mii/usls
1
0
Fork 0
mirror of https://github.com/mii443/usls.git synced 2025-08-22 15:45:41 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add RTMO model for keypoint detection (#2)

Browse Source 
* Add RTMO
This commit is contained in:
Jamjamjon
2024-04-08 20:45:38 +08:00
committed by GitHub
parent a0d410b46d
commit ead175234c
31 changed files with 237 additions and 53 deletions
Download Patch File Download Diff File Expand all files Collapse all files

37
README.md
View File

@ -1,29 +1,30 @@
# usls
A Rust library integrated with **ONNXRuntime**, providing a collection of **Computer Vison** and **Vision-Language** models including [YOLOv8](https://github.com/ultralytics/ultralytics) `(Classification, Segmentation, Detection and Pose Detection)`, [YOLOv9](https://github.com/WongKinYiu/yolov9), [RTDETR](https://arxiv.org/abs/2304.08069), [CLIP](https://github.com/openai/CLIP), [DINOv2](https://github.com/facebookresearch/dinov2), [FastSAM](https://github.com/CASIA-IVA-Lab/FastSAM), [YOLO-World](https://github.com/AILab-CVC/YOLO-World), [BLIP](https://arxiv.org/abs/2201.12086), [PaddleOCR](https://github.com/PaddlePaddle/PaddleOCR) and others. Many execution providers are supported, sunch as `CUDA`, `TensorRT` and `CoreML`.
A Rust library integrated with **ONNXRuntime**, providing a collection of **Computer Vison** and **Vision-Language** models including [YOLOv8](https://github.com/ultralytics/ultralytics), [YOLOv9](https://github.com/WongKinYiu/yolov9), [RTDETR](https://arxiv.org/abs/2304.08069), [CLIP](https://github.com/openai/CLIP), [DINOv2](https://github.com/facebookresearch/dinov2), [FastSAM](https://github.com/CASIA-IVA-Lab/FastSAM), [YOLO-World](https://github.com/AILab-CVC/YOLO-World), [BLIP](https://arxiv.org/abs/2201.12086), [PaddleOCR](https://github.com/PaddlePaddle/PaddleOCR) and others. Many execution providers are supported, sunch as `CUDA`, `TensorRT` and `CoreML`.
## Supported Models
| Model | Example | CUDA<br />f32 | CUDA<br />f16 | TensorRT<br />f32 | TensorRT<br />f16 |
| :---------------------------------------------------------------: | :----------------------: | :-----------: | :-----------: | :------------------------: | :-----------------------: |
| **YOLOv8-detection** | [demo](examples/yolov8) | ✅ | ✅ | ✅ | ✅ |
| **YOLOv8-pose** | [demo](examples/yolov8) | ✅ | ✅ | ✅ | ✅ |
| **YOLOv8-classification** | [demo](examples/yolov8) | ✅ | ✅ | ✅ | ✅ |
| **YOLOv8-segmentation** | [demo](examples/yolov8) | ✅ | ✅ | ✅ | ✅ |
| **YOLOv8-OBB** | TODO | TODO | TODO | TODO | TODO |
| **YOLOv9** | [demo](examples/yolov9) | ✅ | ✅ | ✅ | ✅ |
| **RT-DETR** | [demo](examples/rtdetr) | ✅ | ✅ | ✅ | ✅ |
| **FastSAM** | [demo](examples/fastsam) | ✅ | ✅ | ✅ | ✅ |
| **YOLO-World** | [demo](examples/yolo-world) | ✅ | ✅ | ✅ | ✅ |
| **DINOv2** | [demo](examples/dinov2) | ✅ | ✅ | ✅ | ✅ |
| **CLIP** | [demo](examples/clip) | ✅ | ✅ | ✅ visual<br />❌ textual | ✅ visual<br />❌ textual |
| **BLIP** | [demo](examples/blip) | ✅ | ✅ | ✅ visual<br />❌ textual | ✅ visual<br />❌ textual |
| [**DB(Text Detection)**](https://arxiv.org/abs/1911.08947) | [demo](examples/db) | ✅ | ❌ | ✅ | ✅ |
| [**SVTR(Text Recognition)**](https://arxiv.org/abs/2205.00159) | [demo](examples/svtr) | ✅ | ❌ | ✅ | ✅ |
| Model | Task / Type | Example | CUDA<br />f32 | CUDA<br />f16 | TensorRT<br />f32 | TensorRT<br />f16 |
| :---------------------------------------------------------------: | :----------------------: |:----------------------: | :-----------: | :-----------: | :------------------------: | :-----------------------: |
| **[YOLOv8-detection](https://github.com/ultralytics/ultralytics)** | Object Detection | [demo](examples/yolov8) | ✅ | ✅ | ✅ | ✅ |
| **[YOLOv8-pose](https://github.com/ultralytics/ultralytics)** | Keypoint Detection | [demo](examples/yolov8) | ✅ | ✅ | ✅ | ✅ |
| **[YOLOv8-classification](https://github.com/ultralytics/ultralytics)** | Classification | [demo](examples/yolov8) | ✅ | ✅ | ✅ | ✅ |
| **[YOLOv8-segmentation](https://github.com/ultralytics/ultralytics)** | Instance Segmentation | [demo](examples/yolov8) | ✅ | ✅ | ✅ | ✅ |
| **[YOLOv9](https://github.com/WongKinYiu/yolov9)** | Object Detection | [demo](examples/yolov9) | ✅ | ✅ | ✅ | ✅ |
| **[RT-DETR](https://arxiv.org/abs/2304.08069)** | Object Detection | [demo](examples/rtdetr) | ✅ | ✅ | ✅ | ✅ |
| **[FastSAM](https://github.com/CASIA-IVA-Lab/FastSAM)** | Instance Segmentation | [demo](examples/fastsam) | ✅ | ✅ | ✅ | ✅ |
| **[YOLO-World](https://github.com/AILab-CVC/YOLO-World)** | Object Detection | [demo](examples/yolo-world) | ✅ | ✅ | ✅ | ✅ |
| **[DINOv2](https://github.com/facebookresearch/dinov2)** | Vision-Self-Supervised | [demo](examples/dinov2) | ✅ | ✅ | ✅ | ✅ |
| **[CLIP](https://github.com/openai/CLIP)** | Vision-Language | [demo](examples/clip) | ✅ | ✅ | ✅ visual<br />❌ textual | ✅ visual<br />❌ textual |
| **[BLIP](https://github.com/salesforce/BLIP)** | Vision-Language | [demo](examples/blip) | ✅ | ✅ | ✅ visual<br />❌ textual | ✅ visual<br />❌ textual |
| [**DB**](https://arxiv.org/abs/1911.08947) | Text Detection | [demo](examples/db) | ✅ | ❌ | ✅ | ✅ |
| [**SVTR**](https://arxiv.org/abs/2205.00159) | Text Recognition | [demo](examples/svtr) | ✅ | ❌ | ✅ | ✅ |
| [**RTMO**](https://github.com/open-mmlab/mmpose/tree/main/projects/rtmo) | Keypoint Detection | [demo](examples/rtmo) | ✅ | ✅ | ❌ | ❌ |
## Solution Models
Additionally, this repo also provides some solution models such as pedestrian `fall detection`, `head detection`, `trash detection`, and more.
Additionally, this repo also provides some solution models.
| Model | Example |
| :--------------------------------------------------------------------------------: | :------------------------------: |

BIN
examples/assets/bus.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 134 KiB

BIN
examples/assets/falldown.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 68 KiB

BIN
examples/assets/kids.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 85 KiB

BIN
examples/assets/math.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 54 KiB

BIN
examples/assets/trash.jpg
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 272 KiB

1
examples/blip/README.md
View File

@ -47,7 +47,6 @@ cargo run -r --example blip
## TODO
* [ ] text decode with Top-p sample
* [ ] VQA
* [ ] Retrival
* [ ] TensorRT support for textual model

1
examples/blip/main.rs
View File

@ -10,6 +10,7 @@ fn main() -> Result<(), Box<dyn std::error::Error>> {
// textual
let options_textual = Options::default()
.with_model("../models/blip-textual-base.onnx")
.with_tokenizer("tokenizer-blip.json")
.with_i00((1, 1, 4).into()) // input_id: batch
.with_i01((1, 1, 4).into()) // input_id: seq_len
.with_i10((1, 1, 4).into()) // attention_mask: batch

1
examples/clip/main.rs
View File

@ -10,6 +10,7 @@ fn main() -> Result<(), Box<dyn std::error::Error>> {
// textual
let options_textual = Options::default()
.with_model("../models/clip-b32-textual-dyn.onnx")
.with_tokenizer("tokenizer-clip.json")
.with_i00((1, 1, 4).into())
.with_profile(false);

2
examples/db/README.md
View File

@ -8,7 +8,7 @@ cargo run -r --example db
### 1. Donwload ONNX Model
[ppocr-v3-db-dyn](https://github.com/jamjamjon/assets/releases/download/v0.0.1/ppocr-v3-db-dyn.onnx)
[ppocr-v3-db-dyn](https://github.com/jamjamjon/assets/releases/download/v0.0.1/ppocr-v3-db-dyn.onnx)
[ppocr-v4-db-dyn](https://github.com/jamjamjon/assets/releases/download/v0.0.1/ppocr-v4-db-dyn.onnx)
### 2. Specify the ONNX model path in `main.rs`

1
examples/fastsam/README.md
View File

@ -25,7 +25,6 @@ cargo run -r --example fastsam
```Rust
let options = Options::default()
.with_model("../models/FastSAM-s-dyn-f16.onnx") // <= modify this
.with_saveout("FastSAM")
.with_profile(false);
let mut model = YOLO::new(&options)?;
```

1
examples/rtdetr/README.md
View File

@ -23,7 +23,6 @@ cargo run -r --example rtdetr
```Rust
let options = Options::default()
.with_model("ONNX_MODEL") // <= modify this
.with_saveout("RT-DETR");
```
### 3. Then, run

35
examples/rtmo/README.md Normal file
View File

@ -0,0 +1,35 @@
## Quick Start
```shell
cargo run -r --example rtmo
```
## Or you can manully
### 1. Donwload ONNX Model
[rtmo-s-dyn model](https://github.com/jamjamjon/assets/releases/download/v0.0.1/rtmo-s-dyn.onnx)
[rtmo-m-dyn model](https://github.com/jamjamjon/assets/releases/download/v0.0.1/rtmo-m-dyn.onnx)
[rtmo-l-dyn model](https://github.com/jamjamjon/assets/releases/download/v0.0.1/rtmo-l-dyn.onnx)
[rtmo-s-dyn-f16 model](https://github.com/jamjamjon/assets/releases/download/v0.0.1/rtmo-s-dyn-f16.onnx)
[rtmo-m-dyn-f16 model](https://github.com/jamjamjon/assets/releases/download/v0.0.1/rtmo-m-dyn-f16.onnx)
[rtmo-l-dyn-f16 model](https://github.com/jamjamjon/assets/releases/download/v0.0.1/rtmo-l-dyn-f16.onnx)
### 2. Specify the ONNX model path in `main.rs`
```Rust
let options = Options::default()
.with_model("ONNX_MODEL") // <= modify this
```
### 3. Then, run
```bash
cargo run -r --example rtmo
```
## Results
![](./demo.jpg)

BIN
examples/rtmo/demo.jpg Normal file
View File

Binary file not shown.
Side by Side

After

Width:  |  Height:  |  Size: 242 KiB

26
examples/rtmo/main.rs Normal file
View File

@ -0,0 +1,26 @@
use usls::{models::RTMO, Annotator, DataLoader, Options, COCO_SKELETON_17};
fn main() -> Result<(), Box<dyn std::error::Error>> {
// build model
let options = Options::default()
.with_model("../rtmo-l-dyn-f16.onnx")
.with_i00((1, 2, 8).into())
.with_nk(17)
.with_confs(&[0.3])
.with_kconfs(&[0.5]);
let mut model = RTMO::new(&options)?;
// load image
let x = vec![DataLoader::try_read("./assets/bus.jpg")?];
// run
let y = model.run(&x)?;
// // annotate
let annotator = Annotator::default()
.with_saveout("RTMO")
.with_skeletons(&COCO_SKELETON_17);
annotator.annotate(&x, &y);
Ok(())
}

6
examples/svtr/README.md
View File

@ -8,9 +8,9 @@ cargo run -r --example svtr
### 1. Donwload ONNX Model
[ppocr-v4-server-svtr-ch-dyn](https://github.com/jamjamjon/assets/releases/download/v0.0.1/ppocr-v4-server-svtr-ch-dyn.onnx)
[ppocr-v4-svtr-ch-dyn](https://github.com/jamjamjon/assets/releases/download/v0.0.1/ppocr-v4-svtr-ch-dyn.onnx)
[ppocr-v3-svtr-ch-dyn](https://github.com/jamjamjon/assets/releases/download/v0.0.1/ppocr-v3-svtr-ch-dyn.onnx)
[ppocr-v4-server-svtr-ch-dyn](https://github.com/jamjamjon/assets/releases/download/v0.0.1/ppocr-v4-server-svtr-ch-dyn.onnx)
[ppocr-v4-svtr-ch-dyn](https://github.com/jamjamjon/assets/releases/download/v0.0.1/ppocr-v4-svtr-ch-dyn.onnx)
[ppocr-v3-svtr-ch-dyn](https://github.com/jamjamjon/assets/releases/download/v0.0.1/ppocr-v3-svtr-ch-dyn.onnx)
### 2. Specify the ONNX model path in `main.rs`

2
examples/yolo-world/README.md
View File

@ -10,7 +10,7 @@ cargo run -r --example yolo-world
- Download
[yolov8s-world-v2-shoes](https://github.com/jamjamjon/assets/releases/download/v0.0.1/yolov8s-world-v2-shoes.onnx)
[yolov8s-world-v2-shoes](https://github.com/jamjamjon/assets/releases/download/v0.0.1/yolov8s-world-v2-shoes.onnx)
- Or generate your own `yolo-world` model and then Export
- Installation

2
examples/yolov8-face/README.md
View File

@ -8,7 +8,7 @@ cargo run -r --example yolov8-face
### 1. Donwload ONNX Model
[yolov8-face-dyn-f16](https://github.com/jamjamjon/assets/releases/download/v0.0.1/yolov8-face-dyn-f16.onnx)
[yolov8-face-dyn-f16](https://github.com/jamjamjon/assets/releases/download/v0.0.1/yolov8-face-dyn-f16.onnx)
### 2. Specify the ONNX model path in `main.rs`

2
examples/yolov8-falldown/README.md
View File

@ -8,7 +8,7 @@ cargo run -r --example yolov8-falldown
### 1.Donwload ONNX Model
[yolov8-falldown-f16](https://github.com/jamjamjon/assets/releases/download/v0.0.1/yolov8-falldown-f16.onnx)
[yolov8-falldown-f16](https://github.com/jamjamjon/assets/releases/download/v0.0.1/yolov8-falldown-f16.onnx)
### 2. Specify the ONNX model path in `main.rs`

2
examples/yolov8-head/README.md
View File

@ -8,7 +8,7 @@ cargo run -r --example yolov8-head
### 1. Donwload ONNX Model
[yolov8-head-f16](https://github.com/jamjamjon/assets/releases/download/v0.0.1/yolov8-head-f16.onnx)
[yolov8-head-f16](https://github.com/jamjamjon/assets/releases/download/v0.0.1/yolov8-head-f16.onnx)
### 2. Specify the ONNX model path in `main.rs`

2
examples/yolov8-trash/README.md
View File

@ -10,7 +10,7 @@ cargo run -r --example yolov8-trash
### 1. Donwload ONNX Model
[yolov8-plastic-bag-f16](https://github.com/jamjamjon/assets/releases/download/v0.0.1/yolov8-plastic-bag-f16.onnx)
[yolov8-plastic-bag-f16](https://github.com/jamjamjon/assets/releases/download/v0.0.1/yolov8-plastic-bag-f16.onnx)
### 2. Specify the ONNX model path in `main.rs`

1
examples/yolov8/README.md
View File

@ -38,7 +38,6 @@ yolo export model=yolov8m-seg.pt format=onnx simplify
let options = Options::default()
.with_model("ONNX_PATH") // <= modify this
.with_confs(&[0.4, 0.15]) // person: 0.4, others: 0.15
.with_saveout("YOLOv8");
let mut model = YOLO::new(&options)?;
```

1
examples/yolov8/main.rs
View File

@ -33,6 +33,7 @@ fn main() -> Result<(), Box<dyn std::error::Error>> {
// run & annotate
for (xs, _paths) in dl {
let ys = model.run(&xs)?;
println!("{:?}", ys);
annotator.annotate(&xs, &ys);
}

3
examples/yolov9/README.md
View File

@ -10,7 +10,7 @@ cargo run -r --example yolov9
- **Download**
[yolov9-c-dyn-fp16](https://github.com/jamjamjon/assets/releases/download/v0.0.1/yolov9-c-dyn-f16.onnx)
[yolov9-c-dyn-fp16](https://github.com/jamjamjon/assets/releases/download/v0.0.1/yolov9-c-dyn-f16.onnx)
- **Export**
```shell
@ -31,7 +31,6 @@ cargo run -r --example yolov9
```Rust
let options = Options::default()
.with_model("ONNX_PATH") // <= modify this
.with_saveout("YOLOv9");
```
### 3. Run

8
src/models/blip.rs
View File

@ -4,7 +4,7 @@ use ndarray::{s, Array, Axis, IxDyn};
use std::io::Write;
use tokenizers::Tokenizer;
use crate::{auto_load, ops, LogitsSampler, MinOptMax, Options, OrtEngine, TokenizerStream};
use crate::{ops, LogitsSampler, MinOptMax, Options, OrtEngine, TokenizerStream};
#[derive(Debug)]
pub struct Blip {
@ -27,11 +27,7 @@ impl Blip {
visual.height().to_owned(),
visual.width().to_owned(),
);
let tokenizer = match &options_textual.tokenizer {
None => auto_load("tokenizer-blip.json")?,
Some(tokenizer) => tokenizer.into(),
};
let tokenizer = Tokenizer::from_file(tokenizer).unwrap();
let tokenizer = Tokenizer::from_file(&options_textual.tokenizer.unwrap()).unwrap();
let tokenizer = TokenizerStream::new(tokenizer);
visual.dry_run()?;
textual.dry_run()?;

9
src/models/clip.rs
View File

@ -1,7 +1,6 @@
use crate::{auto_load, ops, MinOptMax, Options, OrtEngine};
use crate::{ops, MinOptMax, Options, OrtEngine};
use anyhow::Result;
use image::DynamicImage;
// use itertools::Itertools;
use ndarray::{Array, Array2, Axis, IxDyn};
use tokenizers::{PaddingDirection, PaddingParams, PaddingStrategy, Tokenizer};
@ -28,11 +27,7 @@ impl Clip {
visual.inputs_minoptmax()[0][2].to_owned(),
visual.inputs_minoptmax()[0][3].to_owned(),
);
let tokenizer = match &options_textual.tokenizer {
None => auto_load("tokenizer-clip.json").unwrap(),
Some(tokenizer) => tokenizer.into(),
};
let mut tokenizer = Tokenizer::from_file(tokenizer).unwrap();
let mut tokenizer = Tokenizer::from_file(&options_textual.tokenizer.unwrap()).unwrap();
tokenizer.with_padding(Some(PaddingParams {
strategy: PaddingStrategy::Fixed(context_length),
direction: PaddingDirection::Right,

2
src/models/mod.rs
View File

@ -3,6 +3,7 @@ mod clip;
mod db;
mod dinov2;
mod rtdetr;
mod rtmo;
mod svtr;
mod yolo;
@ -11,5 +12,6 @@ pub use clip::Clip;
pub use db::DB;
pub use dinov2::Dinov2;
pub use rtdetr::RTDETR;
pub use rtmo::RTMO;
pub use svtr::SVTR;
pub use yolo::YOLO;

1
src/models/rtdetr.rs
View File

@ -41,7 +41,6 @@ impl RTDETR {
.expect("Failed to get num_classes, make it explicit with `--nc`")
.len(),
);
// let annotator = Annotator::default();
let confs = DynConf::new(&options.confs, nc);
engine.dry_run()?;

133
src/models/rtmo.rs Normal file
View File

@ -0,0 +1,133 @@
use anyhow::Result;
use image::DynamicImage;
use ndarray::{Array, Axis, IxDyn};
use crate::{ops, Bbox, DynConf, Keypoint, MinOptMax, Options, OrtEngine, Ys};
#[derive(Debug)]
pub struct RTMO {
engine: OrtEngine,
height: MinOptMax,
width: MinOptMax,
batch: MinOptMax,
confs: DynConf,
kconfs: DynConf,
}
impl RTMO {
pub fn new(options: &Options) -> Result<Self> {
let engine = OrtEngine::new(options)?;
let (batch, height, width) = (
engine.batch().to_owned(),
engine.height().to_owned(),
engine.width().to_owned(),
);
let nc = 1;
let nk = options.nk.unwrap_or(17);
let confs = DynConf::new(&options.kconfs, nc);
let kconfs = DynConf::new(&options.kconfs, nk);
engine.dry_run()?;
Ok(Self {
engine,
confs,
kconfs,
height,
width,
batch,
})
}
pub fn run(&mut self, xs: &[DynamicImage]) -> Result<Vec<Ys>> {
let xs_ = ops::letterbox(xs, self.height() as u32, self.width() as u32, 114.0)?;
let ys = self.engine.run(&[xs_])?;
let ys = self.postprocess(ys, xs)?;
Ok(ys)
}
pub fn postprocess(&self, xs: Vec<Array<f32, IxDyn>>, xs0: &[DynamicImage]) -> Result<Vec<Ys>> {
let mut ys: Vec<Ys> = Vec::new();
let (preds_bboxes, preds_kpts) = if xs[0].ndim() == 3 {
(&xs[0], &xs[1])
} else {
(&xs[1], &xs[0])
};
for (idx, (batch_bboxes, batch_kpts)) in preds_bboxes
.axis_iter(Axis(0))
.zip(preds_kpts.axis_iter(Axis(0)))
.enumerate()
{
let width_original = xs0[idx].width() as f32;
let height_original = xs0[idx].height() as f32;
let ratio =
(self.width() as f32 / width_original).min(self.height() as f32 / height_original);
let mut y_bboxes = Vec::new();
let mut y_kpts: Vec<Vec<Keypoint>> = Vec::new();
for (xyxyc, kpts) in batch_bboxes
.axis_iter(Axis(0))
.zip(batch_kpts.axis_iter(Axis(0)))
{
// bbox
let x1 = xyxyc[0] / ratio;
let y1 = xyxyc[1] / ratio;
let x2 = xyxyc[2] / ratio;
let y2 = xyxyc[3] / ratio;
let confidence = xyxyc[4];
if confidence < self.confs[0] {
continue;
}
let y_bbox = Bbox::new(
(
(
x1.max(0.0f32).min(width_original),
y1.max(0.0f32).min(height_original),
),
(x2, y2),
)
.into(),
0,
confidence,
Some(String::from("Person")),
);
y_bboxes.push(y_bbox);
// keypoints
let mut kpts_ = Vec::new();
for (i, kpt) in kpts.axis_iter(Axis(0)).enumerate() {
let x = kpt[0] / ratio;
let y = kpt[1] / ratio;
let c = kpt[2];
if c < self.kconfs[i] {
kpts_.push(Keypoint::default());
} else {
kpts_.push(Keypoint::new(
(
x.max(0.0f32).min(width_original),
y.max(0.0f32).min(height_original),
)
.into(),
c,
));
}
}
y_kpts.push(kpts_);
}
ys.push(Ys::default().with_bboxes(&y_bboxes).with_keypoints(&y_kpts));
}
Ok(ys)
}
pub fn batch(&self) -> isize {
self.batch.opt
}
pub fn width(&self) -> isize {
self.width.opt
}
pub fn height(&self) -> isize {
self.height.opt
}
}

1
src/models/yolo.rs
View File

@ -34,7 +34,6 @@ pub struct YOLO {
confs: DynConf,
kconfs: DynConf,
iou: f32,
// saveout: Option<String>,
names: Option<Vec<String>>,
apply_nms: bool,
anchors_first: bool,

10
src/options.rs
View File

@ -156,6 +156,11 @@ impl Options {
self
}
pub fn with_tokenizer(mut self, tokenizer: &str) -> Self {
self.tokenizer = Some(auto_load(tokenizer).unwrap());
self
}
pub fn with_unclip_ratio(mut self, x: f32) -> Self {
self.unclip_ratio = x;
self
@ -206,11 +211,6 @@ impl Options {
self
}
pub fn with_tokenizer(mut self, tokenizer: String) -> Self {
self.tokenizer = Some(tokenizer);
self
}
pub fn with_i00(mut self, x: MinOptMax) -> Self {
self.i00 = Some(x);
self