Instructions to use huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507 with Transformers:

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

pipe = pipeline("text-generation", model="huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507", device_map="auto")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507")
model = AutoModelForCausalLM.from_pretrained("huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507", 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]:]))

Inference
Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507

SGLang

How to use huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507 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 "huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507" \
    --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": "huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507",
		"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 "huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507" \
        --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": "huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507 with Docker Model Runner:
```
docker model run hf.co/huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507
```

appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507

This model is a fine-tuned version of Qwen/Qwen3-4B-Instruct-2507 on the appworld_distillation_sft_v2 dataset. It achieves the following results on the evaluation set:

Loss: 0.7486

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: 5e-06
train_batch_size: 4
eval_batch_size: 1
seed: 42
distributed_type: multi-GPU
num_devices: 8
total_train_batch_size: 32
total_eval_batch_size: 8
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: 0.1
num_epochs: 30.0

Training results

Training Loss	Epoch	Step	Validation Loss
1.7313	1.0	2	1.7900
1.6378	2.0	4	1.5365
1.1356	3.0	6	1.1186
0.9041	4.0	8	0.9372
0.8762	5.0	10	0.8529
0.7807	6.0	12	0.8198
0.7323	7.0	14	0.7646
0.6814	8.0	16	0.7229
0.6211	9.0	18	0.6847
0.5738	10.0	20	0.6639
0.5171	11.0	22	0.6499
0.4868	12.0	24	0.6385
0.4371	13.0	26	0.6327
0.407	14.0	28	0.6306
0.3924	15.0	30	0.6330
0.3505	16.0	32	0.6393
0.3339	17.0	34	0.6493
0.3086	18.0	36	0.6623
0.2803	19.0	38	0.6748
0.2687	20.0	40	0.6873
0.25	21.0	42	0.6983
0.2306	22.0	44	0.7100
0.2168	23.0	46	0.7205
0.2125	24.0	48	0.7301
0.2031	25.0	50	0.7378
0.1975	26.0	52	0.7433
0.2001	27.0	54	0.7474
0.1953	28.0	56	0.7487
0.1895	29.0	58	0.7487
0.1976	30.0	60	0.7486

Framework versions

Transformers 4.52.4
Pytorch 2.8.0+cu128
Datasets 3.6.0
Tokenizers 0.21.1

Downloads last month: 9

Safetensors

Model size

4B params

Tensor type

BF16

Model tree for huseyinatahaninan/appworld_distillation_sft_v2-SFT-Qwen3-4B-Instruct-2507

Base model

Qwen/Qwen3-4B-Instruct-2507

Finetuned

(1841)

this model