Instructions to use fenyo/Qwen2.5-7B-base2instruct-DPO with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use fenyo/Qwen2.5-7B-base2instruct-DPO with Transformers:

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

pipe = pipeline("text-generation", model="fenyo/Qwen2.5-7B-base2instruct-DPO")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("fenyo/Qwen2.5-7B-base2instruct-DPO")
model = AutoModelForCausalLM.from_pretrained("fenyo/Qwen2.5-7B-base2instruct-DPO")
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 fenyo/Qwen2.5-7B-base2instruct-DPO with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "fenyo/Qwen2.5-7B-base2instruct-DPO"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "fenyo/Qwen2.5-7B-base2instruct-DPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/fenyo/Qwen2.5-7B-base2instruct-DPO

SGLang

How to use fenyo/Qwen2.5-7B-base2instruct-DPO 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 "fenyo/Qwen2.5-7B-base2instruct-DPO" \
    --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": "fenyo/Qwen2.5-7B-base2instruct-DPO",
		"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 "fenyo/Qwen2.5-7B-base2instruct-DPO" \
        --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": "fenyo/Qwen2.5-7B-base2instruct-DPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use fenyo/Qwen2.5-7B-base2instruct-DPO with Docker Model Runner:
```
docker model run hf.co/fenyo/Qwen2.5-7B-base2instruct-DPO
```

Qwen2.5-7B-base2instruct — étape DPO

Checkpoint intermédiaire (SFT + DPO) du pipeline base→instruct SFT → DPO → RLVR appliqué à Qwen/Qwen2.5-7B. Le DPO aligne le modèle SFT sur des préférences humaines (réponses choisies vs rejetées).

⬅️ Étape précédente (SFT) : fenyo/Qwen2.5-7B-base2instruct-SFT
🏁 Modèle final (après RLVR) : fenyo/Qwen2.5-7B-base2instruct
📦 Code & recette : https://github.com/AlexandreFenyo/qwen2.5-7b-base2instruct

Entraînement

DPO LoRA (TRL DPOTrainer, r=32), référence calculée en direct, 1 epoch, lr 5e-6, β=0.1, seq_len 1536, bf16, gradient checkpointing. Adaptateur fusionné dans le modèle. Données : HuggingFaceH4/ultrafeedback_binarized (10k paires).

Piège résolu : precompute_ref_log_probs=True en bf16 produit des NaN — désactivé. Voir le dépôt GitHub.

Résultats (lm-eval, backend vLLM)

étape	IFEval (prompt strict)	GSM8K (flexible)	MMLU
base Qwen2.5-7B	27.4	83.0	71.8
+ SFT	44.9	77.5	69.1
+ DPO (ce modèle)	44.7	77.1	69.9
+ RLVR	45.1	77.4	69.9

Le DPO sur 10k paires génériques bouge peu les benchmarks ici. Analyse complète : dépôt GitHub.

Usage

Format ChatML, via tokenizer.apply_chat_template. Voir l'exemple sur le modèle final.

Downloads last month: 14

Safetensors

Model size

8B params

Tensor type

BF16

Model tree for fenyo/Qwen2.5-7B-base2instruct-DPO

Base model

Qwen/Qwen2.5-7B

Finetuned

(871)

this model

fenyo
/

Qwen2.5-7B-base2instruct-DPO

Qwen2.5-7B-base2instruct — étape DPO

Entraînement

Résultats (lm-eval, backend vLLM)

Usage

Model tree for fenyo/Qwen2.5-7B-base2instruct-DPO

Datasets used to train fenyo/Qwen2.5-7B-base2instruct-DPO