huggingface/candle
1
0
Fork 0
mirror of https://github.com/huggingface/candle.git synced 2025-06-16 18:48:51 +00:00
Code Issues Packages Projects Releases Wiki Activity

Segment Anything - process images (#766)

Browse Source 
* Start processing images.

* Add LayerNorm2d.

* Properly use LayerNorm2d.

* Tweak eps.

* Use LayerNorm on inputs with a rank different from 3.

* Window partitioning.

* Fix a couple todos.

* More todos.

* Hard-code the einsums.

* More padding support.

* Some sizes tweaks.

* Use the hub to get the weights.

* Use a batch matmul.

* Tweaks.

* More fixes.

* Get some predictions to be generated.
This commit is contained in:
Laurent Mazare
2023-09-07 19:22:45 +01:00
committed by GitHub
parent 7b50f3e106
commit 7396b8ed1a
10 changed files with 303 additions and 105 deletions
Download Patch File Download Diff File Expand all files Collapse all files

153
candle-examples/examples/segment-anything/model_image_encoder.rs
View File

@ -70,6 +70,60 @@ impl Attention {
rel_pos_hw,
})
}
fn add_decomposed_rel_pos(
&self,
attn: Tensor,
q: &Tensor,
(q_h, q_w): (usize, usize),
(k_h, k_w): (usize, usize),
) -> Result<Tensor> {
match &self.rel_pos_hw {
Some((rel_pos_h, rel_pos_w)) => {
let r_h = get_rel_pos(q_h, k_h, rel_pos_h)?;
let r_w = get_rel_pos(q_w, k_w, rel_pos_w)?;
let (b, _, dim) = q.dims3()?;
let r_q = q.reshape((b, q_h, q_w, dim))?;
// rel_h = torch.einsum("bhwc,hkc->bhwk", r_q, Rh)
let rel_h = r_q.matmul(&r_h.broadcast_left(b)?.t()?.contiguous()?)?;
// rel_w = torch.einsum("bhwc,wkc->bhwk", r_q, Rw)
let rel_w = r_q
.transpose(1, 2)? // -> bwhc
.contiguous()?
.matmul(&r_w.broadcast_left(b)?.t()?.contiguous()?)? // bwhc,bwck -> bwhk
.transpose(1, 2)?;
(attn.reshape((b, q_h, q_w, k_h, k_w))?
+ rel_h.unsqueeze(4)?.broadcast_add(&rel_w.unsqueeze(3)?)?)?
.reshape((b, q_h * q_w, k_h * k_w))
}
None => Ok(attn),
}
}
}
fn get_rel_pos(q_size: usize, k_size: usize, rel_pos: &Tensor) -> Result<Tensor> {
let max_rel_dist = 2 * usize::max(q_size, k_size) - 1;
let dev = rel_pos.device();
let rel_pos_resized = if rel_pos.dim(0)? != max_rel_dist {
todo!("interpolation")
} else {
rel_pos
};
let q_coords = Tensor::arange(0u32, q_size as u32, dev)?
.reshape((q_size, 1))?
.to_dtype(DType::F32)?;
let k_coords = Tensor::arange(0u32, k_size as u32, dev)?
.reshape((1, k_size))?
.to_dtype(DType::F32)?;
let q_coords = (q_coords * f64::max(1f64, k_size as f64 / q_size as f64))?;
let k_coords = (k_coords * f64::max(1f64, q_size as f64 / k_size as f64))?;
let relative_coords = (q_coords.broadcast_sub(&k_coords)?
+ (k_size as f64 - 1.) * f64::max(1f64, q_size as f64 / k_size as f64))?;
let (d1, d2) = relative_coords.dims2()?;
let relative_coords = relative_coords.to_dtype(DType::U32)?;
rel_pos_resized
.index_select(&relative_coords.reshape(d1 * d2)?, 0)?
.reshape((d1, d2, rel_pos_resized.dim(1)?))
}
impl Module for Attention {
@ -77,24 +131,22 @@ impl Module for Attention {
let (b, h, w, c) = xs.dims4()?;
let qkv = self
.qkv
.forward(xs)?
.forward(&xs.flatten_to(1)?)?
.reshape((b, h * w, 3, self.num_heads, c / self.num_heads))?
.permute((2, 0, 3, 1, 4))?
.reshape((3, b * self.num_heads, h * w, c / self.num_heads))?;
let q = qkv.i(0)?;
let k = qkv.i(1)?;
let v = qkv.i(2)?;
let attn = (q * self.scale)?.matmul(&k.t()?)?;
if self.use_rel_pos {
todo!()
}
let attn = (&q * self.scale)?.matmul(&k.t()?)?;
let attn = self.add_decomposed_rel_pos(attn, &q, (h, w), (h, w))?;
let attn = candle_nn::ops::softmax_last_dim(&attn)?;
let attn = attn.matmul(&v)?;
let attn = attn
.matmul(&v)?
.reshape((b, self.num_heads, h, w, c / self.num_heads))?
.permute((0, 2, 3, 1, 4))?
.reshape((b, h, w, c / self.num_heads))?;
self.proj.forward(&attn)
.reshape((b, h * w, c))?;
self.proj.forward(&attn)?.reshape((b, h, w, c))
}
}
@ -117,8 +169,8 @@ impl Block {
input_size: (usize, usize),
vb: VarBuilder,
) -> Result<Self> {
let norm1 = layer_norm(dim, 1e-5, vb.pp("norm1"))?;
let norm2 = layer_norm(dim, 1e-5, vb.pp("norm2"))?;
let norm1 = layer_norm(dim, 1e-6, vb.pp("norm1"))?;
let norm2 = layer_norm(dim, 1e-6, vb.pp("norm2"))?;
let input_size_attn = if window_size == 0 {
input_size
} else {
@ -132,7 +184,7 @@ impl Block {
input_size_attn,
vb.pp("attn"),
)?;
let mlp = crate::MlpBlock::new(dim, dim * 4, vb.pp("mlp"))?;
let mlp = crate::MlpBlock::new(dim, dim * 4, candle_nn::Activation::Gelu, vb.pp("mlp"))?;
Ok(Self {
norm1,
attn,
@ -143,17 +195,76 @@ impl Block {
}
}
fn window_partition(xs: Tensor, window_size: usize) -> Result<(Tensor, (usize, usize))> {
let (b, h, w, c) = xs.dims4()?;
let pad_h = (window_size - h % window_size) % window_size;
let pad_w = (window_size - w % window_size) % window_size;
let xs = if pad_h > 0 {
xs.pad_with_zeros(1, 0, pad_h)?
} else {
xs
};
let xs = if pad_w > 0 {
xs.pad_with_zeros(2, 0, pad_w)?
} else {
xs
};
let (h_p, w_p) = (h + pad_h, w + pad_w);
let windows = xs
.reshape((
b,
h_p / window_size,
window_size,
w_p / window_size,
window_size,
c,
))?
.transpose(2, 3)?
.contiguous()?
.flatten_to(2)?;
Ok((windows, (h_p, w_p)))
}
fn window_unpartition(
windows: Tensor,
window_size: usize,
(h_p, w_p): (usize, usize),
(h, w): (usize, usize),
) -> Result<Tensor> {
let b = windows.dim(0)? / (h_p * w_p / window_size / window_size);
let xs = windows
.reshape((
b,
h_p / window_size,
w_p / window_size,
window_size,
window_size,
windows.elem_count() / b / h_p / w_p,
))?
.transpose(2, 3)?
.contiguous()?
.reshape((b, h_p, w_p, windows.elem_count() / b / h_p / w_p))?;
let xs = if h_p > h { xs.narrow(1, 0, h)? } else { xs };
let xs = if w_p > w { xs.narrow(2, 0, w)? } else { xs };
Ok(xs)
}
impl Module for Block {
fn forward(&self, xs: &Tensor) -> Result<Tensor> {
let shortcut = xs;
let xs = self.norm1.forward(xs)?;
if self.window_size > 0 {
todo!()
}
let hw = (xs.dim(1)?, xs.dim(2)?);
let (xs, pad_hw) = if self.window_size > 0 {
window_partition(xs, self.window_size)?
} else {
(xs, (0, 0))
};
let xs = self.attn.forward(&xs)?;
if self.window_size > 0 {
todo!()
}
let xs = if self.window_size > 0 {
window_unpartition(xs, self.window_size, pad_hw, hw)?
} else {
xs
};
let xs = (xs + shortcut)?;
&xs + xs.apply(&self.norm2)?.apply(&self.mlp)?
}
@ -165,9 +276,9 @@ pub struct ImageEncoderViT {
patch_embed: PatchEmbed,
blocks: Vec<Block>,
neck_conv1: candle_nn::Conv2d,
neck_ln1: LayerNorm,
neck_ln1: crate::LayerNorm2d,
neck_conv2: candle_nn::Conv2d,
neck_ln2: LayerNorm,
neck_ln2: crate::LayerNorm2d,
pos_embed: Option<Tensor>,
}
@ -222,13 +333,13 @@ impl ImageEncoderViT {
Default::default(),
vb.pp("neck.0"),
)?;
let neck_ln1 = layer_norm(out_chans, 1e-6, vb.pp("neck.1"))?;
let neck_ln1 = crate::LayerNorm2d::new(out_chans, 1e-6, vb.pp("neck.1"))?;
let cfg = candle_nn::Conv2dConfig {
padding: 1,
..Default::default()
};
let neck_conv2 = candle_nn::conv2d_no_bias(out_chans, out_chans, 3, cfg, vb.pp("neck.2"))?;
let neck_ln2 = layer_norm(out_chans, 1e-6, vb.pp("neck.3"))?;
let neck_ln2 = crate::LayerNorm2d::new(out_chans, 1e-6, vb.pp("neck.3"))?;
let pos_embed = if use_abs_pos {
let p = vb.get(
(1, img_size / patch_size, img_size / patch_size, embed_dim),