Instructions to use NewstaR/7B-Orfini with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use NewstaR/7B-Orfini with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="NewstaR/7B-Orfini")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("NewstaR/7B-Orfini") model = AutoModelForCausalLM.from_pretrained("NewstaR/7B-Orfini", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use NewstaR/7B-Orfini with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "NewstaR/7B-Orfini" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "NewstaR/7B-Orfini", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/NewstaR/7B-Orfini
- SGLang
How to use NewstaR/7B-Orfini 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 "NewstaR/7B-Orfini" \ --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": "NewstaR/7B-Orfini", "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 "NewstaR/7B-Orfini" \ --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": "NewstaR/7B-Orfini", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use NewstaR/7B-Orfini with Docker Model Runner:
docker model run hf.co/NewstaR/7B-Orfini
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,3 +1,63 @@
|
|
| 1 |
---
|
| 2 |
license: mit
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 3 |
---
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
---
|
| 2 |
license: mit
|
| 3 |
+
datasets:
|
| 4 |
+
- Open-Orca/OpenOrca
|
| 5 |
+
- conceptofmind/cot_submix_original
|
| 6 |
+
- conceptofmind/t0_submix_original
|
| 7 |
+
- conceptofmind/niv2_submix_original
|
| 8 |
+
- conceptofmind/flan2021_submix_original
|
| 9 |
+
- ehartford/dolphin
|
| 10 |
+
language:
|
| 11 |
+
- en
|
| 12 |
+
tags:
|
| 13 |
+
- merge
|
| 14 |
+
- slerp
|
| 15 |
+
inference: false
|
| 16 |
+
metrics:
|
| 17 |
+
- accuracy
|
| 18 |
+
- bleu
|
| 19 |
---
|
| 20 |
+
<h1 style="text-align: center">Orfini</h1>
|
| 21 |
+
<h2 style="text-align: center">An experimental model</h2>
|
| 22 |
+
<hr>
|
| 23 |
+
|
| 24 |
+
|
| 25 |
+
## Model Details
|
| 26 |
+
Orfini is an experimental merged model created from the following three foundation models:
|
| 27 |
+
|
| 28 |
+
- stabilityai/StableBeluga-7B
|
| 29 |
+
- pankajmathur/orca_mini_v3_7b
|
| 30 |
+
- AIDC-ai-business/Marcoroni-7B
|
| 31 |
+
|
| 32 |
+
Orfini was created by merging the weights and architectures of these three models using a custom merging technique. No further fine-tuning was performed after the merge.
|
| 33 |
+
|
| 34 |
+
Once the model obtains it's evaluation scores, then we'll know if it works or not.
|
| 35 |
+
|
| 36 |
+
## Intended Use
|
| 37 |
+
As an experimental model, Orfini is intended for testing and research purposes only. It should not be used for production systems or to generate content for public use.
|
| 38 |
+
|
| 39 |
+
## Training Data
|
| 40 |
+
Orfini inherits training data from its three foundation models:
|
| 41 |
+
|
| 42 |
+
- StableBeluga-7B: COT, Niv2, t0, & FLAN2021
|
| 43 |
+
- dolphin-llama2-7b: Dolphin
|
| 44 |
+
- Marcoroni-7B: OpenOrca
|
| 45 |
+
|
| 46 |
+
## Limitations
|
| 47 |
+
As an untested merged model, Orfini has unknown capabilities and limitations. Potential issues include:
|
| 48 |
+
|
| 49 |
+
- Instability due to merged architectures
|
| 50 |
+
- Compounded bias and issues from all three foundation models
|
| 51 |
+
- Decreased performance on some tasks compared to the foundation models
|
| 52 |
+
|
| 53 |
+
Extensive testing is required to characterize Orfini's capabilities and limitations.
|
| 54 |
+
|
| 55 |
+
## Ethical Considerations
|
| 56 |
+
- Orfini may exhibit harmful biases inherited from its training data
|
| 57 |
+
- Output may be unreliable or manipulated due to instability
|
| 58 |
+
- Experimental nature increases potential for misuse
|
| 59 |
+
|
| 60 |
+
Use this model ethically and do not deploy it for sensitive applications.
|
| 61 |
+
|
| 62 |
+
## Contact Information
|
| 63 |
+
Please report issues or concerns with this model to the creator for further investigation.
|