Instructions to use jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale with Transformers:

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

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

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale")
model = AutoModelForCausalLM.from_pretrained("jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale", 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 jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale

SGLang

How to use jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale 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 "jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale" \
    --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": "jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale",
		"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 "jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale" \
        --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": "jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale with Docker Model Runner:
```
docker model run hf.co/jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale
```

instruct_contrastive

This model is a fine-tuned version of jnsffrt/instruct_open_answer on an unknown dataset. It achieves the following results on the evaluation set:

Loss: -13.6334

Model description

More information needed

Intended uses & limitations

More information needed

Training and evaluation data

More information needed

Training procedure

Training hyperparameters

The following hyperparameters were used during training:

learning_rate: 1e-05
train_batch_size: 4
eval_batch_size: 4
seed: 552
gradient_accumulation_steps: 4
total_train_batch_size: 16
optimizer: Use adamw_torch with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_ratio: 0.05
num_epochs: 2

Training results

Training Loss	Epoch	Step	Validation Loss
-4.6212	0.0547	250	-1.3475
-27.3747	0.1094	500	-7.0080
-33.0028	0.1641	750	-8.7426
-51.24	0.2188	1000	-9.9981
-49.2631	0.2735	1250	-11.1210
-49.9911	0.3282	1500	-11.4823
-47.4685	0.3829	1750	-11.5264
-49.8516	0.4376	2000	-11.7210
-52.2479	0.4923	2250	-13.2389
-41.6111	0.5470	2500	-13.4486
-49.9071	0.6017	2750	-11.4862
-50.7067	0.6564	3000	-12.0170
-50.125	0.7111	3250	-13.8635
-64.6233	0.7658	3500	-12.3017
-62.0945	0.8205	3750	-13.0986
-51.1444	0.8752	4000	-13.6334

Framework versions

Transformers 4.52.3
Pytorch 2.7.0+cu126
Datasets 3.6.0
Tokenizers 0.21.1

Downloads last month: 5

Safetensors

Model size

0.6B params

Tensor type

BF16

Model tree for jnsffrt/instruct_mcq_contrastive_full_answer_no_rationale

Base model

Qwen/Qwen3-0.6B-Base

Finetuned

jnsffrt/base_instruct_open_answer

Finetuned

(2)

this model