Instructions to use hypaai/hypa-tiny-keys with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use hypaai/hypa-tiny-keys with Transformers:

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

pipe = pipeline("text-generation", model="hypaai/hypa-tiny-keys")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("hypaai/hypa-tiny-keys")
model = AutoModelForCausalLM.from_pretrained("hypaai/hypa-tiny-keys", 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 hypaai/hypa-tiny-keys with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/hypaai/hypa-tiny-keys

SGLang

How to use hypaai/hypa-tiny-keys 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 "hypaai/hypa-tiny-keys" \
    --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": "hypaai/hypa-tiny-keys",
		"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 "hypaai/hypa-tiny-keys" \
        --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": "hypaai/hypa-tiny-keys",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use hypaai/hypa-tiny-keys with Docker Model Runner:
```
docker model run hf.co/hypaai/hypa-tiny-keys
```

hypa-tiny-keys

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

Loss: 1.1994

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: 0.0003
train_batch_size: 256
eval_batch_size: 256
seed: 3407
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 500
training_steps: 183476

Training results

Training Loss	Epoch	Step	Validation Loss
2.0830	0.0250	4587	1.5913
1.4610	0.0500	9174	1.5099
1.3978	0.0750	13761	1.4691
1.3647	0.1000	18348	1.4470
1.3525	0.1250	22935	1.4381
1.3103	0.1500	27522	1.4193
1.3251	0.1750	32109	1.4158
1.2983	0.2000	36696	1.4092
1.2916	0.2250	41283	1.4117
1.2891	0.2500	45870	1.3917
1.2785	0.2750	50457	1.3796
1.2719	0.3000	55044	1.3855
1.2648	0.3250	59631	1.3817
1.2662	0.3500	64218	1.3654
1.2446	0.3750	68805	1.3616
1.2584	0.4000	73392	1.3651
1.2331	0.4250	77979	1.3469
1.2437	0.4500	82566	1.3525
1.2234	0.4750	87153	1.3285
1.2318	0.5000	91740	1.3416
1.2186	0.5250	96327	1.3188
1.2173	0.5500	100914	1.3146
1.2145	0.5750	105501	1.3058
1.1944	0.6000	110088	1.3002
1.2025	0.6250	114675	1.2898
1.1773	0.6500	119262	1.2806
1.1972	0.6750	123849	1.2718
1.1623	0.7000	128436	1.2427
1.1875	0.7250	133023	1.2537
1.1675	0.7500	137610	1.2442
1.1576	0.7750	142197	1.2418
1.1725	0.8000	146784	1.2313
1.1483	0.8250	151371	1.2312
1.1634	0.8500	155958	1.2177
1.1455	0.8750	160545	1.2165
1.1637	0.9000	165132	1.2105
1.1420	0.9250	169719	1.2045
1.1645	0.9500	174306	1.2024
1.1522	0.9750	178893	1.2000
1.1522	1.0	183476	1.1994

Framework versions

Transformers 5.17.0
Pytorch 2.8.0+cu128
Datasets 5.0.1
Tokenizers 0.23.2

Downloads last month: 8,863

Safetensors

Model size

14.5M params

Tensor type

F32