update fmt

src/models/deepseek_ocr/model.rs

@@ -492,9 +492,11 @@ impl ImageEncoderViT {
     }
     pub fn forward(&self, xs: &Tensor) -> Result<Tensor> {
         let mut x = self.patch_embed.forward(xs)?;
-        if self.pos_embed.is_some() {
+        // if self.pos_embed.is_some() {
+        if let Some(pos_emb) = &self.pos_embed {
             let dim1 = x.dim(1)?;
-            let pos = self.get_abs_pos_sam(self.pos_embed.as_ref().unwrap(), dim1)?;
+            // let pos = self.get_abs_pos_sam(self.pos_embed.as_ref().unwrap(), dim1)?;
+            let pos = self.get_abs_pos_sam(pos_emb, dim1)?;
             x = x.broadcast_add(&pos)?;
         }
         for blk in &self.blocks {

src/models/fun_asr_nano/model.rs

@@ -268,19 +268,15 @@ impl EncoderLayerSANM {
                 self.self_attn
                     .forward(&xs, mask, mask_shfit_chunk, mask_att_chunk_encoder)?;
             let x_concat = Tensor::cat(&[&xs, &attn], D::Minus1)?;
-            if self.in_dim == self.hidden_dim {
-                let x_concat = self
-                    .concat_linear
-                    .as_ref()
-                    .unwrap()
+            if self.in_dim == self.hidden_dim
+                && let Some(concat_linear) = &self.concat_linear
+            {
+                let x_concat = concat_linear
                     .forward(&x_concat)?
                     .affine(stoch_layer_coeff, 0.0)?;
                 xs = residual.add(&x_concat)?;
-            } else {
-                xs = self
-                    .concat_linear
-                    .as_ref()
-                    .unwrap()
+            } else if let Some(concat_linear) = &self.concat_linear {
+                xs = concat_linear
                     .forward(&x_concat)?
                     .affine(stoch_layer_coeff, 0.0)?;
             }
@@ -496,13 +492,10 @@ impl AdaptorEncoderLayer {
         } else {
             xs.clone()
         };
-        if self.concat_linear.is_some() {
+        if let Some(concat_linear) = &self.concat_linear {
             let attn = self.self_attn.forward(&xs, None, None, mask, false)?;
             let x_concat = Tensor::cat(&[&xs, &attn], D::Minus1)?;
-            let x_concat = self
-                .concat_linear
-                .as_ref()
-                .unwrap()
+            let x_concat = concat_linear
                 .forward(&x_concat)?
                 .affine(stoch_layer_coeff, 0.0)?;
             xs = residual.add(&x_concat)?;

src/models/paddleocr_vl/model.rs

@@ -707,11 +707,13 @@ impl PaddleOCRVLModel {
             self.rope_deltas = Some(rope_deltas);
         } else {
             let (bs, seq_len, _) = inputs_embeds.dims3()?;
-            let delta = if let Some(cache_position) = cache_position {
+            let delta = if let Some(cache_position) = cache_position
+                && let Some(rope_deltas) = &self.rope_deltas
+            {
                 cache_position
                     .i(0)?
-                    .to_dtype(self.rope_deltas.as_ref().unwrap().dtype())?
-                    .broadcast_add(self.rope_deltas.as_ref().unwrap())?
+                    .to_dtype(rope_deltas.dtype())?
+                    .broadcast_add(rope_deltas)?
                     .contiguous()?
                     .to_dtype(candle_core::DType::U32)?
             } else {

src/models/qwen2_5vl/model.rs

@@ -1057,11 +1057,13 @@ impl Qwen2_5VLModel {
             self.rope_deltas = Some(rope_deltas);
         } else {
             let (bs, seq_len, _) = inputs_embeds.dims3()?;
-            let delta = if let Some(cache_position) = cache_position {
+            let delta = if let Some(cache_position) = cache_position
+                && let Some(rope_deltas) = &self.rope_deltas
+            {
                 cache_position
                     .i(0)?
-                    .to_dtype(self.rope_deltas.as_ref().unwrap().dtype())?
-                    .broadcast_add(self.rope_deltas.as_ref().unwrap())?
+                    .to_dtype(rope_deltas.dtype())?
+                    .broadcast_add(rope_deltas)?
                     .contiguous()?
                     .to_dtype(candle_core::DType::U32)?
             } else {

src/models/qwen3_5/model.rs

@@ -1316,26 +1316,26 @@ impl Qwen3_5Model {
         video_grid_thw: Option<&Tensor>,
         seqlen_offset: usize,
     ) -> Result<Tensor> {
-        let position_ids = if self.rope_deltas.is_none() {
-            let (position_ids, rope_deltas) =
-                self.get_rope_index(input_ids, image_grid_thw, video_grid_thw, None)?;
-            self.rope_deltas = Some(rope_deltas);
-            position_ids
-        } else {
+        let position_ids = if let Some(rope_deltas) = &self.rope_deltas {
             let (bs, seq_len, _) = inputs_embeds.dims3()?;
             Tensor::arange(
                 seqlen_offset as i64,
                 (seqlen_offset + seq_len) as i64,
                 input_ids.device(),
             )?
-            .to_dtype(self.rope_deltas.as_ref().unwrap().dtype())?
+            .to_dtype(rope_deltas.dtype())?
             .unsqueeze(0)?
             .broadcast_as((bs, seq_len))?
-            .broadcast_add(self.rope_deltas.as_ref().unwrap())?
+            .broadcast_add(rope_deltas)?
             .unsqueeze(0)?
             .broadcast_as((3, bs, seq_len))?
             .contiguous()?
             .to_dtype(candle_core::DType::U32)?
+        } else {
+            let (position_ids, rope_deltas) =
+                self.get_rope_index(input_ids, image_grid_thw, video_grid_thw, None)?;
+            self.rope_deltas = Some(rope_deltas);
+            position_ids
         };
         Ok(position_ids)
     }

src/models/qwen3vl/model.rs

@@ -1013,11 +1013,13 @@ impl Qwen3VLModel {
             self.rope_deltas = Some(rope_deltas);
         } else {
             let (bs, seq_len, _) = inputs_embeds.dims3()?;
-            let delta = if let Some(cache_position) = cache_position {
+            let delta = if let Some(cache_position) = cache_position
+                && let Some(rope_deltas) = &self.rope_deltas
+            {
                 cache_position
                     .i(0)?
-                    .to_dtype(self.rope_deltas.as_ref().unwrap().dtype())?
-                    .broadcast_add(self.rope_deltas.as_ref().unwrap())?
+                    .to_dtype(rope_deltas.dtype())?
+                    .broadcast_add(rope_deltas)?
                     .contiguous()?
                     .to_dtype(candle_core::DType::U32)?
             } else {

src/models/qwen3vl/processor.rs

@@ -9,6 +9,9 @@ use candle_core::{DType, Device, IndexOp, Shape, Tensor};
 use ffmpeg_next as ffmpeg;
 use image::DynamicImage;
 
+#[cfg(feature = "ffmpeg")]
+use anyhow::anyhow;
+
 #[cfg(feature = "ffmpeg")]
 use crate::utils::video_utils::video_smart_resize;
 use crate::{

src/models/rmbg2_0/model.rs

@@ -63,10 +63,10 @@ impl PatchEmbed {
             xs = xs.pad_with_zeros(2, 0, self.patch_size - h % self.patch_size)?;
         }
         xs = self.proj.forward(&xs)?;
-        if self.norm.is_some() {
+        if let Some(norm) = &self.norm {
             let (_, _, ph, pw) = xs.dims4()?;
             xs = xs.flatten_from(2)?.transpose(1, 2)?;
-            xs = self.norm.as_ref().unwrap().forward(&xs)?;
+            xs = norm.forward(&xs)?;
             xs = xs.transpose(1, 2)?.reshape(((), self.embed_dim, ph, pw))?;
         }
         Ok(xs)
@@ -1018,9 +1018,12 @@ impl BasicDecBlk {
     pub fn new(vb: VarBuilder, in_c: usize, out_c: usize) -> Result<Self> {
         let inter_channels = 64;
         let conv_in = get_conv2d(vb.pp("conv_in"), in_c, inter_channels, 3, 1, 1, 1, 1, true)?;
-        let dec_att = ASPPDeformable::new(vb.pp("dec_att"), inter_channels, inter_channels, vec![
-            1, 3, 7,
-        ])?;
+        let dec_att = ASPPDeformable::new(
+            vb.pp("dec_att"),
+            inter_channels,
+            inter_channels,
+            vec![1, 3, 7],
+        )?;
         let conv_out = get_conv2d(
             vb.pp("conv_out"),
             inter_channels,

src/position_embed/sinusoidal_pe.rs

@@ -33,11 +33,11 @@ impl SinusoidalPositionEncoderCat {
             device,
         )?
         .reshape((seq_len, 1))?; // (seq_len, 1)
-        let inv_freq = if self.inv_freq.is_none() {
+        let inv_freq = if let Some(inv_freq) = &self.inv_freq {
+            inv_freq.clone()
+        } else {
             let inv_freq = compute_default_rope_parameters(head_dim, 10000.0);
             Tensor::from_slice(&inv_freq, (1, inv_freq.len()), device)?
-        } else {
-            self.inv_freq.as_ref().unwrap().clone()
         };
         let freqs = positions.matmul(&inv_freq)?; // (seq_len, dim / 2)
         let sin = freqs.sin()?;