tfjs
tfjs copied to clipboard
tensorflow
Error: Requested texture size [-1536x512] is invalid
Please make sure that this is a bug. As per our GitHub Policy, we only address code/doc bugs, performance issues, feature requests and build/installation issues on GitHub. tag:bug_template
System information
- Have I written custom code (as opposed to using a stock example script provided in TensorFlow.js): No
- OS Platform and Distribution (e.g., Linux Ubuntu 16.04): Windows
- TensorFlow.js installed from (npm or script link): script link
- TensorFlow.js version (use command below): 3.9.0
- Browser version: Chrome 94.0.4606.54
- Tensorflow.js Converter Version: Most recent
Describe the current behavior I am running two object detection models. The first locates where a sudoku board is in an image, and the second model locates where each grid is on the sudoku board. The first model works fine. However, when I call executeAsync to the model to find where each grid is on sliced tensor (This comes from the predictions made from the first model) I get an error. So the first object detection models works, but upon calling the second object detection model I get the error : tf.min.js:17 Uncaught (in promise) Error: Requested texture size [-1536x512] is invalid. at tf.min.js:17 at jX (tf.min.js:17) at tf.min.js:17 at e.t.createFloat32MatrixTexture (tf.min.js:17) at e.t.acquireTexture (tf.min.js:17) at t.n.acquireTexture (tf.min.js:17) at t.n.uploadToGPU (tf.min.js:17) at t.n.runWebGLProgram (tf.min.js:17) at Object.kernelFunc (tf.min.js:17) at n (tf.min.js:17)
Describe the expected behavior It is should work exactly how the first model works. Its weird because the first model is able to make predictions but for the second model it fails.
Standalone code to reproduce the issue Provide a reproducible test case that is the bare minimum necessary to generate the problem. If possible, please share a link to Colab/CodePen/any notebook.--This is the image I was loading to the html input
Other info / logs Include any logs or source code that would be helpful to diagnose the problem. If including tracebacks, please include the full traceback. Large logs and files should be attached.
Code: HTML:
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8">
<title>TF JS example</title>
<script src="https://cdn.jsdelivr.net/npm/@tensorflow/[email protected]/dist/tf.min.js"></script>
</head>
<body>
<form method="POST" action="/upload" enctype="multipart/form-data">
<p><input type="file" id="file1" name="file"></p>
</form>
<canvas style="border: 10px solid black " id="canvas" >
<img id ="display" width = "..." height="..." >
</canvas>
<script src="main.js"></script>
</body>
</html>
main.js:
async function load() {
const model = await tf.loadGraphModel('https://big-g.s3.us-east.cloud-object-storage.appdomain.cloud/model.json');
return model;
};
async function load_s(){
const model_s = await tf.loadGraphModel('https://small-g.s3.us-east.cloud-object-storage.appdomain.cloud/model.json');
return model_s
}
async function load_class(){
const model_class = await tf.loadGraphModel("https://classy.s3.us-east.cloud-object-storage.appdomain.cloud/model.json");
return model_class
}
var model = load();
var model_s = load_s();
//var model_class =load_class();
console.log("laoded")
document.querySelector('input').addEventListener('change', function() {
console.log("change")
var reader = new FileReader();
const fileInput = document.getElementById("file1").files[0];
console.log(fileInput);
readerDim = new FileReader();
readerDim.readAsDataURL(fileInput);
readerDim.onload = function (e) {
//Initiate the JavaScript Image object.
var image = new Image();
//Set the Base64 string return from FileReader as source.
image.src = e.target.result;
//Validate the File Height and Width.
image.onload = function () {
var height = this.height;
var width = this.width;
console.log(width,height);//width,height
var a = tf.browser.fromPixels(image);
//convert image into 3 channel grayscale
var g_scale =a.mean(2).expandDims(-1);
var a = tf.image.grayscaleToRGB(g_scale).cast("int32");
a=a.expandDims();
model.then(model => {
async function pred() {
console.log('Start');
const result = await model.executeAsync(a);
console.log('End')
//console.log( await result[6].array());
//const classes = await result[2].array();
//const accuracy = await result[4].array();
const boxes = await result[6].array(); //y,x,height,width
//Time to draw
const canvas = document.querySelector('#canvas');
const ctx = canvas.getContext('2d');
//resize
canvas.height = image.height;
canvas.width= image.width;
// ctx.drawImage(image,0,0);
tf.browser.toPixels(a.squeeze(),canvas);
const [y,x,height,width] =boxes[0][0];
//console.log(y);
console.log("drawing...");
const xc = Math.round(x*image.width);
const yc = Math.round(y*image.height);
const x2c = Math.round(width*image.width);
const y2c = Math.round(height*image.height);
console.log(xc,yc,x2c ,y2c);//x1, y1, x2, y2, not height or width
//ctx.rect(xc, yc, x2c-xc, y2c-yc);
//ctx.stroke();
console.log(a.shape)
sliced = a.slice([0,yc,xc,0],[1,y2c-yc,x2c-xc,3] );
console.log(sliced)
//await new Promise(r => setTimeout(r, 1000));
//ctx.clearRect(0, 0, canvas.width, canvas.height)
tf.browser.toPixels(sliced.squeeze(),canvas);
model_s.then(model_s =>{
async function small_g(sliced){
console.log("predicting small",sliced);
const result_s = await model_s.executeAsync(sliced);
return result_s
}
const result_s = small_g(sliced);
console.log(result_s);
});
};
console.log('Predicting')
var box = pred();
});
};
};
}, false);
Logs:
change
main.js:27 File {name: 'test_img.png', lastModified: 1631819149502, lastModifiedDate: Thu Sep 16 2021 15:05:49 GMT-0400 (Eastern Daylight Time), webkitRelativePath: '', size: 259253, …}
main.js:45 1268 784
main.js:104 Predicting
main.js:56 Start
main.js:58 End
main.js:76 drawing...
main.js:82 117 81 660 624
main.js:85 (4) [1, 784, 1268, 3]
main.js:87 e {kept: false, isDisposedInternal: false, shape: Array(4), dtype: 'int32', size: 884547, …}
main.js:95 predicting small e {kept: false, isDisposedInternal: false, shape: Array(4), dtype: 'int32', size: 884547, …}
main.js:100 Promise {
![google-ml-butler[bot] avatar](https://avatars.githubusercontent.com/in/43729?v=4 "Published by a pub.dev verified publisher"){kind=small}
Any tensor I use even a random one gives this error. I tried recreating the model from the tfjs-converter but that didn't work. I think the error is that this is a center-net model. so I will just use a different model
Hi, @mkaufmann84
Apologize for the delayed response and we're re-visiting our older issues and checking whether those issues got resolved or not as of now so May I know are you still looking for the solution or your issue got resolved ?
If issue still persists after trying with latest version of TFJs please let us know with error log and code snippet to replicate the same issue from our end ?
Could you please confirm if this issue is resolved for you ? Please feel free to close the issue if it is resolved ? Thank you!
I think what I did is I used a different model. Idc about this issue anymore u can resolve it
On Tue, May 23, 2023 at 3:30 AM gaikwadrahul8 @.***> wrote:
Hi, @mkaufmann84 https://github.com/mkaufmann84
Apologize for the delayed response and we're re-visiting our older issues and checking whether those issues got resolved or not as of now so May I know are you still looking for the solution or your issue got resolved ?
If issue still persists after trying with latest version of TFJs please let us know with error log and code snippet to replicate the same issue from our end ?
Could you please confirm if this issue is resolved for you ? Please feel free to close the issue if it is resolved ? Thank you!
— Reply to this email directly, view it on GitHub https://github.com/tensorflow/tfjs/issues/5664#issuecomment-1558694285, or unsubscribe https://github.com/notifications/unsubscribe-auth/AE7SXDSIG3JCL4HS5FVWJNLXHRRQTANCNFSM5EX6UGBQ . You are receiving this because you were mentioned.Message ID: @.***>
Hi, @mkaufmann84
Apologize for the delayed response and Good to hear that your problem got solved by using different model. Could you please confirm if this issue is resolved for you ? Please feel free to close the issue if it is resolved ? Thank you!
I same error:
Unhandled Promise Rejection: Error: Requested texture size [0x0] is invalid.
