Instructions to use formalmathatepfl/apertus-cpt-sft-feedback with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use formalmathatepfl/apertus-cpt-sft-feedback with Transformers:

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

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

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("formalmathatepfl/apertus-cpt-sft-feedback")
model = AutoModelForCausalLM.from_pretrained("formalmathatepfl/apertus-cpt-sft-feedback")
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-cpt-sft-feedback with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "formalmathatepfl/apertus-cpt-sft-feedback"
# 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-cpt-sft-feedback",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/formalmathatepfl/apertus-cpt-sft-feedback

SGLang

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

Docker Model Runner
How to use formalmathatepfl/apertus-cpt-sft-feedback with Docker Model Runner:
```
docker model run hf.co/formalmathatepfl/apertus-cpt-sft-feedback
```

Apertus-feedback

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

Loss: 0.0303

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: 1e-05
train_batch_size: 2
eval_batch_size: 2
seed: 42
distributed_type: multi-GPU
num_devices: 7
total_train_batch_size: 14
total_eval_batch_size: 14
optimizer: Use OptimizerNames.ADAMW_TORCH 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: 0.05
num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss
0.0609	0.0777	1000	0.0622
0.0428	0.1553	2000	0.0455
0.04	0.2330	3000	0.0402
0.0349	0.3107	4000	0.0371
0.0384	0.3883	5000	0.0352
0.0319	0.4660	6000	0.0337
0.0313	0.5437	7000	0.0326
0.0312	0.6214	8000	0.0318
0.0307	0.6990	9000	0.0312
0.0303	0.7767	10000	0.0308
0.0283	0.8544	11000	0.0305
0.0279	0.9320	12000	0.0304

Framework versions

Transformers 4.57.6
Pytorch 2.9.0+cu128
Datasets 4.0.0
Tokenizers 0.22.2

Downloads last month: 30

Safetensors

Model size

274k params

Tensor type

BF16

Model tree for formalmathatepfl/apertus-cpt-sft-feedback

Base model

swiss-ai/Apertus-8B-2509

Finetuned

(18)

this model