Instructions to use sapientinc/HRM-Text-1B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use sapientinc/HRM-Text-1B with Transformers:

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

pipe = pipeline("text-generation", model="sapientinc/HRM-Text-1B", trust_remote_code=True)

# Load model directly
from transformers import AutoTokenizer, AutoModelForCausalLM

tokenizer = AutoTokenizer.from_pretrained("sapientinc/HRM-Text-1B", trust_remote_code=True)
model = AutoModelForCausalLM.from_pretrained("sapientinc/HRM-Text-1B", trust_remote_code=True)

Notebooks
Google Colab
Kaggle
Local Apps

vLLM

How to use sapientinc/HRM-Text-1B with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "sapientinc/HRM-Text-1B"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "sapientinc/HRM-Text-1B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/sapientinc/HRM-Text-1B

SGLang

How to use sapientinc/HRM-Text-1B 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 "sapientinc/HRM-Text-1B" \
    --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": "sapientinc/HRM-Text-1B",
		"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 "sapientinc/HRM-Text-1B" \
        --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": "sapientinc/HRM-Text-1B",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use sapientinc/HRM-Text-1B with Docker Model Runner:
```
docker model run hf.co/sapientinc/HRM-Text-1B
```

Why HRM?

by GCBQDevTeam - opened 1 day ago

Discussion

GCBQDevTeam

1 day ago

From my understanding, TRM was a more efficient implementation of HRM that was functionally equivalent. Is there a technical reason why HRM was chosen for this project, or is it simply logistical?

abcd1927

Sapient AI org 1 day ago

We tried, but TRM is unstable and underperformant on language

GCBQDevTeam

1 day ago

•

edited 1 day ago

Interesting. I wonder if that is due to HRM having more layers. Thanks for the quick reply.

abcd1927

Sapient AI org about 23 hours ago

Thanks for the interest! We just put out our paper, you can find more details in it :)

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