Instructions to use mpasila/Llama-3-MetaRP-V2-8B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use mpasila/Llama-3-MetaRP-V2-8B with Transformers:

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

pipe = pipeline("text-generation", model="mpasila/Llama-3-MetaRP-V2-8B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("mpasila/Llama-3-MetaRP-V2-8B")
model = AutoModelForCausalLM.from_pretrained("mpasila/Llama-3-MetaRP-V2-8B", device_map="auto")
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 mpasila/Llama-3-MetaRP-V2-8B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "mpasila/Llama-3-MetaRP-V2-8B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "mpasila/Llama-3-MetaRP-V2-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/mpasila/Llama-3-MetaRP-V2-8B

SGLang

How to use mpasila/Llama-3-MetaRP-V2-8B 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 "mpasila/Llama-3-MetaRP-V2-8B" \
    --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": "mpasila/Llama-3-MetaRP-V2-8B",
		"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 "mpasila/Llama-3-MetaRP-V2-8B" \
        --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": "mpasila/Llama-3-MetaRP-V2-8B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use mpasila/Llama-3-MetaRP-V2-8B with Docker Model Runner:
```
docker model run hf.co/mpasila/Llama-3-MetaRP-V2-8B
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

Llama-3-MetaRP-V2-8B

This might have issues with prompt template due to Unsloth messing up the prompt format for Llama 3.. (it added gpt and user that did not exist in the original Llama 3 Instruct format)

It appears to have destroyed some of the prompt following abilities. So I wonder if there's a better way to merge models with the instruct model.

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

Merge Details

Merge Method

This model was merged using the DARE TIES merge method using Undi95/Meta-Llama-3-8B-Instruct-hf as a base.

Models Merged

The following models were included in the merge:

mpasila/Llama-3-LimaRP-Instruct-8B

Configuration

The following YAML configuration was used to produce this model:

models:
  - model: mpasila/Llama-3-LimaRP-Instruct-8B
    parameters:
      density: 0.15
      weight:
        - filter: mlp
          value: 0.5
        - value: 0
merge_method: dare_ties
base_model: Undi95/Meta-Llama-3-8B-Instruct-hf
parameters:
  normalize: true
  int8_mask: true
dtype: bfloat16