Instructions to use matrixglitch/SD3_prompt-llama_8b with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use matrixglitch/SD3_prompt-llama_8b with Transformers:

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

pipe = pipeline("text-generation", model="matrixglitch/SD3_prompt-llama_8b")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("matrixglitch/SD3_prompt-llama_8b")
model = AutoModelForCausalLM.from_pretrained("matrixglitch/SD3_prompt-llama_8b")
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

vLLM

How to use matrixglitch/SD3_prompt-llama_8b with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "matrixglitch/SD3_prompt-llama_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": "matrixglitch/SD3_prompt-llama_8b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/matrixglitch/SD3_prompt-llama_8b

SGLang

How to use matrixglitch/SD3_prompt-llama_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 "matrixglitch/SD3_prompt-llama_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": "matrixglitch/SD3_prompt-llama_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 "matrixglitch/SD3_prompt-llama_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": "matrixglitch/SD3_prompt-llama_8b",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use matrixglitch/SD3_prompt-llama_8b with Docker Model Runner:
```
docker model run hf.co/matrixglitch/SD3_prompt-llama_8b
```
Browse Quantizations to use this model in llama.cpp, Ollama, LM Studio, or any compatible app.

this finetune is based on llama3 8b and operates under the same license

the model aims to convert non-natural language prompts to natural language automatically while retaining the general idea of the prompt because SD3 performs poorly on non-natural language prompts.

this model was trained for 1500 steps on an RTX 3090 for 2.5 hours, training any longer gave a deminishing result because of the low batch size. the dataset has 90k+ original prompts (non-natural language questions) and 220k modified prompts (natural language answers). it has not been trained on the whole dataset because of lack of compute but the results are already amazing.

Downloads last month: 7

Safetensors

Model size

8B params

Tensor type

BF16

Model tree for matrixglitch/SD3_prompt-llama_8b

Quantizations

1 model