Image-Text-to-Text
Transformers
Safetensors
qwen3_5
vllm
video
multimodal
reinforcement-learning
temporal-grounding
object-tracking
video-segmentation
visual-question-answering
spatial-reasoning
qwen3.5
conversational
Instructions to use OraRL/Video-ORA-4B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use OraRL/Video-ORA-4B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("image-text-to-text", model="OraRL/Video-ORA-4B") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] pipe(text=messages)# Load model directly from transformers import AutoProcessor, AutoModelForMultimodalLM processor = AutoProcessor.from_pretrained("OraRL/Video-ORA-4B") model = AutoModelForMultimodalLM.from_pretrained("OraRL/Video-ORA-4B", device_map="auto") messages = [ { "role": "user", "content": [ {"type": "image", "url": "https://huggingface.co/datasets/huggingface/documentation-images/resolve/main/p-blog/candy.JPG"}, {"type": "text", "text": "What animal is on the candy?"} ] }, ] inputs = processor.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(processor.decode(outputs[0][inputs["input_ids"].shape[-1]:])) - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use OraRL/Video-ORA-4B with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "OraRL/Video-ORA-4B" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/chat/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "OraRL/Video-ORA-4B", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'Use Docker
docker model run hf.co/OraRL/Video-ORA-4B
- SGLang
How to use OraRL/Video-ORA-4B 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 "OraRL/Video-ORA-4B" \ --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": "OraRL/Video-ORA-4B", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }'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 "OraRL/Video-ORA-4B" \ --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": "OraRL/Video-ORA-4B", "messages": [ { "role": "user", "content": [ { "type": "text", "text": "Describe this image in one sentence." }, { "type": "image_url", "image_url": { "url": "https://cdn.britannica.com/61/93061-050-99147DCE/Statue-of-Liberty-Island-New-York-Bay.jpg" } } ] } ] }' - Docker Model Runner
How to use OraRL/Video-ORA-4B with Docker Model Runner:
docker model run hf.co/OraRL/Video-ORA-4B
File size: 3,445 Bytes
0185029 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 | #!/usr/bin/env python3
"""Move legacy evaluation outputs into paper-level task families."""
from __future__ import annotations
import argparse
from pathlib import Path
from typing import Iterable
TASK_FAMILIES = {
"video_qa": (
"videomme",
"videommev2",
"videommmu",
"mmvu",
"mvbench",
"videoholmes",
"longvideobench",
"lvbench",
"mlvu",
),
"spatial_intelligence": ("vsi", "mmsi", "mindcube", "revsi"),
"spatial_temporal_grounding": ("stvg",),
}
def _exists(path: Path) -> bool:
return path.exists() or path.is_symlink()
def planned_moves(root: Path) -> list[tuple[Path, Path]]:
moves = []
for family, tasks in TASK_FAMILIES.items():
for task in tasks:
source = root / task
if source.is_dir():
moves.append((source, root / family / task))
return moves
def _preflight_merge(source: Path, destination: Path) -> None:
for child in source.iterdir():
target = destination / child.name
if not _exists(target):
continue
if child.is_dir() and not child.is_symlink():
if not target.is_dir() or target.is_symlink():
raise FileExistsError(f"cannot merge directory into {target}")
_preflight_merge(child, target)
continue
raise FileExistsError(f"refusing to overwrite existing output: {target}")
def _merge(source: Path, destination: Path) -> None:
destination.mkdir(parents=True, exist_ok=True)
for child in source.iterdir():
target = destination / child.name
if child.is_dir() and not child.is_symlink() and target.is_dir():
_merge(child, target)
else:
child.rename(target)
source.rmdir()
def organize(root: Path, moves: Iterable[tuple[Path, Path]]) -> None:
moves = list(moves)
for source, destination in moves:
if _exists(destination):
_preflight_merge(source, destination)
for source, destination in moves:
destination.parent.mkdir(parents=True, exist_ok=True)
if _exists(destination):
_merge(source, destination)
else:
source.rename(destination)
def create_parser() -> argparse.ArgumentParser:
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument(
"root",
type=Path,
help="One model's output root, for example outputs/Video-ORA-9B.",
)
parser.add_argument(
"--apply",
action="store_true",
help="Perform the moves. Without this flag, only print the plan.",
)
return parser
def main() -> int:
args = create_parser().parse_args()
root = args.root.expanduser().resolve()
if not root.is_dir():
raise SystemExit(f"output root is not a directory: {root}")
moves = planned_moves(root)
if not moves:
print(f"Already organized: {root}")
return 0
action = "MOVE" if args.apply else "PLAN"
for source, destination in moves:
print(f"{action} {source.relative_to(root)} -> {destination.relative_to(root)}")
if args.apply:
organize(root, moves)
print(f"Organized {len(moves)} task director{'y' if len(moves) == 1 else 'ies'}.")
else:
print("Dry run only; pass --apply to move these directories.")
return 0
if __name__ == "__main__":
raise SystemExit(main())
|