huggingface/candle
1
0
Fork 0
mirror of https://github.com/huggingface/candle.git synced 2025-06-21 12:20:46 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add negative prompts to segment-anything. (#1000)

Browse Source
This commit is contained in:
Laurent Mazare
2023-09-30 07:17:42 +02:00
committed by GitHub
parent 34842fb234
commit 6203ced495
3 changed files with 34 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
candle-transformers/src/models/segment_anything/sam.rs
View File

@ -130,7 +130,7 @@ impl Sam {
pub fn forward(
&self,
img: &Tensor,
points: &[(f64, f64)],
points: &[(f64, f64, bool)],
multimask_output: bool,
) -> Result<(Tensor, Tensor)> {
let (_c, original_h, original_w) = img.dims3()?;
@ -150,12 +150,17 @@ impl Sam {
Ok((mask, iou))
}
/// Generate the mask and IOU predictions from some image embeddings and prompt.
///
/// The prompt is specified as a list of points `(x, y, b)`. `x` and `y` are the point
/// coordinates (between 0 and 1) and `b` is `true` for points that should be part of the mask
/// and `false` for points that should be part of the background and so excluded from the mask.
pub fn forward_for_embeddings(
&self,
img_embeddings: &Tensor,
original_h: usize,
original_w: usize,
points: &[(f64, f64)],
points: &[(f64, f64, bool)],
multimask_output: bool,
) -> Result<(Tensor, Tensor)> {
let image_pe = self.prompt_encoder.get_dense_pe()?;
@ -165,14 +170,18 @@ impl Sam {
let n_points = points.len();
let xys = points
.iter()
.flat_map(|(x, y)| {
.flat_map(|(x, y, _b)| {
let x = (*x as f32) * (original_w as f32);
let y = (*y as f32) * (original_h as f32);
[x, y]
})
.collect::<Vec<_>>();
let labels = points
.iter()
.map(|(_x, _y, b)| if *b { 1f32 } else { 0f32 })
.collect::<Vec<_>>();
let points = Tensor::from_vec(xys, (1, n_points, 2), img_embeddings.device())?;
let labels = Tensor::ones((1, n_points), DType::F32, img_embeddings.device())?;
let labels = Tensor::from_vec(labels, (1, n_points), img_embeddings.device())?;
Some((points, labels))
};
let points = points.as_ref().map(|(x, y)| (x, y));