Instructions to use W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614 with Transformers:

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

pipe = pipeline("text-generation", model="W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614")
model = AutoModelForCausalLM.from_pretrained("W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614", 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 W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614

SGLang

How to use W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614 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 "W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614" \
    --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": "W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614",
		"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 "W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614" \
        --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": "W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614 with Docker Model Runner:
```
docker model run hf.co/W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614
```

qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614

This model is a fine-tuned version of jackf857/qwen3-8b-base-sft-ultrachat-4xh200-batch-128 on the HuggingFaceH4/ultrafeedback_binarized dataset. It achieves the following results on the evaluation set:

Loss: 0.4315
Rewards/chosen: -0.6019
Logps/chosen: -340.639
Rewards/rejected: -1.3436
Logps/rejected: -399.0197
Rewards/margins: 0.7417
Kl: 0.0
Logits/chosen: 145031824.0
Logits/rejected: 149160000.0

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-07
train_batch_size: 4
eval_batch_size: 2
seed: 42
distributed_type: multi-GPU
num_devices: 4
gradient_accumulation_steps: 8
total_train_batch_size: 128
total_eval_batch_size: 8
optimizer: Use OptimizerNames.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: 1

Training results

Training Loss	Epoch	Step	Validation Loss	Rewards/chosen	Logps/chosen	Rewards/rejected	Logps/rejected	Rewards/margins	Logits/chosen	Logits/rejected
3.8868	0.2094	200	0.4847	0.0356	-276.8863	-0.0882	-273.4840	0.1239	173188400.0	171871392.0
3.45	0.4188	400	0.4441	-0.6119	-341.6399	-1.1696	-381.6226	0.5577	138682288.0	141790496.0
3.4128	0.6282	600	0.4341	-0.6728	-347.7317	-1.3754	-402.2034	0.7026	138811552.0	142210000.0
3.4942	0.8376	800	0.4315	-0.6019	-340.639	-1.3436	-399.0197	0.7417	145031824.0	149160000.0

Framework versions

Transformers 4.51.0
Pytorch 2.3.1+cu121
Datasets 2.21.0
Tokenizers 0.21.4

Downloads last month: 7

Safetensors

Model size

8B params

Tensor type

F32

Model tree for W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614

Base model

Qwen/Qwen3-8B-Base

Finetuned

jackf857/qwen3-8b-base-sft-ultrachat-4xh200-batch-128

Finetuned

(30)

this model

W-61
/

qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614

qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614

Model description

Intended uses & limitations

Training and evaluation data

Training procedure

Training hyperparameters

Training results

Framework versions

Model tree for W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614

Dataset used to train W-61/qwen3-8b-base-kto-ultrafeedback-4xh200-batch-128-20260426-105614