Instructions to use nagbhaskar55/slm125mlive-qa-DPO with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use nagbhaskar55/slm125mlive-qa-DPO with Transformers:

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

pipe = pipeline("text-generation", model="nagbhaskar55/slm125mlive-qa-DPO")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("nagbhaskar55/slm125mlive-qa-DPO")
model = AutoModelForCausalLM.from_pretrained("nagbhaskar55/slm125mlive-qa-DPO", 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=256)
print(tokenizer.decode(outputs[0][inputs["input_ids"].shape[-1]:]))

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use nagbhaskar55/slm125mlive-qa-DPO with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "nagbhaskar55/slm125mlive-qa-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": "nagbhaskar55/slm125mlive-qa-DPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/nagbhaskar55/slm125mlive-qa-DPO

SGLang

How to use nagbhaskar55/slm125mlive-qa-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 "nagbhaskar55/slm125mlive-qa-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": "nagbhaskar55/slm125mlive-qa-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 "nagbhaskar55/slm125mlive-qa-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": "nagbhaskar55/slm125mlive-qa-DPO",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use nagbhaskar55/slm125mlive-qa-DPO with Docker Model Runner:
```
docker model run hf.co/nagbhaskar55/slm125mlive-qa-DPO
```

slm125mlive-qa-DPO

DPO applied to nagbhaskar55/slm125mlive-bhaskar-finetune, a QA-SFT model, to sharpen answer quality on grounded (passage + question) prompts.

Results

metric	before	after
preference accuracy (held-out)	0.6	1.0
chosen-vs-rejected logprob margin	56.0583	75.5719

Trained 2 epochs, beta 0.1, 64 steps on 260 pairs (40 held out), in 38s.

How it was trained

DPO (Direct Preference Optimization) optimises the preference objective in closed form against a frozen copy of the starting model. No reward model and no sampling are involved at training time, which is why it is fast and stable.

Preference data

600 triplets (prompt / chosen / rejected), 300 grounded and 300 closed-book. This model trained on the grounded half, because that is the distribution it was supervised-fine-tuned on.

The rejected answer in each pair carries exactly one deliberately induced flaw, drawn evenly from six modes that these models actually exhibit: a wrong figure, an unsupported claim, vagueness, rambling repetition, a partial answer, and a false refusal. An LLM judge then confirmed the ordering with the two answers shown in randomised positions, so it could not score well by always choosing the first. Pairs the judge disagreed with, or was unconfident about, were dropped.

Prompt format

This model expects slm_chat:

<|bos|><|system|>{system}<|eos|><|user|>{passage}\n\n{question}<|eos|><|assistant|>

The format was established empirically, by scoring known-good answers under competing templates and comparing generation behaviour -- not assumed from the tokenizer.

Limits

600 preference pairs is a small budget for preference optimization. Expect sharper formatting, less padding and more consistent refusals -- not new capability. Preference labels come from an LLM judge, so the model inherits that judge's blind spots. Not legal or financial advice.

Downloads last month: 143

Safetensors

Model size

0.1B params

Tensor type

F32

Model tree for nagbhaskar55/slm125mlive-qa-DPO

Base model

nagbhaskar55/slm125mlive-base

Finetuned

nagbhaskar55/slm125mlive-bhaskar-finetune

Finetuned

(2)

this model