsolete)

Note: A much-improved version is available at jsgreenawalt/gemma-2-9B-it-advanced-v2.1

Experimental merge #1, attempting to combine some of the advanced Gemma fine-tunes

Quants are available here: https://huggingface.co/QuantFactory/gemma-advanced-v1-GGUF

Notes and observations:

Recommended temperature 0.15 or lower , the model is more temperature sensitive than the parent models
Recommended Q8_0 quant, Q6* and lower quants lose more than quality than expected
The model writes coherently (at lower temperatures) and has a different writing style than the parent models

This is a merge of pre-trained language models created using mergekit.

Merge Details

Merge Method

This model was merged using the della merge method using google/google-gemma-2-9b-it as a base.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: google/gemma-2-9b-it 
    # no parameters necessary for base model
  - model: princeton-nlp/gemma-2-9b-it-SimPO 
    parameters:
      density: 0.5
      weight: 0.5
  - model: wzhouad/gemma-2-9b-it-WPO-HB
    parameters:
      density: 0.5
      weight: 0.5
merge_method: della
base_model: google/gemma-2-9b-it
parameters:
  normalize: true
dtype: float16

Downloads last month: 4

Safetensors

Model size

10B params

Tensor type

F16

Model tree for jsgreenawalt/gemma-advanced-v1

google/gemma-2-9b-it

princeton-nlp/gemma-2-9b-it-SimPO

wzhouad/gemma-2-9b-it-WPO-HB

Merge model

this model

Collection including jsgreenawalt/gemma-advanced-v1

Gemma-2-9B-it-Advanced

Collection

Merges of the advanced research fine tunes of gemma-2 9B it • 3 items • Updated Oct 20, 2024 • 3