qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128

Instructions to use seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128 with Transformers:

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

pipe = pipeline("text-generation", model="seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128")
model = AutoModelForCausalLM.from_pretrained("seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128", 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 seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128

SGLang

How to use seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128 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 "seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128" \
    --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": "seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128",
		"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 "seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128" \
        --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": "seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128 with Docker Model Runner:
```
docker model run hf.co/seanyhan/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128
```

qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128 / all_results.json

seanyhan

Duplicate from jackf857/qwen3-8b-base-simpo-ultrafeedback-4xH200-batch-128

345b7c8 3 months ago

Raw

History Blame Contribute Delete

782 Bytes

	{
	"epoch": 0.9989528795811519,
	"eval_logits/chosen": 2.1443471908569336,
	"eval_logits/rejected": 2.1642587184906006,
	"eval_logps/chosen": -1.0595875978469849,
	"eval_logps/rejected": -1.4812482595443726,
	"eval_loss": 1.0770305395126343,
	"eval_rewards/accuracies": 0.6660000085830688,
	"eval_rewards/chosen": -2.1191751956939697,
	"eval_rewards/margins": 0.8433210253715515,
	"eval_rewards/rejected": -2.962496519088745,
	"eval_runtime": 45.3759,
	"eval_samples": 2000,
	"eval_samples_per_second": 44.076,
	"eval_steps_per_second": 5.51,
	"total_flos": 0.0,
	"train_loss": 9.188309121681709,
	"train_runtime": 5401.7608,
	"train_samples": 61135,
	"train_samples_per_second": 11.318,
	"train_steps_per_second": 0.088
	}