Instructions to use formalmathatepfl/Apertus-8B-cpt with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use formalmathatepfl/Apertus-8B-cpt with Transformers:

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

pipe = pipeline("text-generation", model="formalmathatepfl/Apertus-8B-cpt")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("formalmathatepfl/Apertus-8B-cpt")
model = AutoModelForCausalLM.from_pretrained("formalmathatepfl/Apertus-8B-cpt")
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 formalmathatepfl/Apertus-8B-cpt with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/formalmathatepfl/Apertus-8B-cpt

SGLang

How to use formalmathatepfl/Apertus-8B-cpt 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 "formalmathatepfl/Apertus-8B-cpt" \
    --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": "formalmathatepfl/Apertus-8B-cpt",
		"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 "formalmathatepfl/Apertus-8B-cpt" \
        --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": "formalmathatepfl/Apertus-8B-cpt",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use formalmathatepfl/Apertus-8B-cpt with Docker Model Runner:
```
docker model run hf.co/formalmathatepfl/Apertus-8B-cpt
```

Apertus-8B-cpt

This model is a fine-tuned version of /mnt/task_runtime/models/Apertus-8B on the cpt_data dataset. It achieves the following results on the evaluation set:

Loss: 0.9989

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: 3e-06
train_batch_size: 4
eval_batch_size: 4
seed: 42
distributed_type: multi-GPU
num_devices: 8
total_train_batch_size: 32
total_eval_batch_size: 32
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: cosine
lr_scheduler_warmup_ratio: 0.03
num_epochs: 2.0

Training results

Training Loss	Epoch	Step	Validation Loss
1.2574	0.1311	100	1.2728
1.1402	0.2621	200	1.1717
1.1055	0.3932	300	1.1226
1.0727	0.5242	400	1.0898
1.0481	0.6553	500	1.0660
1.0351	0.7864	600	1.0474
0.977	0.9174	700	1.0336
0.9933	1.0485	800	1.0230
0.9646	1.1796	900	1.0147
0.945	1.3106	1000	1.0083
1.0034	1.4417	1100	1.0039
0.9788	1.5727	1200	1.0012
0.9938	1.7038	1300	0.9997
1.007	1.8349	1400	0.9991
1.0129	1.9659	1500	0.9989

Framework versions

Transformers 4.57.6
Pytorch 2.9.0+cu128
Datasets 4.0.0
Tokenizers 0.22.2

Downloads last month: 3

Safetensors

Model size

274k params

Tensor type

BF16