Instructions to use Gryphe/MergeMonster-13b-20231124 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Gryphe/MergeMonster-13b-20231124 with Transformers:

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

pipe = pipeline("text-generation", model="Gryphe/MergeMonster-13b-20231124")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Gryphe/MergeMonster-13b-20231124")
model = AutoModelForCausalLM.from_pretrained("Gryphe/MergeMonster-13b-20231124")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use Gryphe/MergeMonster-13b-20231124 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Gryphe/MergeMonster-13b-20231124"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Gryphe/MergeMonster-13b-20231124",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Gryphe/MergeMonster-13b-20231124

SGLang

How to use Gryphe/MergeMonster-13b-20231124 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 "Gryphe/MergeMonster-13b-20231124" \
    --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": "Gryphe/MergeMonster-13b-20231124",
		"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 "Gryphe/MergeMonster-13b-20231124" \
        --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": "Gryphe/MergeMonster-13b-20231124",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Gryphe/MergeMonster-13b-20231124 with Docker Model Runner:
```
docker model run hf.co/Gryphe/MergeMonster-13b-20231124
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

I received a request to upload the FP16 version of my first MergeMonster 13b merge, which had the sole focus of reducing overused ChatGPT roleplaying phrases. Many ministrations were murdered throughout this process.

GGUF quantizations can be found in my main MergeMonster repo.

Merge Composition

Unsurprisingly, not much remained of MythoMax. I'm (obviously!) very fond of it, but it's not perfect.

Reminder: MergeMonster works with hundreds of pieces scattered throughout the final merge. If a model takes a large part of the composition this does not mean it is merged in a manner that might be logical to humans.

Base model: MythoMax-L2-13b

------ FINAL MERGE COMPOSITION ------
LLaMA2-13B-Tiefighter: 0.27
Emerhyst-13B: 0.26
Orca-2-13b: 0.24
Stheno-1.8-L2-13B: 0.09
MythoMax-L2-13b: 0.09
Athena-v4: 0.04

Prompt Format

As always, I suggest sticking to the Alpaca format.

### Instruction:
Your instruction or question here.

### Response:

Downloads last month: 18

Safetensors

Model size

13B params

Tensor type

F16

Model tree for Gryphe/MergeMonster-13b-20231124

Quantizations

3 models