Instructions to use Swindl/GLUS-S-partial with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Swindl/GLUS-S-partial with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-generation", model="Swindl/GLUS-S-partial")# Load model directly from transformers import AutoProcessor, AutoModelForCausalLM processor = AutoProcessor.from_pretrained("Swindl/GLUS-S-partial") model = AutoModelForCausalLM.from_pretrained("Swindl/GLUS-S-partial", device_map="auto") - Notebooks
- Google Colab
- Kaggle
- Local Apps Settings
- vLLM
How to use Swindl/GLUS-S-partial with vLLM:
Install from pip and serve model
# Install vLLM from pip: pip install vllm # Start the vLLM server: vllm serve "Swindl/GLUS-S-partial" # Call the server using curl (OpenAI-compatible API): curl -X POST "http://localhost:8000/v1/completions" \ -H "Content-Type: application/json" \ --data '{ "model": "Swindl/GLUS-S-partial", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }'Use Docker
docker model run hf.co/Swindl/GLUS-S-partial
- SGLang
How to use Swindl/GLUS-S-partial 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 "Swindl/GLUS-S-partial" \ --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": "Swindl/GLUS-S-partial", "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 "Swindl/GLUS-S-partial" \ --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": "Swindl/GLUS-S-partial", "prompt": "Once upon a time,", "max_tokens": 512, "temperature": 0.5 }' - Docker Model Runner
How to use Swindl/GLUS-S-partial with Docker Model Runner:
docker model run hf.co/Swindl/GLUS-S-partial
Update README.md
Browse files
README.md
CHANGED
|
@@ -1,6 +1,5 @@
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
| 3 |
-
pipeline_tag: video-segmentation
|
| 4 |
library_name: transformers
|
| 5 |
---
|
| 6 |
|
|
@@ -22,11 +21,6 @@ library_name: transformers
|
|
| 22 |
[](https://huggingface.co/Swindl/GLUS-A)
|
| 23 |
|
| 24 |
|
| 25 |
-
|
| 26 |
-
<div align=center>
|
| 27 |
-
<img src="assets/teaserfig.png" style="width:100%;">
|
| 28 |
-
</div>
|
| 29 |
-
|
| 30 |
## Overview
|
| 31 |
|
| 32 |
**RefVOS in complex scenarios** places high demands on models' video understanding and fine-grained localization capabilities. Recently, numerous models leveraging **MLLM-based** comprehension and reasoning abilities have been proposed to address this challenge. Our **GLUS** advances further along this methodological path.
|
|
@@ -37,11 +31,6 @@ library_name: transformers
|
|
| 37 |
|
| 38 |
📌 **GLUS is simple.** It elegantly integrates the approach for complex-scenario RefVOS tasks within a single MLLM framework, eliminating the necessity of utilizing other independent modules.
|
| 39 |
|
| 40 |
-
<div align=center>
|
| 41 |
-
<img src="assets/pipeline_00.png" style="width:100%;">
|
| 42 |
-
</div>
|
| 43 |
-
|
| 44 |
-
## News
|
| 45 |
|
| 46 |
## Installation
|
| 47 |
```shell
|
|
|
|
| 1 |
---
|
| 2 |
license: apache-2.0
|
|
|
|
| 3 |
library_name: transformers
|
| 4 |
---
|
| 5 |
|
|
|
|
| 21 |
[](https://huggingface.co/Swindl/GLUS-A)
|
| 22 |
|
| 23 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 24 |
## Overview
|
| 25 |
|
| 26 |
**RefVOS in complex scenarios** places high demands on models' video understanding and fine-grained localization capabilities. Recently, numerous models leveraging **MLLM-based** comprehension and reasoning abilities have been proposed to address this challenge. Our **GLUS** advances further along this methodological path.
|
|
|
|
| 31 |
|
| 32 |
📌 **GLUS is simple.** It elegantly integrates the approach for complex-scenario RefVOS tasks within a single MLLM framework, eliminating the necessity of utilizing other independent modules.
|
| 33 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
|
| 35 |
## Installation
|
| 36 |
```shell
|