Instructions to use KnutJaegersberg/Galpaca-30b-MiniOrca with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use KnutJaegersberg/Galpaca-30b-MiniOrca with Transformers:

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

pipe = pipeline("text-generation", model="KnutJaegersberg/Galpaca-30b-MiniOrca")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("KnutJaegersberg/Galpaca-30b-MiniOrca")
model = AutoModelForCausalLM.from_pretrained("KnutJaegersberg/Galpaca-30b-MiniOrca")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use KnutJaegersberg/Galpaca-30b-MiniOrca with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "KnutJaegersberg/Galpaca-30b-MiniOrca"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "KnutJaegersberg/Galpaca-30b-MiniOrca",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/KnutJaegersberg/Galpaca-30b-MiniOrca

SGLang

How to use KnutJaegersberg/Galpaca-30b-MiniOrca 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 "KnutJaegersberg/Galpaca-30b-MiniOrca" \
    --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": "KnutJaegersberg/Galpaca-30b-MiniOrca",
		"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 "KnutJaegersberg/Galpaca-30b-MiniOrca" \
        --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": "KnutJaegersberg/Galpaca-30b-MiniOrca",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use KnutJaegersberg/Galpaca-30b-MiniOrca with Docker Model Runner:
```
docker model run hf.co/KnutJaegersberg/Galpaca-30b-MiniOrca
```

Galpaca trained for 2.7 epochs on the 50k shortest records of miniorca dataset with NEFTune.

Prompt Example:

### System:
You are an AI assistant. You will be given a task. You must generate a detailed and long answer.

### User:
What is AGI?

### Assistant:

Open LLM Leaderboard Evaluation Results

Detailed results can be found here

Metric	Value
Avg.	42.23
AI2 Reasoning Challenge (25-Shot)	48.89
HellaSwag (10-Shot)	57.80
MMLU (5-Shot)	43.72
TruthfulQA (0-shot)	41.10
Winogrande (5-shot)	60.06
GSM8k (5-shot)	1.82

Downloads last month: 88

Safetensors

Model size

30B params

Tensor type

F16

Dataset used to train KnutJaegersberg/Galpaca-30b-MiniOrca

Evaluation results

normalized accuracy on AI2 Reasoning Challenge (25-Shot)
test set Open LLM Leaderboard

48.890
normalized accuracy on HellaSwag (10-Shot)
validation set Open LLM Leaderboard

57.800
accuracy on MMLU (5-Shot)
test set Open LLM Leaderboard

43.720
mc2 on TruthfulQA (0-shot)
validation set Open LLM Leaderboard

41.100
accuracy on Winogrande (5-shot)
validation set Open LLM Leaderboard

60.060
accuracy on GSM8k (5-shot)
test set Open LLM Leaderboard

1.820