robbert-v2-dutch-base-hebban-reviews5

robbert-v2-dutch-base-hebban-reviews5

Dataset

• dataset_name: BramVanroy/hebban-reviews
• dataset_config: filtered_rating
• dataset_revision: 2.0.0
• labelcolumn: review_rating0
• textcolumn: review_text_without_quotes

Training

• optim: adamw_hf
• learning_rate: 5e-05
• per_device_train_batch_size: 64
• per_device_eval_batch_size: 64
• gradient_accumulation_steps: 1
• max_steps: 5001
• save_steps: 500
• metric_for_best_model: qwk

Best checkedpoint based on validation

• best_metric: 0.7480754124116261
• best_model_checkpoint: trained/hebban-reviews5/robbert-v2-dutch-base/checkpoint-3000

Test results of best checkpoint

• accuracy: 0.624457593688363
• f1: 0.625518585787774
• precision: 0.6295608657909847
• qwk: 0.7517620387343015
• recall: 0.624457593688363

Confusion matrix

Normalized confusion matrix

Environment

• cuda_capabilities: 8.0; 8.0
• cuda_device_count: 2
• cuda_devices: NVIDIA A100-SXM4-80GB; NVIDIA A100-SXM4-80GB
• finetuner_commit: 8159b4c1d5e66b36f68dd263299927ffb8670ebd
• platform: Linux-4.18.0-305.49.1.el8_4.x86_64-x86_64-with-glibc2.28
• python_version: 3.9.5
• toch_version: 1.10.0
• transformers_version: 4.21.0