Instructions to use Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit with Transformers:

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

pipe = pipeline("text-generation", model="Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit")
model = AutoModelForCausalLM.from_pretrained("Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit", 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 Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit

SGLang

How to use Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit 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 "Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit" \
    --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": "Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit",
		"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 "Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit" \
        --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": "Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit with Docker Model Runner:
```
docker model run hf.co/Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit
```

merged

This is the 4.5bit EXL2 Quantized version of SnowyRP's first Llama 3 model. Quality of this model is unknown, since it is a Llama 3 model that I have not had the chance to test much.

.BF16

Merge Details

For this model I used the most uncensored and smartest finetunes of LLama 3, since LLama 3 has the largest ammount of training tokens I know of, so I expect it to have more than enough of a knowledge base.

This model is under both the LLama 3 licince and a few other ones that only affect conmerical uses, But I am not a lawer so use in commerical settings at your own risk.

Model Use:

Here is an modified Alpaca prompt I use with this model:

Below is an instruction that describes a task. Write a response that appropriately completes the request.

### Instruction: Take the role of {{char}} in a play that leaves a lasting impression on {{user}}.  Write {{char}}'s next reply. Never skip or gloss over {{char}}’s actions. Progress the scene at a naturally slow pace. Avoid repetition, don't loop. Develop the plot slowly, always stay in character. Describe all actions in full, elaborate, explicit, graphic, and vivid detail. Mention all relevant sensory perceptions.

Merge Method

This model was merged using the Model Stock merge method using kuotient/Meta-Llama-3-8B-Instruct as a base.

Models Merged

The following models were included in the merge:

Configuration

The following YAML configuration was used to produce this model:

base_model: kuotient/Meta-Llama-3-8B-Instruct
dtype: bfloat16
merge_method: model_stock
slices:
- sources:
  - layer_range: [0, 32]
    model: Undi95/Llama-3-Unholy-8B-e4
  - layer_range: [0, 32]
    model: nbeerbower/llama-3-dragonmaid-8B
  - layer_range: [0, 32]
    model: openlynn/Llama-3-Soliloquy-8B
  - layer_range: [0, 32]
    model: Locutusque/llama-3-neural-chat-v1-8b
  - layer_range: [0, 32]
    model: Dogge/llama-3-8B-instruct-Bluemoon-Freedom-RP
  - layer_range: [0, 32]
    model: kuotient/Meta-Llama-3-8B-Instruct

Downloads last month: 4

Model tree for Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit

Dogge/llama-3-8B-instruct-Bluemoon-Freedom-RP

Locutusque/llama-3-neural-chat-v1-8b

Undi95/Llama-3-Unholy-8B-e4

elyn-dev/Llama-3-Soliloquy-8B-v1

kuotient/Meta-Llama-3-8B-Instruct

nbeerbower/llama-3-dragonmaid-8B

Merge model

this model

Paper for Masterjp123/Llama-3-SnowyRP-8B-V1-EXL2-4.5bit

Model Stock: All we need is just a few fine-tuned models

Paper • 2403.19522 • Published Mar 28, 2024 • 15