maktzgls/bert-jepa-infonce-bilingual-15lang-opus100-15lang-infonce-bilingual
Checkpoint: checkpoint_epoch_10.pt (epoch 10, step 21870) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: best_model.pt (epoch 9, step 19683) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: checkpoint_epoch_8.pt (epoch 8, step 17496) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: best_model.pt (epoch 7, step 15309) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: checkpoint_epoch_6.pt (epoch 6, step 13122) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: best_model.pt (epoch 5, step 10935) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: checkpoint_epoch_4.pt (epoch 4, step 8748) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: best_model.pt (epoch 3, step 6561) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: checkpoint_epoch_2.pt (epoch 2, step 4374) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: best_model.pt (epoch 1, step 2187) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: checkpoint_epoch_10.pt (epoch 10, step 20310) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: checkpoint_epoch_8.pt (epoch 8, step 16248) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: best_model.pt (epoch 7, step 14217) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: checkpoint_epoch_6.pt (epoch 6, step 12186) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: best_model.pt (epoch 5, step 10155) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: checkpoint_epoch_4.pt (epoch 4, step 8124) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: best_model.pt (epoch 3, step 6093) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: checkpoint_epoch_2.pt (epoch 2, step 4062) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: best_model.pt (epoch 1, step 2031) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: best_model.pt (epoch 1, step 2031) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
Checkpoint: best_model.pt (epoch 1, step 2031) | Experiment: opus100-15lang-infonce-bilingual | Training with 15 diverse language pairs using InfoNCE contrastive loss and Bilingual MLM strategy
initial commit