Instructions to use openvla/openvla-7b-finetuned-libero-spatial with libraries, inference providers, notebooks, and local apps. Follow these links to get started.

Libraries

How to use openvla/openvla-7b-finetuned-libero-spatial with Transformers:

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

pipe = pipeline("image-text-to-text", model="openvla/openvla-7b-finetuned-libero-spatial", trust_remote_code=True)

# Load model directly
from transformers import AutoModelForVision2Seq
model = AutoModelForVision2Seq.from_pretrained("openvla/openvla-7b-finetuned-libero-spatial", trust_remote_code=True, device_map="auto")

Notebooks
Google Colab
Kaggle
Local Apps Settings

vLLM

How to use openvla/openvla-7b-finetuned-libero-spatial with vLLM:

Install from pip and serve model

# Install vLLM from pip:
pip install vllm
# Start the vLLM server:
vllm serve "openvla/openvla-7b-finetuned-libero-spatial"
# Call the server using curl (OpenAI-compatible API):
curl -X POST "http://localhost:8000/v1/completions" \
	-H "Content-Type: application/json" \
	--data '{
		"model": "openvla/openvla-7b-finetuned-libero-spatial",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Use Docker

docker model run hf.co/openvla/openvla-7b-finetuned-libero-spatial

SGLang

How to use openvla/openvla-7b-finetuned-libero-spatial 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 "openvla/openvla-7b-finetuned-libero-spatial" \
    --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": "openvla/openvla-7b-finetuned-libero-spatial",
		"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 "openvla/openvla-7b-finetuned-libero-spatial" \
        --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": "openvla/openvla-7b-finetuned-libero-spatial",
		"prompt": "Once upon a time,",
		"max_tokens": 512,
		"temperature": 0.5
	}'

Docker Model Runner
How to use openvla/openvla-7b-finetuned-libero-spatial with Docker Model Runner:
```
docker model run hf.co/openvla/openvla-7b-finetuned-libero-spatial
```

openvla-7b-finetuned-libero-spatial / config.json

moojink

Modify unnorm_key in config.json, dataset_statistics.json

d923f6e almost 2 years ago

Raw

History Blame Contribute Delete

3.51 kB

	{
	"_name_or_path": "openvla/openvla-7b",
	"arch_specifier": "no-align+fused-gelu-mlp",
	"architectures": [
	"OpenVLAForActionPrediction"
	],
	"auto_map": {
	"AutoConfig": "openvla/openvla-7b--configuration_prismatic.OpenVLAConfig",
	"AutoModelForVision2Seq": "openvla/openvla-7b--modeling_prismatic.OpenVLAForActionPrediction"
	},
	"hf_llm_id": "meta-llama/Llama-2-7b-hf",
	"image_resize_strategy": "resize-naive",
	"image_sizes": [
	224,
	224
	],
	"llm_backbone_id": "llama2-7b-pure",
	"llm_max_length": 2048,
	"model_type": "openvla",
	"n_action_bins": 256,
	"norm_stats": {
	"libero_spatial": {
	"action": {
	"mean": [
	0.15312479436397552,
	0.13707277178764343,
	-0.15526802837848663,
	-0.005176450591534376,
	-0.01120874285697937,
	-0.020194264128804207,
	0.4578818082809448
	],
	"std": [
	0.41272708773612976,
	0.34724321961402893,
	0.50869220495224,
	0.037266165018081665,
	0.07244449853897095,
	0.05762382969260216,
	0.49827873706817627
	],
	"max": [
	0.9375,
	0.9375,
	0.9375,
	0.1971428543329239,
	0.33642858266830444,
	0.375,
	1.0
	],
	"min": [
	-0.9375,
	-0.9375,
	-0.9375,
	-0.1875,
	-0.3675000071525574,
	-0.36000001430511475,
	0.0
	],
	"q01": [
	-0.7454732114076613,
	-0.6616071462631226,
	-0.9375,
	-0.1071428582072258,
	-0.20678570866584778,
	-0.1842857152223587,
	0.0
	],
	"q99": [
	0.9375,
	0.8758928775787354,
	0.9321428537368774,
	0.1039285734295845,
	0.17678570747375488,
	0.14571428298950195,
	1.0
	],
	"mask": [
	true,
	true,
	true,
	true,
	true,
	true,
	false
	]
	},
	"proprio": {
	"mean": [
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	0.0
	],
	"std": [
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	0.0
	],
	"max": [
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	0.0
	],
	"min": [
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	0.0
	],
	"q01": [
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	0.0
	],
	"q99": [
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	0.0,
	0.0
	]
	},
	"num_transitions": 52970,
	"num_trajectories": 432
	}
	},
	"output_projector_states": false,
	"pad_to_multiple_of": 64,
	"pad_token_id": 32000,
	"text_config": {
	"model_type": "llama",
	"pad_token_id": 32000,
	"torch_dtype": "bfloat16",
	"vocab_size": 32064
	},
	"timm_model_ids": [
	"vit_large_patch14_reg4_dinov2.lvd142m",
	"vit_so400m_patch14_siglip_224"
	],
	"timm_override_act_layers": [
	null,
	null
	],
	"torch_dtype": "bfloat16",
	"transformers_version": "4.40.1",
	"use_fused_vision_backbone": true,
	"vision_backbone_id": "dinosiglip-vit-so-224px"
	}