Instructions to use zeng123/Depth-attention-1.5B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use zeng123/Depth-attention-1.5B with Transformers:

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

pipe = pipeline("text-generation", model="zeng123/Depth-attention-1.5B")
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoTokenizer, AutoModelForMultimodalLM

tokenizer = AutoTokenizer.from_pretrained("zeng123/Depth-attention-1.5B")
model = AutoModelForMultimodalLM.from_pretrained("zeng123/Depth-attention-1.5B")
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 zeng123/Depth-attention-1.5B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "zeng123/Depth-attention-1.5B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/chat/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "zeng123/Depth-attention-1.5B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Use Docker

docker model run hf.co/zeng123/Depth-attention-1.5B

SGLang

How to use zeng123/Depth-attention-1.5B 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 "zeng123/Depth-attention-1.5B" \
    --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": "zeng123/Depth-attention-1.5B",
		"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 "zeng123/Depth-attention-1.5B" \
        --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": "zeng123/Depth-attention-1.5B",
		"messages": [
			{
				"role": "user",
				"content": "What is the capital of France?"
			}
		]
	}'

Docker Model Runner
How to use zeng123/Depth-attention-1.5B with Docker Model Runner:
```
docker model run hf.co/zeng123/Depth-attention-1.5B
```

1b_depth_softmax_s24_qknorm_gqa6kv_smallpile_lr1p5e3

This model is a fine-tuned version of Llama_config/1b_depth_softmax_s24_qknorm_gqa6kv_lr1p5e3 on the smallpile dataset. It achieves the following results on the evaluation set:

Loss: 2.0506

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: 0.0015
train_batch_size: 4
eval_batch_size: 4
seed: 42
distributed_type: multi-GPU
num_devices: 32
gradient_accumulation_steps: 8
total_train_batch_size: 1024
total_eval_batch_size: 128
optimizer: Use OptimizerNames.ADAMW_TORCH with betas=(0.9,0.95) and epsilon=1e-08 and optimizer_args=No additional optimizer arguments
lr_scheduler_type: cosine_with_min_lr
lr_scheduler_warmup_ratio: 0.02
num_epochs: 1.0

Training results

Training Loss	Epoch	Step	Validation Loss
16.8546	0.64	10000	2.1235

Framework versions

Transformers 4.46.1
Pytorch 2.6.0
Datasets 3.1.0
Tokenizers 0.20.3

Downloads last month: 18

Safetensors

Model size

1B params

Tensor type

BF16