Instructions to use grimjim/Gemma2-Nephilim-v3-9B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use grimjim/Gemma2-Nephilim-v3-9B with Transformers:

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

pipe = pipeline("text-generation", model="grimjim/Gemma2-Nephilim-v3-9B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("grimjim/Gemma2-Nephilim-v3-9B")
model = AutoModelForCausalLM.from_pretrained("grimjim/Gemma2-Nephilim-v3-9B")
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 Settings

vLLM

How to use grimjim/Gemma2-Nephilim-v3-9B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "grimjim/Gemma2-Nephilim-v3-9B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "grimjim/Gemma2-Nephilim-v3-9B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/grimjim/Gemma2-Nephilim-v3-9B

SGLang

How to use grimjim/Gemma2-Nephilim-v3-9B 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 "grimjim/Gemma2-Nephilim-v3-9B" \
    --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": "grimjim/Gemma2-Nephilim-v3-9B",
		"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 "grimjim/Gemma2-Nephilim-v3-9B" \
        --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": "grimjim/Gemma2-Nephilim-v3-9B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use grimjim/Gemma2-Nephilim-v3-9B with Docker Model Runner:
```
docker model run hf.co/grimjim/Gemma2-Nephilim-v3-9B
```

Gemma2-Nephilim-v3-9B

File size: 2,136 Bytes

---
base_model:
- UCLA-AGI/Gemma-2-9B-It-SPPO-Iter3
- princeton-nlp/gemma-2-9b-it-SimPO
library_name: transformers
tags:
- mergekit
- merge
license: gemma
pipeline_tag: text-generation
---
# Gemma2-Nephilim-v3-9B

This repo contains a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).

Though none of the components of this merge were trained for roleplay nor intended for it, the model can be used effectively in that role.

Tested with temperature 1 and minP 0.01. This model leans toward being creative, so adjust temperature upward or downward as desired. The Instruct template used during testing can be found below:

- [context template](https://huggingface.co/debased-ai/SillyTavern-settings/blob/main/advanced_formatting/context_template/Gemma2%20Unleashed3.json)
- [instruct prompt](https://huggingface.co/debased-ai/SillyTavern-settings/blob/main/advanced_formatting/instruct_mode/Gemma2%20Unleashed3.json)

Afterword: In subsequent testing, I encountered an occasional lapse in tracking context for complex scenarios, which seems to originate in the SPPO model. This lapse is not present in [grimjim/Kitsunebi-v1-Gemma2-8k-9B](https://huggingface.co/grimjim/Kitsunebi-v1-Gemma2-8k-9B).

## Merge Details
### Merge Method

This model was merged using the SLERP merge method.

### Models Merged

The following models were included in the merge:
* [UCLA-AGI/Gemma-2-9B-It-SPPO-Iter3](https://huggingface.co/UCLA-AGI/Gemma-2-9B-It-SPPO-Iter3)
* [princeton-nlp/gemma-2-9b-it-SimPO](https://huggingface.co/princeton-nlp/gemma-2-9b-it-SimPO)

### Configuration

The following YAML configuration was used to produce this model:

```yaml
slices:
- sources:
  - model: princeton-nlp/gemma-2-9b-it-SimPO
    layer_range:
    - 0
    - 42
  - model: UCLA-AGI/Gemma-2-9B-It-SPPO-Iter3
    layer_range:
    - 0
    - 42
merge_method: slerp
base_model: UCLA-AGI/Gemma-2-9B-It-SPPO-Iter3
parameters:
  t:
  - filter: self_attn
    value:
    - 0
    - 0.5
    - 0.3
    - 0.7
    - 1
  - filter: mlp
    value:
    - 1
    - 0.5
    - 0.7
    - 0.3
    - 0
  - value: 0.5
dtype: bfloat16

```