Instructions to use saishf/Merge-Mayhem-L3-V2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use saishf/Merge-Mayhem-L3-V2 with Transformers:

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

pipe = pipeline("text-generation", model="saishf/Merge-Mayhem-L3-V2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("saishf/Merge-Mayhem-L3-V2")
model = AutoModelForCausalLM.from_pretrained("saishf/Merge-Mayhem-L3-V2")
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]:]))

Inference
Local Apps Settings

vLLM

How to use saishf/Merge-Mayhem-L3-V2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "saishf/Merge-Mayhem-L3-V2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "saishf/Merge-Mayhem-L3-V2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/saishf/Merge-Mayhem-L3-V2

SGLang

How to use saishf/Merge-Mayhem-L3-V2 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 "saishf/Merge-Mayhem-L3-V2" \
    --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": "saishf/Merge-Mayhem-L3-V2",
		"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 "saishf/Merge-Mayhem-L3-V2" \
        --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": "saishf/Merge-Mayhem-L3-V2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use saishf/Merge-Mayhem-L3-V2 with Docker Model Runner:
```
docker model run hf.co/saishf/Merge-Mayhem-L3-V2
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Quants

mradermacher has kindly provided quants here: mradermacher/Merge-Mayhem-L3-V2-GGUF

merge

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

Merge Details

This is quite an interesting model, it's fun so far. But quite harsh, so if that's something you don't like, this model isn't for you :3

it's an attempt at loosely recreating ResplendentAI/SOVL_Llama3_8B but trying to keep it smarter, with the lovely openlynn/Llama-3-Soliloquy-8B-v2 holding it together. I'm personally enjoying this model, it's different from most llama-3 models.

Merge Method

This model was merged using the Model Stock merge method using openlynn/Llama-3-Soliloquy-8B-v2 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: Undi95/Meta-Llama-3-8B-Instruct-hf+ResplendentAI/Aura_Llama3
  - model: Undi95/Meta-Llama-3-8B-Instruct-hf+ResplendentAI/Smarts_Llama3
  - model: Undi95/Meta-Llama-3-8B-Instruct-hf+ResplendentAI/Luna_Llama3
  - model: Undi95/Meta-Llama-3-8B-Instruct-hf+ResplendentAI/BlueMoon_Llama3
  - model: Undi95/Meta-Llama-3-8B-Instruct-hf+ResplendentAI/RP_Format_QuoteAsterisk_Llama3
merge_method: model_stock
base_model: openlynn/Llama-3-Soliloquy-8B-v2
dtype: float16