Instructions to use kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2 with Transformers:

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

pipe = pipeline("text-generation", model="kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2")
model = AutoModelForCausalLM.from_pretrained("kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2", 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 kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2

SGLang

How to use kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2 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 "kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2" \
    --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": "kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2",
		"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 "kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2" \
        --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": "kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2 with Docker Model Runner:
```
docker model run hf.co/kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2
```

Configuration Parsing Warning:In config.json: "quantization_config.bits" must be an integer

EXL2 quants of ryzen88/Llama-3-70b-Arimas-story-RP-V2.1

2.50 bits per weight
2.75 bits per weight
3.00 bits per weight
3.25 bits per weight
3.50 bits per weight
4.00 bits per weight
4.50 bits per weight
5.00 bits per weight
6.00 bits per weight
8.00 bits per weight

Created using the defaults from exllamav2 0.1.6 convert.py
2.5bpw to 6.0bpw head bits = 6
8.0bpw head bits = 8
length = 8192
dataset rows = 200
measurement rows = 32
measurement length = 8192

Merge_XL_model_Stock

This model switches to the Smaug instruct 32K for the base bodel. Expanded with Giraffe and Gradient to keep a robuust long context window. Higgs and cat for most of the story and RP aspects. Hermes and Chinese chat are for overall intelligence and understanding.

Merge Details

Merge Method

This model was merged using the Model Stock merge method using \Smaug-Llama-3-70B-Instruct-32K as a base.

Models Merged

The following models were included in the merge:

\Llama-3-Giraffe-70B-Instruct
\Llama-3-70B-Instruct-Gradient-262k
\Hermes-2-Theta-Llama-3-70B
\Higgs-Llama-3-70B
\Llama3-70B-Chinese-Chat
\Meta-LLama-3-Cat-A-LLama-70b

Configuration

The following YAML configuration was used to produce this model:

models:
    - model: \Smaug-Llama-3-70B-Instruct-32K
    - model: \Llama-3-70B-Instruct-Gradient-262k
    - model: \Llama-3-Giraffe-70B-Instruct
    - model: \Higgs-Llama-3-70B
    - model: \Llama3-70B-Chinese-Chat
    - model: \Meta-LLama-3-Cat-A-LLama-70b
    - model: \Hermes-2-Theta-Llama-3-70B
merge_method: model_stock
base_model: \Smaug-Llama-3-70B-Instruct-32K
dtype: bfloat16

Downloads last month: 9

Paper for kim512/Llama-3-70b-Arimas-story-RP-V2.1-3.25bpw-h6-exl2

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

Paper • 2403.19522 • Published Mar 28, 2024 • 15