Instructions to use nvidia/Hymba-1.5B-Base with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use nvidia/Hymba-1.5B-Base with Transformers:

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

pipe = pipeline("text-generation", model="nvidia/Hymba-1.5B-Base", trust_remote_code=True)
messages = [
    {"role": "user", "content": "Who are you?"},
]
pipe(messages)

# Load model directly
from transformers import AutoModelForCausalLM
model = AutoModelForCausalLM.from_pretrained("nvidia/Hymba-1.5B-Base", trust_remote_code=True, dtype="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use nvidia/Hymba-1.5B-Base with vLLM:

Install from pip and serve model

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

Use Docker

docker model run hf.co/nvidia/Hymba-1.5B-Base

SGLang

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

Docker Model Runner
How to use nvidia/Hymba-1.5B-Base with Docker Model Runner:
```
docker model run hf.co/nvidia/Hymba-1.5B-Base
```

A collection of observations in the code

by AntonV - opened Nov 22, 2024

Discussion

AntonV

Nov 22, 2024

Note: don't mind this if you're aware that the code is not finished

I went through the custom code given here (https://huggingface.co/nvidia/Hymba-1.5B-Base/blob/main/modeling_hymba.py) and I noticed a few things:

The fast path (one kernel for the whole mamba op) doesn't consider the attention given in the slow path
It seems like we only use one "head" of mamba, i.e. index is fixed to 0
self_attn_weights returns torch empty if I see it correctly
The batch size stuff seems like the problem we discovered in hf, i.e. conv1d + mamba need to zero padded tokens + left padding

YongganFu

NVIDIA org Nov 25, 2024

Thank you for your interest in our work and for pointing these out!

Yes, you are correct that we didn't use the fast path of Mamba since we have both attention and Mamba in one Hymba block. Building a fused kernel to support the entire Hymba block is our ongoing work.
Yes, we use a single Mamba block as SSM heads, treating each inner dimension of Mamba as a separate head. As an analogy to Mamba2, where each head has a separate A value, each inner dimension in Mamba1 has its own distinct A value and thus can be viewed as a separate head.
Good catch! We will not output self_attn_weights as we use FlexAttention and FlashAttention. We will remove this in the next version.
Yes, you are correct. The batch size issue arises because of the order of meta tokens and padding, which may violate the padding rules of some operators. We will update this soon.

pmolchanov

NVIDIA org Nov 26, 2024

@AntonV thanks for looking into the code! Let us know if you have other observations, or have feedback. We appreciate you looking into this.

pmolchanov changed discussion status to closed Nov 29, 2024

Upload images, audio, and videos by dragging in the text input, pasting, or clicking here.

Tap or paste here to upload images

· Sign up or log in to comment