FastSAM

FastSAM copied to clipboard

尽管推理速度很快，但在promt 阶段有个函数严重拖累了，导致每秒只能处理3张图。 plot_to_result plot_to_result cost too much time （round 300ms per ）  请问有什么方法优化么？

dd-rongfa

Hi, I've trained the model on custom dataset, I was wondering which script I need to use to make inference?

Samareh-Az

Hi, The license for FastSAM previously was Apache 2.0 and now it shows AGPL 3.0. Can we still use this in commercial product? Thanks

sulaimanvesal

MatplotlibDeprecationWarning: The tostring_rgb function was deprecated in Matplotlib 3.8 and will be removed two minor releases later. Use buffer_rgba instead. buf

1

File prompt.py: ```python try: buf = fig.canvas.tostring_rgb() except AttributeError: fig.canvas.draw() buf = fig.canvas.tostring_rgb() cols, rows = fig.canvas.get_width_height() img_array = np.frombuffer(buf, dtype=np.uint8).reshape(rows, cols, 3) result = cv2.cvtColor(img_array, cv2.COLOR_RGB2BGR) plt.close() return result...

pixkk

Thanks for your awesome project, Does it have a web demo like SAM https://github.com/facebookresearch/segment-anything/tree/main/demo

sswink

Hello, thank you for providing such excellent results. I want to display only the contour of the segmentation result without the color mask. How should I modify the code?

zxcvmmm

First of all, let me thank and congratulate the authors for this wok. I feel it is a smart way of combining what SAM and the existing semantic segmentation work...

akshay-iyer

Thanks

Qidian213

In the postprocess method of FastSAMPredictor, when critical_iou_index is > 1, it raises the following error: ``` Error: expand(torch.FloatTensor{[2]}, size=[]): the number of sizes provided (0) must be greater or...

javierganan99

Import clip to "prompt.py" and "Inference.py" to avoid error "name 'clip' is not defined"

Xiaoyu68