Instructions to use stage-babylm/llama-256-12L-pairwise with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use stage-babylm/llama-256-12L-pairwise with Transformers:

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

pipe = pipeline("text-generation", model="stage-babylm/llama-256-12L-pairwise")

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("stage-babylm/llama-256-12L-pairwise")
model = AutoModelForCausalLM.from_pretrained("stage-babylm/llama-256-12L-pairwise", device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use stage-babylm/llama-256-12L-pairwise with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "stage-babylm/llama-256-12L-pairwise"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "stage-babylm/llama-256-12L-pairwise",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/stage-babylm/llama-256-12L-pairwise

SGLang

How to use stage-babylm/llama-256-12L-pairwise 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 "stage-babylm/llama-256-12L-pairwise" \
    --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": "stage-babylm/llama-256-12L-pairwise",
		"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 "stage-babylm/llama-256-12L-pairwise" \
        --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": "stage-babylm/llama-256-12L-pairwise",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use stage-babylm/llama-256-12L-pairwise with Docker Model Runner:
```
docker model run hf.co/stage-babylm/llama-256-12L-pairwise
```

llama-256-12L-pairwise

This model is a fine-tuned version of on an unknown dataset. It achieves the following results on the evaluation set:

Loss: 1.7027

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: 0.0018
train_batch_size: 32
eval_batch_size: 32
seed: 42
optimizer: Use OptimizerNames.ADAMW_TORCH_FUSED with betas=(0.9,0.95) and epsilon=1e-06 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine
lr_scheduler_warmup_steps: 0.05
num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss
No log	0	0	6.9752
2.3540	0.0500	1973	2.3841
2.1531	0.1000	3946	2.1744
2.0805	0.1500	5919	2.0891
2.0298	0.2000	7892	2.0497
1.9944	0.2500	9865	2.0164
1.9774	0.3000	11838	1.9913
1.9386	0.3500	13811	1.9674
1.9219	0.4000	15784	1.9453
1.9060	0.4500	17757	1.9251
1.8754	0.5000	19730	1.8996
1.8544	0.5500	21703	1.8780
1.8337	0.6000	23676	1.8502
1.8023	0.6500	25649	1.8250
1.7808	0.7000	27622	1.7986
1.7482	0.7500	29595	1.7729
1.7302	0.8000	31568	1.7482
1.7060	0.8500	33541	1.7268
1.6923	0.9000	35514	1.7119
1.6893	0.9500	37487	1.7042
1.6835	1.0	39459	1.7027

Framework versions

Transformers 5.14.1
Pytorch 2.13.0+cu130
Datasets 5.0.0
Tokenizers 0.22.2

Downloads last month: 703

Safetensors

Model size

9.95M params

Tensor type

F32