René's URL Explorer Experiment

{"@context":"https://schema.org","@type":"DiscussionForumPosting","headline":"Error with only captcha digit datasets","articleBody":"Hello,\r\n\r\nI'm testing your captcha to text tutorial and it works fine with alphanum captcha datasets, but fails with only digits captcha datasets (I'm testing a 4-digit captcha dataset):\r\n\r\n`\r\nC:\\Projects\\Captcha Solver\\mltu-main\\Tutorials\\02_captcha_to_text\u003epython inferenceModel.py\r\n  0%|                                                                                           | 0/28 [00:00\u003c?, ?it/s]\r\nTraceback (most recent call last):\r\n  File \"C:\\Projects\\Captcha Solver\\mltu-main\\Tutorials\\02_captcha_to_text\\inferenceModel.py\", line 41, in \u003cmodule\u003e\r\n    cer = get_cer(prediction_text, label)\r\n  File \"C:\\Users\\mudro\\AppData\\Local\\Programs\\Python\\Python310\\lib\\site-packages\\mltu\\utils\\text_utils.py\", line 79, in get_cer\r\n    for pred_tokens, tgt_tokens in zip(preds, target):\r\nTypeError: 'int' object is not iterable\r\n`\r\n\r\nI fixed the code in inferenceModel.py with:\r\n`\r\nfor image_path, label in tqdm(df):\r\n        image = cv2.imread(image_path.replace(\"\\\\\", \"/\"))\r\n\r\n        prediction_text = model.predict(image)\r\n\r\n        cer = get_cer(prediction_text, str(label).zfill(4))\r\n        print(f\"Image: {image_path}, Label: {str(label).zfill(4)}, Prediction: {prediction_text}, CER: {cer}\")\r\n\r\n        accum_cer.append(cer)\r\n`\r\n\r\nThis code runs correctly but gives now bad predition results.\r\n","author":{"url":"https://github.com/Mudr0x","@type":"Person","name":"Mudr0x"},"datePublished":"2024-08-05T15:44:34.000Z","interactionStatistic":{"@type":"InteractionCounter","interactionType":"https://schema.org/CommentAction","userInteractionCount":0},"url":"https://github.com/59/mltu/issues/59"}