diff options
| -rw-r--r-- | training/conf/experiment/conv_transformer_paragraphs.yaml | 10 |
1 files changed, 5 insertions, 5 deletions
diff --git a/training/conf/experiment/conv_transformer_paragraphs.yaml b/training/conf/experiment/conv_transformer_paragraphs.yaml index af27a9e..d3311fd 100644 --- a/training/conf/experiment/conv_transformer_paragraphs.yaml +++ b/training/conf/experiment/conv_transformer_paragraphs.yaml @@ -40,7 +40,7 @@ lr_scheduler: max_lr: 3.0e-4 total_steps: null epochs: *epochs - steps_per_epoch: 5037 + steps_per_epoch: 3358 pct_start: 0.15 anneal_strategy: cos cycle_momentum: true @@ -55,7 +55,7 @@ lr_scheduler: monitor: val/cer datamodule: - batch_size: 4 + batch_size: 6 train_fraction: 0.95 network: @@ -63,9 +63,9 @@ network: num_classes: *num_classes pad_index: *ignore_index encoder: - depth: 5 + depth: 6 decoder: - depth: 4 + depth: 5 pixel_embedding: shape: [18, 80] @@ -75,4 +75,4 @@ model: trainer: gradient_clip_val: 0.5 max_epochs: *epochs - accumulate_grad_batches: 2 + accumulate_grad_batches: 1 |