Instructions to use Lilin004/luxun-gpt2-optimized2 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use Lilin004/luxun-gpt2-optimized2 with Transformers:

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

pipe = pipeline("text-generation", model="Lilin004/luxun-gpt2-optimized2")

# pip install -U transformers accelerate
# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("Lilin004/luxun-gpt2-optimized2")
model = AutoModelForCausalLM.from_pretrained("Lilin004/luxun-gpt2-optimized2", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use Lilin004/luxun-gpt2-optimized2 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "Lilin004/luxun-gpt2-optimized2"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Lilin004/luxun-gpt2-optimized2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/Lilin004/luxun-gpt2-optimized2

SGLang

How to use Lilin004/luxun-gpt2-optimized2 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 "Lilin004/luxun-gpt2-optimized2" \
    --host 0.0.0.0 \
    --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Lilin004/luxun-gpt2-optimized2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

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 "Lilin004/luxun-gpt2-optimized2" \
        --host 0.0.0.0 \
        --port 30000
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:30000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "Lilin004/luxun-gpt2-optimized2",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use Lilin004/luxun-gpt2-optimized2 with Docker Model Runner:
```
docker model run hf.co/Lilin004/luxun-gpt2-optimized2
```

luxun-gpt2-optimized2

This model is a fine-tuned version of uer/gpt2-chinese-cluecorpussmall on the None dataset. It achieves the following results on the evaluation set:

Loss: 2.7527

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-05
train_batch_size: 4
eval_batch_size: 8
seed: 42
gradient_accumulation_steps: 4
total_train_batch_size: 16
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.999) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: linear
lr_scheduler_warmup_steps: 0.1
num_epochs: 10
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
21.3680	0.1361	20	4.9754
21.1391	0.2721	40	4.7909
19.4552	0.4082	60	4.6759
19.1040	0.5442	80	4.5777
19.4053	0.6803	100	4.4904
18.4277	0.8163	120	4.4121
18.2665	0.9524	140	4.3433
17.9532	1.0884	160	4.2720
17.4830	1.2245	180	4.2146
17.1754	1.3605	200	4.1561
17.0501	1.4966	220	4.0919
16.7780	1.6327	240	4.0369
16.6452	1.7687	260	3.9848
16.4966	1.9048	280	3.9273
15.5722	2.0408	300	3.8965
15.8068	2.1769	320	3.8402
15.7545	2.3129	340	3.8010
15.8805	2.4490	360	3.7666
15.0228	2.5850	380	3.7225
14.9995	2.7211	400	3.6866
14.4941	2.8571	420	3.6518
14.9387	2.9932	440	3.6127
14.3115	3.1293	460	3.5830
14.3165	3.2653	480	3.5469
14.0445	3.4014	500	3.5137
13.7494	3.5374	520	3.4837
13.7659	3.6735	540	3.4434
13.9861	3.8095	560	3.4142
14.1536	3.9456	580	3.3838
13.1904	4.0816	600	3.3558
12.8714	4.2177	620	3.3279
13.3145	4.3537	640	3.3033
13.2822	4.4898	660	3.2687
12.6844	4.6259	680	3.2473
12.7821	4.7619	700	3.2254
12.8397	4.8980	720	3.1974
12.0735	5.0340	740	3.1780
12.0329	5.1701	760	3.1471
12.2971	5.3061	780	3.1263
12.0071	5.4422	800	3.1036
12.0631	5.5782	820	3.0795
11.7063	5.7143	840	3.0671
11.8699	5.8503	860	3.0479
11.9551	5.9864	880	3.0283
11.7066	6.1224	900	3.0070
11.6296	6.2585	920	2.9910
11.2781	6.3946	940	2.9774
11.4764	6.5306	960	2.9600
11.6617	6.6667	980	2.9460
11.2081	6.8027	1000	2.9248
10.9386	6.9388	1020	2.9123
11.4737	7.0748	1040	2.8967
10.6856	7.2109	1060	2.8874
10.6605	7.3469	1080	2.8752
10.6367	7.4830	1100	2.8627
11.3121	7.6190	1120	2.8528
10.8436	7.7551	1140	2.8385
10.5651	7.8912	1160	2.8303
10.3416	8.0272	1180	2.8211
10.7753	8.1633	1200	2.8104
10.5844	8.2993	1220	2.8035
10.2085	8.4354	1240	2.7933
10.4647	8.5714	1260	2.7872
10.4550	8.7075	1280	2.7822
10.7316	8.8435	1300	2.7765
10.3152	8.9796	1320	2.7728
10.3639	9.1156	1340	2.7677
10.3007	9.2517	1360	2.7638
10.4514	9.3878	1380	2.7591
10.4867	9.5238	1400	2.7576
10.1711	9.6599	1420	2.7557
10.2853	9.7959	1440	2.7538
10.5973	9.9320	1460	2.7527

Framework versions

Transformers 5.0.0
Pytorch 2.10.0+cu128
Datasets 4.0.0
Tokenizers 0.22.2

Downloads last month: 9

Safetensors

Model size

0.1B params

Tensor type

F32

Model tree for Lilin004/luxun-gpt2-optimized2

Base model

uer/gpt2-chinese-cluecorpussmall

Finetuned

(7)

this model