Instructions to use cstr/phi-3-orpo-v9_16 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use cstr/phi-3-orpo-v9_16 with Transformers:

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

pipe = pipeline("text-generation", model="cstr/phi-3-orpo-v9_16")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("cstr/phi-3-orpo-v9_16")
model = AutoModelForCausalLM.from_pretrained("cstr/phi-3-orpo-v9_16", 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 cstr/phi-3-orpo-v9_16 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "cstr/phi-3-orpo-v9_16"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "cstr/phi-3-orpo-v9_16",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/cstr/phi-3-orpo-v9_16

SGLang

How to use cstr/phi-3-orpo-v9_16 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 "cstr/phi-3-orpo-v9_16" \
    --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": "cstr/phi-3-orpo-v9_16",
		"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 "cstr/phi-3-orpo-v9_16" \
        --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": "cstr/phi-3-orpo-v9_16",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Unsloth Studio

How to use cstr/phi-3-orpo-v9_16 with Unsloth Studio:

Install Unsloth Studio (macOS, Linux, WSL)

curl -fsSL https://unsloth.ai/install.sh | sh
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for cstr/phi-3-orpo-v9_16 to start chatting

Install Unsloth Studio (Windows)

irm https://unsloth.ai/install.ps1 | iex
# Run unsloth studio
unsloth studio -H 0.0.0.0 -p 8888
# Then open http://localhost:8888 in your browser
# Search for cstr/phi-3-orpo-v9_16 to start chatting

Using HuggingFace Spaces for Unsloth

# No setup required
# Open https://huggingface.co/spaces/unsloth/studio in your browser
# Search for cstr/phi-3-orpo-v9_16 to start chatting

Load model with FastModel

pip install unsloth
from unsloth import FastModel
model, tokenizer = FastModel.from_pretrained(
    model_name="cstr/phi-3-orpo-v9_16",
    max_seq_length=2048,
)

Docker Model Runner
How to use cstr/phi-3-orpo-v9_16 with Docker Model Runner:
```
docker model run hf.co/cstr/phi-3-orpo-v9_16
```

Model details

This is a quick experiment on llamafied phi-3 with only 1000 orpo steps from an azureml translated german orca binarized-dataset (johannhartmann/mistralorpo), with original phi-3 prompt template. The immediate result is not really good, but also not bad enough to disencourage further experiments.

Benchmark results

This was an experiment on a german dataset snippet which, as expected, worsened results on english benchmarks:

Metric	Value
Avg.	64.40
AI2 Reasoning Challenge (25-Shot)	60.41
HellaSwag (10-Shot)	78.37
MMLU (5-Shot)	65.26
TruthfulQA (0-shot)	49.76
Winogrande (5-shot)	70.24
GSM8k (5-shot)	62.32

On german EQ-Bench (v2_de) 51.82 (insignificant over 51.41 for original llamafied but significantly better than intermediate cstr/phi-3-orpo-v8_16 which after initial 150 test steps achieved 46.38) but with still only 164/171 correctly parsed.

Note: We can improve the correctness of parsing, i.a., by only a few SFT steps, as shown with cas/phi3-mini-4k-llamafied-sft-v3 (170/171 correct but with then only 39.46 score in v2_de, which was also an experiment in changing the prompt template). All that was quickly done with bnb and q4 quants only, which might, in theory, affect especially such small dense models significantly. But it served the intention for both proof-of-concept-experiments at least. Probably it would easily be possible to further improve results, but that would take some time and compute.

Training setup

This llama model was trained 2x faster with Unsloth and Huggingface's TRL library.

EU AI Act Art. 53 — provider obligations

Added 2026-08-02 during an account-wide provenance review.

This is a fine-tune, not a format conversion. Most cstr/* repositories are GGUF conversions where the upstream research team remains the provider of the model and conversion changes only the numeric representation of the weights. Training changes the model itself, so under Regulation (EU) 2024/1689 the maintainer of this repository is plausibly the provider of the model it produced, and the obligations that survive the Art. 53(2) free-and-open-source exemption — Art. 53(1)(c) and 53(1)(d) — attach here.

Art. 53(1)(c) — copyright policy. The ORPO training used johannhartmann/mistralorpo, an AzureML-translated German binarized Orca preference set, as named in the card above. Questions about the provenance of that dataset's content attach to its publisher. No text or data mining was performed to assemble a corpus for this repository, so no rights reservation under Art. 4(3) of Directive (EU) 2019/790 was engaged by this step. Any credible claim that this model was trained on material without the necessary rights will be acted on — contact via the Community tab.

Art. 53(1)(d) — training content. Approximately 1000 ORPO steps on johannhartmann/mistralorpo (AzureML-translated German binarized Orca preference data), applied to cstr/phi-3-orpo-v8_16. No other corpus was introduced. The remaining training content is inherited through that model from Microsoft's Phi-3-mini and is documented by its provider.

Base model. cstr/phi-3-orpo-v8_16 — the training built on those weights and inherits their terms.

Downloads last month: 27

Model tree for cstr/phi-3-orpo-v9_16

Base model

vonjack/Phi-3-mini-4k-instruct-LLaMAfied

Finetuned

cstr/phi-3-orpo-v8_16

Finetuned

(1)

this model