Instructions to use StudioDesign/sft-qwen3 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use StudioDesign/sft-qwen3 with PEFT:

from peft import PeftModel
from transformers import AutoModelForCausalLM

base_model = AutoModelForCausalLM.from_pretrained("/root/.cache/huggingface/hub/models--Qwen--Qwen3-8B/snapshots/b968826d9c46dd6066d109eabc6255188de91218")
model = PeftModel.from_pretrained(base_model, "StudioDesign/sft-qwen3")

Transformers

How to use StudioDesign/sft-qwen3 with Transformers:

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

pipe = pipeline("text-generation", model="StudioDesign/sft-qwen3")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModel
model = AutoModel.from_pretrained("StudioDesign/sft-qwen3", dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use StudioDesign/sft-qwen3 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "StudioDesign/sft-qwen3"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "StudioDesign/sft-qwen3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/StudioDesign/sft-qwen3

SGLang

How to use StudioDesign/sft-qwen3 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 "StudioDesign/sft-qwen3" \
    --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": "StudioDesign/sft-qwen3",
		"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 "StudioDesign/sft-qwen3" \
        --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": "StudioDesign/sft-qwen3",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use StudioDesign/sft-qwen3 with Docker Model Runner:
```
docker model run hf.co/StudioDesign/sft-qwen3
```

YuYamada commited on Mar 17

Commit

ad16856

verified ·

1 Parent(s): 4faf6ef

Training in progress, epoch 1

Browse files

Files changed (1) hide show

logging.jsonl +2 -0

logging.jsonl CHANGED Viewed

@@ -1,2 +1,4 @@
 {"loss": 0.32817921, "grad_norm": 0.15581589, "learning_rate": 1e-05, "token_acc": 0.92229585, "epoch": 1.0, "global_step/max_steps": "1/1", "percentage": "100.00%", "elapsed_time": "11s", "remaining_time": "0s", "memory(GiB)": 22.22, "train_speed(iter/s)": 0.086502}
 {"eval_loss": 0.22503923, "eval_runtime": 3.86, "eval_samples_per_second": 0.518, "eval_steps_per_second": 0.259, "eval_token_acc": 0.94509159, "epoch": 1.0, "global_step/max_steps": "1/1", "percentage": "100.00%", "elapsed_time": "15s", "remaining_time": "0s", "memory(GiB)": 39.15, "train_speed(iter/s)": 0.064698}

 {"loss": 0.32817921, "grad_norm": 0.15581589, "learning_rate": 1e-05, "token_acc": 0.92229585, "epoch": 1.0, "global_step/max_steps": "1/1", "percentage": "100.00%", "elapsed_time": "11s", "remaining_time": "0s", "memory(GiB)": 22.22, "train_speed(iter/s)": 0.086502}
 {"eval_loss": 0.22503923, "eval_runtime": 3.86, "eval_samples_per_second": 0.518, "eval_steps_per_second": 0.259, "eval_token_acc": 0.94509159, "epoch": 1.0, "global_step/max_steps": "1/1", "percentage": "100.00%", "elapsed_time": "15s", "remaining_time": "0s", "memory(GiB)": 39.15, "train_speed(iter/s)": 0.064698}
+{"eval_loss": 0.22503923, "eval_runtime": 3.7326, "eval_samples_per_second": 0.536, "eval_steps_per_second": 0.268, "eval_token_acc": 0.94509159, "epoch": 1.0, "global_step/max_steps": "1/1", "percentage": "100.00%", "elapsed_time": "21s", "remaining_time": "0s", "memory(GiB)": 39.15, "train_speed(iter/s)": 0.046011}
+{"train_runtime": 26.378, "train_samples_per_second": 0.341, "train_steps_per_second": 0.038, "total_flos": 10981327503360.0, "train_loss": 0.32817921, "epoch": 1.0, "global_step/max_steps": "1/1", "percentage": "100.00%", "elapsed_time": "24s", "remaining_time": "0s", "memory(GiB)": 39.15, "train_speed(iter/s)": 0.041489}