[BUG] KeyError: 'target' in Image Classification - autotrain-advanced Colab

[Theivaprakasham]

Prerequisites

• [X] I have read the documentation.
• [X] I have checked other issues for similar problems.

Backend

Colab

Interface Used

UI

CLI Command

No response

UI Screenshots & Parameters

Error Logs

Training is starting... Please wait! INFO | 2024-09-30 11:48:29 | autotrain.cli.autotrain:main:60 - Using AutoTrain configuration: config.yml INFO | 2024-09-30 11:48:29 | autotrain.parser:post_init__:148 - Running task: image_multi_class_classification INFO | 2024-09-30 11:48:29 | autotrain.parser:post_init:149 - Using backend: local WARNING | 2024-09-30 11:48:29 | autotrain.trainers.common:init:180 - Parameters supplied but not used: image, label INFO | 2024-09-30 11:48:29 | autotrain.parser:run:212 - {'data_path': '/content/dummy', 'model': 'microsoft/resnet-50', 'username': None, 'lr': 5e-05, 'epochs': 3, 'batch_size': 8, 'warmup_ratio': 0.1, 'gradient_accumulation': 1, 'optimizer': 'adamw_torch', 'scheduler': 'linear', 'weight_decay': 0.0, 'max_grad_norm': 1.0, 'seed': 42, 'train_split': 'train', 'valid_split': None, 'logging_steps': -1, 'project_name': 'autotrain-m82sg-85s3j', 'auto_find_batch_size': False, 'mixed_precision': 'fp16', 'save_total_limit': 1, 'token': None, 'push_to_hub': True, 'eval_strategy': 'epoch', 'image_column': 'image', 'target_column': 'target', 'log': 'tensorboard', 'early_stopping_patience': 5, 'early_stopping_threshold': 0.01} INFO | 2024-09-30 11:48:29 | autotrain.backends.local:create:8 - Starting local training... INFO | 2024-09-30 11:48:29 | autotrain.commands:launch_command:501 - ['accelerate', 'launch', '--num_machines', '1', '--num_processes', '1', '--mixed_precision', 'fp16', '-m', 'autotrain.trainers.image_classification', '--training_config', 'autotrain-m82sg-85s3j/training_params.json'] INFO | 2024-09-30 11:48:29 | autotrain.commands:launch_command:502 - {'data_path': '/content/dummy', 'model': 'microsoft/resnet-50', 'username': None, 'lr': 5e-05, 'epochs': 3, 'batch_size': 8, 'warmup_ratio': 0.1, 'gradient_accumulation': 1, 'optimizer': 'adamw_torch', 'scheduler': 'linear', 'weight_decay': 0.0, 'max_grad_norm': 1.0, 'seed': 42, 'train_split': 'train', 'valid_split': None, 'logging_steps': -1, 'project_name': 'autotrain-m82sg-85s3j', 'auto_find_batch_size': False, 'mixed_precision': 'fp16', 'save_total_limit': 1, 'token': None, 'push_to_hub': True, 'eval_strategy': 'epoch', 'image_column': 'image', 'target_column': 'target', 'log': 'tensorboard', 'early_stopping_patience': 5, 'early_stopping_threshold': 0.01} The following values were not passed to accelerate launch and had defaults used instead: --dynamo_backend was set to a value of 'no' To avoid this warning pass in values for each of the problematic parameters or run accelerate config. INFO:albumentations.check_version:A new version of Albumentations is available: 1.4.16 (you have 1.4.7). Upgrade using: pip install --upgrade albumentations INFO | 2024-09-30 11:48:41 | main:train:82 - Train data: Dataset({ features: ['image', 'label'], num_rows: 374 }) INFO | 2024-09-30 11:48:41 | main:train:83 - Valid data: None ERROR | 2024-09-30 11:48:41 | autotrain.trainers.common:wrapper:120 - train has failed due to an exception: Traceback (most recent call last): File "/usr/local/lib/python3.10/dist-packages/autotrain/trainers/common.py", line 117, in wrapper return func(*args, **kwargs) File "/usr/local/lib/python3.10/dist-packages/autotrain/trainers/image_classification/main__.py", line 85, in train classes = train_data.features[config.target_column].names KeyError: 'target'

ERROR | 2024-09-30 11:48:41 | autotrain.trainers.common:wrapper:121 - 'target' INFO | 2024-09-30 11:48:42 | autotrain.parser:run:217 - Job ID: 6923 Training completed successfully!

Additional Information

I am trying to train a image classification model using a dataset of 5 classes. However, I'm receiving a KeyError: 'target' error while starting to train. I have followed the dataset structure as mentioned in the Documentation.

Config:

{ "lr": 5e-05, "epochs": 3, "batch_size": 8, "warmup_ratio": 0.1, "gradient_accumulation": 1, "optimizer": "adamw_torch", "scheduler": "linear", "weight_decay": 0.0, "max_grad_norm": 1.0, "seed": 42, "logging_steps": -1, "auto_find_batch_size": false, "mixed_precision": "fp16", "save_total_limit": 1, "eval_strategy": "epoch", "early_stopping_patience": 5, "early_stopping_threshold": 0.01, "push_to_hub": true }

[abhishekkrthakur]

how does the dataset look like?

[Theivaprakasham]

how does the dataset look like?

/content/dummy

├── Aglais_io │ ├── 20240210230338280122_large.jpg │ ├── 20240210230338437501_large.jpg │ ├── 20240210230338679894_large.jpeg │ ├── 20240210230338859195_large.jpeg │ ├── 20240210230339371897_large.jpeg │ ├── 20240210230339481055_large.jpeg ........... ├── Anthocharis_cardamines │ ├── 20240210230356269418_large.jpg │ ├── 20240210230356624279_large.jpg │ ├── 20240210230356792814_large.jpg │ ├── 20240210230356861192_large.jpg │ ├── 20240210230357034893_large.jpeg │ ├── 20240210230357081746_large.jpeg ........ ...... └── Argynnis_pandora ├── 20240210230426055731_large.jpg ├── 20240210230426075113_large.JPG ├── 20240210230426142114_large.jpg ├── 20240210230426332549_large.jpg ├── 20240210230426976274_large.jpg

6 directories, 374 files

[abhishekkrthakur]

could you please try this colab instead: https://colab.research.google.com/github/huggingface/autotrain-advanced/blob/main/colabs/image_classification.ipynb ?

[Theivaprakasham]

could you please try this colab instead: https://colab.research.google.com/github/huggingface/autotrain-advanced/blob/main/colabs/image_classification.ipynb ?

I tried running this colab. I got this error.

ERROR | 2024-09-30 12:50:52 | autotrain.trainers.common:wrapper:120 - train has failed due to an exception: Traceback (most recent call last): File "/usr/local/lib/python3.10/dist-packages/autotrain/trainers/common.py", line 117, in wrapper return func(*args, **kwargs) File "/usr/local/lib/python3.10/dist-packages/autotrain/trainers/image_classification/main.py", line 85, in train classes = train_data.features[config.target_column].names KeyError: 'labels'

ERROR | 2024-09-30 12:50:52 | autotrain.trainers.common:wrapper:121 - 'labels' INFO | 2024-09-30 12:50:54 | autotrain.parser:run:217 - Job ID: 3343

However, I just changed this

column_mapping: # do not change image_column: image target_column: labels

to

_targetcolumn: label

And this Colab works seamlessly now.

[abhishekkrthakur]

ohh. ill fix it and publish task specific colabs. the colab ui is quite un-maintainable. thank you for letting me know.

[Theivaprakasham]

ohh. ill fix it and publish task specific colabs. the colab ui is quite un-maintainable. thank you for letting me know.

Thank you! Cheers.. Have a good day!