ultralytics/hub

Ultralytics HUB tutorials and support

中文 | 한국어 | 日本語 | Русский | Deutsch | Français | Español | Português | Türkçe | Tiếng Việt | हिन्दी | العربية

👋 Hello from the Ultralytics Team! We've been working hard these last few months to launch Ultralytics HUB, a new web tool for training and deploying all your YOLOv5 and YOLOv8 🚀 models from one spot!

We hope that the resources here will help you get the most out of HUB. Please browse the HUB Docs for details, raise an issue on GitHub for support, and join our Discord community for questions and discussions!

🗂️ 1. Upload a Dataset

Ultralytics HUB datasets align with the format used by YOLOv5 and YOLOv8. They adhere to the same structural and labeling conventions, ensuring a seamless experience.

Dataset Preparation:

Ensure that the YAML file describing your dataset is placed in the root directory of your dataset, as illustrated below. Once in place, zip the directory for uploading to Ultralytics HUB. The dataset YAML, its directory, and the zip file should all bear the identical name.

For instance, with a dataset named 'coco8', as shown in ultralytics/hub/example_datasets/coco8.zip, include a coco8.yaml within the coco8/ directory. Zip this to form coco8.zip for upload with the command:

# Zipping the dataset directory for upload
zip -r coco8.zip coco8

Explore the provided example_datasets/coco8.zip for insights on structuring your custom dataset.

YAML Specifications:

The dataset YAML follows the established YOLOv5 and YOLOv8 format. For exhaustive instructions, refer to the Train Custom Data tutorial.

# YAML configuration example for custom datasets
path: # dataset root dir (leave empty for HUB integration)
train: images/train # train images (relative to 'path') 8 images
val: images/val # val images (relative to 'path') 8 images
test: # test images (optional)

# Class labels
names:
  0: person
  1: bicycle
  2: car
  3: motorcycle
  # Add more classes as needed

Upload your zipped dataset to Ultralytics HUB by logging in, navigating to the 'Datasets' tab, and selecting 'Upload Dataset'. This lets you scan and view your dataset prior to training YOLOv5 or YOLOv8 models.

🚀 2. Train a Model

Connect to the Ultralytics HUB notebook and employ your model API key to embark on model training!

🌐 3. Deploy to the Real World

Transition your model to 13 different formats including TensorFlow, ONNX, OpenVINO, CoreML, Paddle, and more. Operate your models directly on your iOS or Android mobile device by downloading the Ultralytics App!

❓ Have Issues or Questions?

For those just embarking on their Ultralytics HUB journey, the Issues tab is your go-to resource for support. Click the 'New Issue' button and share your thoughts or questions. Our aim is to enhance your experience with invaluable solutions and improvements! 😃

hub's Issues

Upload not working dataset

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

No response

Bug

error: Dataset Error : data.yaml/images/train not found, data.yaml/labels/train not found, data.yaml/images/valid not found, data.yaml/labels/valid not found
i tried to make the data file exactly like in the coco6.yalm
data.yalm :

path:
train: images/train
val: images/valid

nc: 4
names: ['black_measles', 'black_root', 'blight', 'healthy']

one document annotation:
1 0.60703125 0.45 0.03671875 0.034375
1 0.5375 0.490625 0.034375 0.03125
1 0.6875 0.534375 0.04375 0.04765625
1 0.63984375 0.55703125 0.03515625 0.03203125
1 0.4390625 0.559375 0.02890625 0.025
1 0.671875 0.64140625 0.03203125 0.02734375
1 0.6140625 0.6421875 0.03203125 0.03125
1 0.82109375 0.71640625 0.053125 0.05
1 0.44375 0.16640625 0.0421875 0.0359375
1 0.453125 0.275 0.03203125 0.0328125
1 0.43984375 0.359375 0.04296875 0.03515625

I tried to make the zip file with the same name with the yalm file still not working and getting the same error

Environment

No response

Minimal Reproducible Example

No response

Additional

No response

How to raise an Issue about Ultralytics HUB

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

If you are a new Ultralytics HUB user and have questions or comments for Ultralytics, you are in the right place! Please click the New Issue button in the Issues tab and let us know what we can do to improve your experience on Ultralytics HUB 😃!

Additional

No response

How to Use

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

Hi is there a detailed guide on how to use this program for beginners? I am trying to annotate a dataset, train a custom model, and have yolo identify and count objects in a set of photos

Additional

No response

Not all models are available for training

Search before asking

• I have searched the HUB issues and found no similar feature requests.

Description

I can't train the models: YOLOv5n6, YOLOv5s6, YOLOv5m6, YOLOv5l6, YOLOv5x6

Use case

No response

Additional

No response

Model won’t complete download on app

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

Models, Export

Bug

I have the Ultralytics app downloaded on an iOS device running iOS 15.1. The app works great when running the Ultralytics demo models. But when I try to download my own trained model the model seems to stall out repeatedly when 3/4 of the way complete and eventually a red exclamation point appears. I have checked the available memory on my device and should have plenty available.

Environment

-mobile device iPhone XR RUNNING iOS 15.1

Minimal Reproducible Example

Download Ultralytics app from AppStore sign into my Ultralytics hub account and download custom model. This is where I eventually stall out.

Additional

Keep up the awesome work!

Upload dataset

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

Dataset upload shows error. but what error not clear. make dataset as per coco6 example. coco6 is uploading correct but , other could not be upload.

Additional

No response

Product Roadmap Feature List

Hi All,

We are always looking to improve the Ultralytics HUB and to provide features that are most useful to our users. 🚀

In Progress

• Ultralytics Cloud training with GPU (K80, P4, T4, P100, V100, A100)
• Plan upgrades to meet your requirements
• Custom Dataset Validation. We are working to improve the validation of your custom datasets to prevent simple errors.
• UI/UX Updates.

Coming Soon

• Marketplace. A one stop shop to datasets and models created by other users, or share your creations with the world.
• Teams. Collaborate within your team by sharing projects, models and datasets privately.
• Public Previews. Share your model publicly with limited access so your friends can see what you have created.
• Resume Training. If your training stops midway resume from the last checkpoint to prevent lost time.
• Concat Datasets. Combine datasets or create a custom one by selecting only the classes you need from existing datasets.

No Storage Available

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

Datasets

Bug

When uploading a dataset, upload fails with the error "No Storage Available".

This happens even after deleting all my datasets and all models save one then trying to re-upload a dataset that previously worked.

Additionally my dashboard shows 177.3 GB / 100.0 GB for Storage Used. I currently have zero datasets uploaded and a single model that lists its size as 146 MB.

Environment

Linux (Arch)
Brave

Minimal Reproducible Example

1. Login to my Hub account
2. Try to upload a dataset
3. Get incorrect storage used numbers and an error message mentioning "No Storage"

Additional

No response

Model upload

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

can I upload my already trained model to your service?

Additional

didn't see how to do it

Can't Login in Android App Ultralytics

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

No response

Bug

I can't log in my acc in gmail and also github acc. My log in details are correct, I'm sure. Can you help me please?

Environment

No response

Minimal Reproducible Example

No response

Additional

No response

HUB- local agent: CUDA not detected

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

Hello,

When using the HUB local agent (on Windows)- how can I get it to utilise the graphics card ?
I get an error 'CUDA not detected' message

• I have tried COLAB (I have a Pro subscription)- but as is my usual experience with COLAB- it gets 90% finished, then disconnects, and won't let me back on with a GPU- very annoying....!
• I do also have a Linux/RTX3060 machine which I can try, when it has finished on another job

thank you

Andrew

Additional

No response

Dataset Guidelines

Ultralytics HUB datasets are just like YOLOv5 🚀 datasets, they use the same structure and the same label formats to keep everything simple.

1. Upload a Dataset

When you upload a dataset to Ultralytics HUB, make sure to place your dataset YAML inside the dataset root directory as in the example shown below, and then zip for upload to https://hub.ultralytics.com/. Your dataset YAML, directory and zip should all share the same name. For example, if your dataset is called 'coco6' as in our example ultralytics/hub/coco6.zip, then you should have a coco6.yaml inside your coco6/ directory, which should zip to create coco6.zip for upload:

zip -r coco6.zip coco6

The example coco6.zip dataset in this repository can be downloaded and unzipped to see exactly how to structure your custom dataset.

The dataset YAML is the same standard YOLOv5 YAML format. See the YOLOv5 Train Custom Data tutorial for full details.

# Train/val/test sets as 1) dir: path/to/imgs, 2) file: path/to/imgs.txt, or 3) list: [path/to/imgs1, path/to/imgs2, ..]
path:  # dataset root dir (leave empty for HUB)
train: images/train  # train images (relative to 'path') 8 images
val: images/val  # val images (relative to 'path') 8 images
test:  # test images (optional)

# Classes
names:
  0: person
  1: bicycle
  2: car
  3: motorcycle
  ...

After zipping your dataset, sign in to Ultralytics HUB and click the Datasets tab. Click 'Upload Dataset' to upload, scan and visualize your new dataset before training new YOLOv5 models on it!

Connect Google Colab JSONDecodeError: Expecting value: line 1 column 1 (char 0)

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

Models

Bug

First Step: I have create colab and run:

%pip install ultralytics  # install
import ultralytics
ultralytics.checks()  # checks

-> Success

Second step:

ultralytics.start('XXXXXX')  # my API key

-> Get error

[/usr/local/lib/python3.7/dist-packages/ultralytics/__init__.py](https://localhost:8080/#) in start(key)
     59     # key = 'ac0ab020186aeb50cc4c2a5272de17f58bbd2c0.RqFCDNBxgU4mOLmaBrcd'  # example
     60     # api_key='ac0ab020186aeb50cc4c2a5272de17f58bbd2c0', model_key='RqFCDNBxgU4mOLmaBrcd'  # example
---> 61     train_model(api_key=api_key, model_key=model_key)

[/usr/local/lib/python3.7/dist-packages/ultralytics/main.py](https://localhost:8080/#) in train_model(api_key, model_key)
     15     # connect_to_hub(api_key=api_key)  # deprecated
     16     AUTH.api_key = api_key
---> 17     trainer = Trainer(model_id=model_key, auth=AUTH)  # No model so next in train queue is fetched
     18     if trainer.model is not None:
     19         trainer.start()

[/usr/local/lib/python3.7/dist-packages/ultralytics/trainer.py](https://localhost:8080/#) in __init__(self, model_id, auth)
     12     def __init__(self, model_id, auth):
     13         self.auth = auth
---> 14         self.model = self._get_model(model_id)
     15         if self.model is not None:
     16             self._connect_callbacks()

[/usr/local/lib/python3.7/dist-packages/ultralytics/trainer.py](https://localhost:8080/#) in _get_model(self, model_id)
     33         try:
     34             r = requests.post(api_url, json=payload)
---> 35             res = r.json()
     36 
     37             if res["data"] is None:

[/usr/local/lib/python3.7/dist-packages/requests/models.py](https://localhost:8080/#) in json(self, **kwargs)
    896                     # used.
    897                     pass
--> 898         return complexjson.loads(self.text, **kwargs)
    899 
    900     @property

[/usr/lib/python3.7/json/__init__.py](https://localhost:8080/#) in loads(s, encoding, cls, object_hook, parse_float, parse_int, parse_constant, object_pairs_hook, **kw)
    346             parse_int is None and parse_float is None and
    347             parse_constant is None and object_pairs_hook is None and not kw):
--> 348         return _default_decoder.decode(s)
    349     if cls is None:
    350         cls = JSONDecoder

[/usr/lib/python3.7/json/decoder.py](https://localhost:8080/#) in decode(self, s, _w)
    335 
    336         """
--> 337         obj, end = self.raw_decode(s, idx=_w(s, 0).end())
    338         end = _w(s, end).end()
    339         if end != len(s):

[/usr/lib/python3.7/json/decoder.py](https://localhost:8080/#) in raw_decode(self, s, idx)
    353             obj, end = self.scan_once(s, idx)
    354         except StopIteration as err:
--> 355             raise JSONDecodeError("Expecting value", s, err.value) from None
    356         return obj, end

JSONDecodeError: Expecting value: line 1 column 1 (char 0)

Environment

Ultralytics HUB Version
v0.1.14-beta.1
Client User Agent
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/101.0.4951.67 Safari/537.36
Operating System
Win32
Browser Window Size
1920 x 969
Server Timestamp
1653926780

Minimal Reproducible Example

No response

Additional

No response

Training Models to Show More Information

Search before asking

• I have searched the HUB issues and found no similar feature requests.

Description

When a model is in training it would be helpful to see how many epochs are completed versus the total. It would also be awesome to see an estimated time to completion and perhaps a total train time to date. It would be great to preview the model as it is in the last uploaded epoch.

• See current epoch
• See estimated training completion time
• Preview models currently in training

Use case

No response

Additional

No response

Dataset upload is not working with non standard annotations

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

Hi, I'm trying to upload my dataset on Ultralytics Hub but I keep getting the processing error. Am pretty sure I have already followed all the steps according to coco6.zip but it still doesnt work. My yaml file is written like this:

`path:
train: images/train
val: images/val

nc: 16
names: ['animal_crab', 'animal_eel', 'animal_etc', 'animal_fish', 'animal_shells', 'animal_starfish', 'plant', 'rov', 'trash_etc', 'trash_fabric', 'trash_fishing_gear', 'trash_metal', 'trash_paper', 'trash_plastic', 'trash_rubber', 'trash_wood']`

Additional

No response

Dataset processing

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

Datasets

Bug

Processing times are over 48 hours to upload relatively small datasets <500 images. Datasets have been downloaded straight from Roboflow using the YOLOv5 PyTorch option. I have had to manually change the .YAML file name to match the dataset and zipped file name. While trying to trouble shoot the issue I noticed that the downloaded file from Roboflow has a different file structure than the COCO6 dataset used as an example of what our dataset should look like. I understand that Ultralytics has no control over how the Roboflow datasets are configured....

Correcting the file structure would take manual file dragging and dropping while requiring edits to the .YAML file. If this is what is required I will just have to suck it up.. Perhaps I am missing something right in front of my face at the same time...

Environment

Ultralytics HUB Version
v0.1.10-beta.2
Client User Agent
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/15.2 Safari/605.1.15
Operating System
MacIntel
Browser Window Size
1850 x 878
Server Timestamp
1641964834

Minimal Reproducible Example

No response

Additional

No response

The word seems to be wrong

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

The word seems to be wrong

Additional

The word seems to be wrong

Sign on

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

I would like biometrics enable to sign in. Thank you.
I use an IPad Pro, Linux Ubuntu, kali, window all.
But I am sick and I am just going to suffer a long incurable tortures life. So the less I have to navigate to get to my distractions from my continuous walk hell of nonstop total body pain the better. Yours suffering and really grumpy.

Additional

Apple facial, android finger print and the recent build has facial recognition.

Datasets Processing Error

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

Datasets

Bug

@ultralytics HUB Recently,I have uploaded my custom dataset completely.But it can't be unzipped and is shown failed,processing error.I wanna to konw what leads to fail.Meanwhile,I tried to prepare my custom datasets correctly.But the result is still failed.

Environment

Environment
Ultralytics HUB Version
v0.1.14-beta.1
Client User Agent
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/101.0.4951.64 Safari/537.36
Operating System
Win32
Browser Window Size
1536 x 746
Server Timestamp
1652809519

Minimal Reproducible Example

No response

Additional

No response

Use examples don't seem to work

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

Use examples don't seem to work

Additional

No response

Slow Inference on XIAOMI 11T Pro

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

No response

Bug

Hello, we purchased a XIAOMI 11T pro in order to replicate the tests that you have perfomrmed and for which you sent a video https://user-images.githubusercontent.com/26566365/165805393-5e632cb1-7bbb-4b7e-8489-cfcef6341a1e.mp4

However on our XIAOMI 11T pro we get slower inferences and no indication on inference time:
https://drive.google.com/file/d/1-4GuUWYiGMX5ztwgE6G-0QMEz-xn7ecL/view?usp=sharing

what could it be due to?

The phone is the same and the app is from the play store.

Note that on a Pixel 3i and another Pixel 4x we do get to see the inference time

Thaks

Environment

XIAOMI 11T pro

Minimal Reproducible Example

See video

Additional

None

Can't get step 2 on Colab to work. Needs more than API key?

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

Training, API Key

Bug

Must the model name also be included in step 2 in the Google Colab? With API key inserted, it produces error, saying "Name is not defined."

Environment

Computer: PC, i7 with 16 Gig RAM, SSD
Ubuntu 20.04
Firefox browser, updated.

Minimal Reproducible Example

Started new model, selected format, selected dataset, went to Colab to train.

Additional

Ultralytics HUB Version
v0.1.14-beta.12
Client User Agent
Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:102.0) Gecko/20100101 Firefox/102.0
Operating System
Linux x86_64
Browser Window Size
1271 x 968
Server Timestamp
1658112208

Do datasets have a size limit?

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

Sorry for another question but I am not sure what the limit is on the dataset size ? I have a dataset of around 14GB and I uploaded it but the interface shows "Processing ... " for more than 4 h . Is that normal ? Can I actually train such a dataset ? with 100k images ? Image size I am seeking to train is 640 px. Please let me know or please point me in the right direction where the limits are set.

Additional

No response

Converting a custom model to mlmodel is 100% no matter what object it recognizes

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

Inference

Bug

Hi, I found a strange problem. It is normal for the customized models on the Hub to be converted into online test recognition. The models downloaded from the Ultralytics App are 100% recognized, and objects beyond the model can be recognized. The same is true for exporting Ultralytics iOS models. If you have any ideas or answers, please let me know, thank you very much!

this is datasets.
key.zip
)

Environment

Macbook Pro,Xcode

Minimal Reproducible Example

No response

Additional

No response

Filter Datasets by Class

Search before asking

• I have searched the HUB issues and found no similar feature requests.

Description

It would be great to be able to filter the classes shown in the dataset preview. At the moment there is a lot of scrolling involved to get to even the second class. I would like to be able to toggle classes on and off so that I can look at specific images and identify problems.

Use case

No response

Additional

No response

Timeframe for TensorRT export from HUB ??

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

Hi,

The TensorRT export from HUB currently says' coming soon'
I was wondering what the timeframe for this is ?
Also, I'm assuming it will export a .engine file ?

thanks

Andrew

Additional

No response

Dataset upload yaml file missing

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

I am new to the Ultralytics HUB. I thought I give it a try and I uploaded a custom dataset after creating a project etc. The upload finished (big dataset around 14 GB), then I had the "processing " phase for a few minutes then it failed saying that it didn't find the yaml file for my dataset. That's is right because I didn't upload. My dataset comes in the format required by the colab example. I didn't know up front that the yaml has to be included and now that I know that I have no idea where to add it?

Can you please clarify at what stage or where to upload the yaml file. I'll try again once that is clear because the upload take quite a while like 2h .Apart from that this looks awesome I can hardly wait to see if I can train a model.

Additional

No response

Feature Roadmap May 2022

Search before asking

• I have searched the HUB issues and found no similar feature requests.

Description

• Hard drive space check before dataset download
• Exponential standoff in final checkpoint upload fails with time cutoff, i.e. 60 min.
• HUB training CI for ultralytics staged changes and for official package
• Re-submit metrics on next metrics sync (TODO: requires payload/backend changes)

Use case

No response

Additional

No response

Dataset upload fails with large number of images

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

Hello,

I am having a problem uploading a dataset:

• I've followed the instructions, and I'm pretty certain I've got the folder structure and naming correct, because a test set with 20 images/1 class works fine.
• However, exactly the same data structure, YAML file etc but with 4000 images/4 classes fails- with a 'processing error'
  i.e in the first screenshot below constructaiv6 & constructaiv7 are identical, except for the number of images and labels
• The same dataset works fine with a local yolov5 (Docker)
• I've been through the file naming, counts etc, class IDs etc and I can't see anything wrong

I'm loving yolov5 and finding it very easy to use- but would like to try the hub option.
(A separate question- but I was wondering if there was any way to use a locally trained model on the phone app ?)

I'd be very grateful for any help with this

Andrew

Additional

No response

why isn't my dataset uploading? it is saying failed/process error

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

why isn't my dataset uploading? it is saying failed/process error

Additional

No response

Object Counter

Search before asking

• I have searched the HUB issues and found no similar feature requests.

Description

Hi,
I would like to use Ultralytics app as real time traffic counter. - I would like to count how many different objects (cars, bicycles, pedestrians,...) cross the street.

Use case

This feature should be potentionally use for urban development and smart cities. I would like to measure where is efficient to built a bicycle lanes, other infrastructure etc. - based on the counter.

Additional

No response

Is it possible to upload a pretrained YOLOv5 model to HUB?

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

We have trained a model for wildlife species detection for 71 animal species in Australian national parks. Is it possible for us to upload this into the hub and use it there?

Additional

No response

How should I format my custom dataset?

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

No response

Bug

Every time I upload my dataset i get an error says Zip Formatted Incorrectly
I tried to modify my zip file and the same I even tried to download some random dataset from roboflow and upload it and still same problem

Environment

Browser Chorme last version windows 10

Minimal Reproducible Example

Zip Formatted Incorrectly

Additional

No response

The hub does not work for Russian users.

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

Signin, Datasets

Bug

It is impossible to download dataset and also to log in or register to the account from the Russian IP address.

Environment

• Windows 11
• Chrome 100.0.4896.127

Minimal Reproducible Example

1. Login to hub #use russian ip

Additional

No response

Linear Regression Head

Search before asking

• I have searched the HUB issues and found no similar feature requests.

Description

Easy to make linear regression head and object detection head. A way to find an object in a picture and give you a linear regression model about the characteristics of that model.

Use case

If you want to find a rose and how many days until it is fully bloomed would be a great example.

Additional

No response

Dataset Processing failed

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

Datasets

Bug

I am working on a new dataset and tried to upload it to the HUB. The dataset seemed to upload just fine and quickly reached the processing stage. Shortly after beginning the processing stage the progress bar turned red and failed with the text "processing error" appearing below. I'm trying to determine if there is something wrong with my dataset(very possible as I'm new to this) or if the processing error is on the HUB's end. Any guidance would be greatly appreciated! I have attached the screenshot as well as my zipped dataset file below.

07.zip
.

Environment

-Computer: Mac Mini Late 2014. Processor: 2.6 GHz Dual-Core Intel Core i5. Memory: 8GB 1600 MHz DDR3. MacOS: Big Sur. Graphics: Intel Iris 1536 MB

Minimal Reproducible Example

Uploaded new dataset, let it reach 100%, Processing begins and shortly fails.

Additional

Keep up the AWESOME work and tools! This technology is helping so many people with various tasks and makes machine learning to the average Joe.

EdgeTpu Export model inference

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

How can we use Exported edgetpu model for inference. In windows 10 with google coral usb accelerator.

Additional

No response

Unable to upload dataset

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

Datasets

Bug

After selecting a custom dataset "custom.zip" or the sample one "coco6.zip" included in the tutorial create a dataset, a dialog screen appears with this content: Datasets must be uploaded as a zip or yaml file!

Environment

Ultralytics HUB Version
v0.1.13-beta.2
Client User Agent
Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/100.0.4896.75 Safari/537.36
Operating System
Win32
Browser Window Size
1366 x 657
Server Timestamp
1649275461

Minimal Reproducible Example

Steps to reproduce the problem:

1. Sign in to HUB
2. Click on the Datasets tab.
3. Click 'Upload Dataset'
4. Select the file coco6.zip

Additional

No response

Problems using custom model with Pytorch

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

Hello,
I trained a model via Ultralytics Hub and a custom dataset that I uploaded. I am now looking to utilize the model via Pytorch or Tensorflow to predict images. I exported the model via Pytorch and Tensorflow saved model but I can't make it work. I tried the functions torch.load("path") but I keep obtaining the same error which is : ModuleNotFoundError: No module named 'models'.

Can you help me ?

Thanks for reading me,
Adam.

Additional

No response

How to raise an Ultralytics HUB Issue

Ultralytics HUB Issues

Hello from the Ultralytics Team! We've been working hard to launch Ultralytics HUB, a new web tool for training and deploying YOLOv5 🚀 models.

If you are a new Ultralytics HUB user and have questions or comments for Ultralytics, you are in the right place! Please click the New Issue button in the Issues tab and let us know what we can do to improve your experience on Ultralytics HUB 😃!

Pixel 3i

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

Datasets

Bug

Does not load downloaded data sets and does not start recognition

Environment

No response

Minimal Reproducible Example

No response

Additional

No response

Dataset not Processing

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

i want to train my dataset why is my dataset is processing take time too much

Additional

No response

Error Uploading a new dataset on ultralytics

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

Hello,
I am having problems uploading a dataset on the hub. I followed the steps on this link https://github.com/ultralytics/hub#1-create-a-dataset to create my dataset with roboflow but I tried to upload several times and this error appears:

this is the last dataset that I tried to upload:
projeto.zip

Additional

No response

Can I upload existing models to Ultralytics Hub?

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

Hey,

I have already done some training with yolov5s to use it with the tensorflowlite app. I have some models already trained, with their weights (best.pt and respective *.tflite). Is there a way to upload the models I have already trained? They were all trained with the steps of Yolov5 from Git Hub.

Thanks

Additional

No response

Mobile app sign in freezes

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

Signin

Bug

Tried to create a user thought the app ([email protected]). No error messages just did not progress. Managed to use the Google button to sign in with my g-account.

Environment

Android on Samsung s21+

Minimal Reproducible Example

Install the app
Select create account

Additional

No response

How to raise an Ultralytics HUB Issue

Search before asking

• I have searched the HUB issues and discussions and found no similar questions.

Question

Hello from the Ultralytics Team! We've been working hard to launch Ultralytics HUB, a new web tool for training and deploying YOLOv5 🚀 models.

If you are a new Ultralytics HUB user and have questions or comments for Ultralytics, you are in the right place! Please click the New Issue button in the Issues tab in this ultralytics/hub repo and let us know what we can do to improve your experience on Ultralytics HUB 😃!

Additional

No response

Minor GUI comments on HUB

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

No response

Bug

Minor comments on HUB GUI (not really a bug)

1. 'Sign Up' button should possibly say 'Sign In' i.e. if you've already got an account
   

2. In the models page- the 'Device' and 'Cache' items don't seem to be correct- I'm certain I used GPU and RAM cache (with a local agent). Without an overload of info, it might be useful to display which method was used (COLAB. local agent etc & how long training took) I haven't used WandB with the agent- but perhaps a link to this ?

Andrew

Environment

No response

Minimal Reproducible Example

No response

Additional

No response

Automatically cache dataset

Search before asking

• I have searched the HUB issues and found no similar feature requests.

Description

I have found myself training a model in Google Colab without any caching strategy enabled. The model trains quite slow, and I have also found myself with about 20GB of free disk space and 9GB of free RAM. From my point of view, making an automatic choice for caching when available would be quite useful instead of keeping the default off

Use case

This feature would be used by people that want to train their models faster

Additional

No response

Can I upload custom weights for training?

Is there a way to add a pretrained weight file
So you don't need to start from scratch?

Unable to Edit Model Training Parameters

Search before asking

• I have searched the HUB issues and found no similar bug report.

HUB Component

Models, Training

Bug

I'm not able to edit training parameters here. When I click the edit icon the fields become editable, but then when I click the cursor in the edit field the field disappears and is replaced with the text value.

Environment

Ultralytics HUB Version
v0.1.10-beta.2
Client User Agent
Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/97.0.4692.71 Safari/537.36
Operating System
MacIntel
Browser Window Size
1311 x 705
Server Timestamp
1641945930

Minimal Reproducible Example

No response

Additional

First reported in #9