Views
No views yet
⚠️ DEPRECATED — kept for legacy compatibility. This model was trained withmax_seq_length=256on a constrained-VRAM laptop GPU as a preliminary step. For new work, usejongador/bert-imdb-512, which covers ~95–98% of IMDB reviews (vs. ~85–90%) and achieves higher accuracy (94.14% vs. 92.20%).
bert-base-uncased fine-tuned on the IMDB sentiment classification dataset with max_seq_length=256. Trained as a victim model for adversarial NLP research (TextBugger / TextFooler / DeepWordBug-style attacks).bert-base-uncased (12 layers, 768 hidden, 12 heads, ~110M parameters)bert-base-uncased on the IMDB train split (25,000 examples) using TextAttack 0.3.x.| Hyperparameter | Value |
|---|---|
| Epochs | 5 |
| Per-device batch size | 2 |
| Gradient accumulation | 8 (effective batch 16) |
| Learning rate | 2e-5 |
| Weight decay | 0.01 |
| Warmup steps | 500 |
| Random seed | 786 |
| Hardware | NVIDIA RTX 3050 Laptop (4 GB VRAM) |
jongador/bert-imdb-512 variant uses batch 8 × accum 2 on a cluster RTX 3090 (24 GB).| Metric | Value |
|---|---|
| Accuracy | 92.20% |
1from transformers import AutoTokenizer, AutoModelForSequenceClassification
2
3tokenizer = AutoTokenizer.from_pretrained("jongador/bert-imdb-256")
4model = AutoModelForSequenceClassification.from_pretrained("jongador/bert-imdb-256")