Instructions to use danielkosyra/pretraining5 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use danielkosyra/pretraining5 with Transformers:

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

pipe = pipeline("text-generation", model="danielkosyra/pretraining5")

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

tokenizer = AutoTokenizer.from_pretrained("danielkosyra/pretraining5")
model = AutoModelForCausalLM.from_pretrained("danielkosyra/pretraining5", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use danielkosyra/pretraining5 with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "danielkosyra/pretraining5"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "danielkosyra/pretraining5",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/danielkosyra/pretraining5

SGLang

How to use danielkosyra/pretraining5 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 "danielkosyra/pretraining5" \
    --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": "danielkosyra/pretraining5",
		"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 "danielkosyra/pretraining5" \
        --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": "danielkosyra/pretraining5",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use danielkosyra/pretraining5 with Docker Model Runner:
```
docker model run hf.co/danielkosyra/pretraining5
```

pretraining5

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

Loss: 2.8364

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: 0.0006
train_batch_size: 32
eval_batch_size: 32
seed: 42
gradient_accumulation_steps: 10
total_train_batch_size: 320
optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
lr_scheduler_type: polynomial
lr_scheduler_warmup_steps: 250
training_steps: 2500
mixed_precision_training: Native AMP

Training results

Training Loss	Epoch	Step	Validation Loss
9.1525	0.1830	50	7.3936
6.7279	0.3660	100	6.2230
5.9675	0.5490	150	5.6680
5.4986	0.7321	200	5.2719
5.135	0.9151	250	4.9299
4.7982	1.0981	300	4.6292
4.4937	1.2811	350	4.3658
4.2185	1.4641	400	4.0518
3.9681	1.6471	450	3.7770
3.76	1.8302	500	3.6076
3.5945	2.0132	550	3.4863
3.389	2.1962	600	3.4087
3.3484	2.3792	650	3.3396
3.2843	2.5622	700	3.2784
3.2286	2.7452	750	3.2334
3.1736	2.9283	800	3.1845
3.0711	3.1113	850	3.1490
2.9997	3.2943	900	3.1204
2.9726	3.4773	950	3.0938
2.9603	3.6603	1000	3.0655
2.9585	3.8433	1050	3.0443
2.916	4.0264	1100	3.0194
2.7717	4.2094	1150	3.0058
2.7737	4.3924	1200	2.9916
2.7665	4.5754	1250	2.9743
2.7632	4.7584	1300	2.9588
2.7579	4.9414	1350	2.9429
2.642	5.1245	1400	2.9346
2.6045	5.3075	1450	2.9260
2.6107	5.4905	1500	2.9138
2.6073	5.6735	1550	2.9052
2.6015	5.8565	1600	2.8938
2.5693	6.0395	1650	2.8881
2.4441	6.2225	1700	2.8888
2.4742	6.4056	1750	2.8801
2.4652	6.5886	1800	2.8728
2.47	6.7716	1850	2.8639
2.4721	6.9546	1900	2.8525
2.3633	7.1376	1950	2.8629
2.3339	7.3206	2000	2.8579
2.3426	7.5037	2050	2.8542
2.3494	7.6867	2100	2.8483
2.3555	7.8697	2150	2.8431
2.3068	8.0527	2200	2.8460
2.2349	8.2357	2250	2.8463
2.2399	8.4187	2300	2.8441
2.243	8.6018	2350	2.8406
2.24	8.7848	2400	2.8379
2.236	8.9678	2450	2.8344
2.1823	9.1508	2500	2.8364

Framework versions

Transformers 4.40.1
Pytorch 2.3.0+cu121
Datasets 2.19.0
Tokenizers 0.19.1

Downloads last month: 8

Safetensors

Model size

0.1B params

Tensor type

F32

Model tree for danielkosyra/pretraining5

Base model

openai-community/gpt2

Finetuned

(2331)

this model