Downloads · 30 days
7
39% of all-time downloads
grasp-anything-9040
charlesH777/grasp-anything-9040
grasp-anything-9040 is a image-text-to-text model from charlesH777. Use it for the image-text-to-text task on the model card, and read the license before you ship it in a product. The card lists the license as apache-2.0.
基于 NVIDIA LocateAnything-3B 的语言引导二维抓取模型,Phase 2 SFT 阶段产物。
All-time downloads
18
Public
Parameters
3.6B
7.8 GB on disk
Likes
0
Public
Repo makeup
Click a slice to open those files.
.safetensors7.8 GB · 100%
At a glance
- Task
- Image-Text-to-Text
- License
- apache-2.0
- Model type
- locateanything
- Access
- Public
- Created
- Aug 1, 2026
- Updated
- Aug 1, 2026
- SHA
- 47ef3e8d
Try a prompt
Model card
From the Hugging Face model README
Grasp Anything 2D — Checkpoint 9040 (grasp_rect 专用)
基于 NVIDIA LocateAnything-3B 的语言引导二维抓取模型,Phase 2 SFT 阶段产物。
任务
grasp_rect — 输出 (cx, cy, θ, width) 矩形抓取参数,按 RealVLG 固定 40 px depth 展开矩形。
评测结果(RealVLG 官方 652 样本)
| Split | n | gAcc_corrected_strict | mIoU_strict | angle_seam_gAcc |
|---|---|---|---|---|
| seen | 253 | 69.96% | 43.08% | 65.00% |
| similar | 235 | 57.02% | 35.48% | 81.25% |
| novel | 164 | 27.44% | 17.02% | 33.33% |
- 评测协议:
evaluate_realvlg_grasp.py,fast 模式 - gAcc_corrected_strict: IoU>0.25 且 angle<30° (mod 180, 修正了官方 radian/degree bug)
- 对比 RealVLG-R1 GRPO Grasp 3B: seen gAcc +28.5%, similar +22.1%
使用方法
from locate_anything_service.model import LocateAnythingRuntime
from locate_anything_service.config import Settings
settings = Settings(model_id="charlesH777/grasp-anything-9040")
runtime = LocateAnythingRuntime(settings)
result = runtime.predict("image.jpg", "抓取红色杯子", mode="grasp_rect")
训练配置
- 阶段: Phase 2 SFT (grasp_rect 专用)
- LLM LoRA: rank 32, target Qwen2.5-3B-Instruct
- 训练参数:
grasp_rect_task_enabled: true,contact_loss_enabled: false
- Task
- Image-Text-to-Text
- Type
- locateanything
- License
- apache-2.0
- Languages
- en, zh
- Created
- Aug 1, 2026
- Updated
- Aug 1, 2026