Instructions to use notstoic/OPT-13B-Nerybus-Mix-4bit-128g with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use notstoic/OPT-13B-Nerybus-Mix-4bit-128g with Transformers:

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

pipe = pipeline("text-generation", model="notstoic/OPT-13B-Nerybus-Mix-4bit-128g")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("notstoic/OPT-13B-Nerybus-Mix-4bit-128g")
model = AutoModelForCausalLM.from_pretrained("notstoic/OPT-13B-Nerybus-Mix-4bit-128g")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use notstoic/OPT-13B-Nerybus-Mix-4bit-128g with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "notstoic/OPT-13B-Nerybus-Mix-4bit-128g"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "notstoic/OPT-13B-Nerybus-Mix-4bit-128g",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/notstoic/OPT-13B-Nerybus-Mix-4bit-128g

SGLang

How to use notstoic/OPT-13B-Nerybus-Mix-4bit-128g 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 "notstoic/OPT-13B-Nerybus-Mix-4bit-128g" \
    --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": "notstoic/OPT-13B-Nerybus-Mix-4bit-128g",
		"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 "notstoic/OPT-13B-Nerybus-Mix-4bit-128g" \
        --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": "notstoic/OPT-13B-Nerybus-Mix-4bit-128g",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use notstoic/OPT-13B-Nerybus-Mix-4bit-128g with Docker Model Runner:
```
docker model run hf.co/notstoic/OPT-13B-Nerybus-Mix-4bit-128g
```

OPT-13B-Nerybus-Mix-4bit-128g

File size: 1,107 Bytes

50f0c26

{
  "_name_or_path": "facebook/opt-13b",
  "_remove_final_layer_norm": false,
  "activation_dropout": 0.0,
  "activation_function": "relu",
  "antemplate": "[Genre: <|>]",
  "architectures": [
    "OPTForCausalLM"
  ],
  "attention_dropout": 0.0,
  "bos_token_id": 2,
  "do_layer_norm_before": true,
  "dropout": 0.1,
  "enable_bias": true,
  "eos_token_id": 2,
  "ffn_dim": 20480,
  "hidden_size": 5120,
  "init_std": 0.02,
  "layer_norm_elementwise_affine": true,
  "layerdrop": 0.0,
  "max_position_embeddings": 2048,
  "model_type": "opt",
  "num_attention_heads": 40,
  "num_hidden_layers": 40,
  "output_projection": true,
  "pad_token_id": 1,
  "prefix": "</s>",
  "torch_dtype": "float16",
  "transformers_version": "4.25.1",
  "use_cache": false,
  "vocab_size": 50265,
  "welcome": "## Warning: This model has a very heavy NSFW bias and is not suitable for use by minors!\n\nYou are currently running the model `OPT-13B-Nerybus-Mix.`\n\nThis is an experimental model containing a parameter-wise 50/50 blend (weighted average) of NerysV2-13B and ErebusV1-13B.\n\n",
  "word_embed_proj_dim": 5120
}