summaryrefslogtreecommitdiff
path: root/training/experiments/image_transformer.yaml
diff options
context:
space:
mode:
authorGustaf Rydholm <gustaf.rydholm@gmail.com>2021-04-05 23:14:24 +0200
committerGustaf Rydholm <gustaf.rydholm@gmail.com>2021-04-05 23:14:24 +0200
commit532286b516b17d279c321358bf03dddc8adc8029 (patch)
tree9409789d2a3c093b499253f2fd101a74be1da452 /training/experiments/image_transformer.yaml
parent370c3eb35f47e64eab7926a5f995947c63c6b208 (diff)
Completed first draft for training loop with PyTorch Lightning
Diffstat (limited to 'training/experiments/image_transformer.yaml')
-rw-r--r--training/experiments/image_transformer.yaml70
1 files changed, 70 insertions, 0 deletions
diff --git a/training/experiments/image_transformer.yaml b/training/experiments/image_transformer.yaml
new file mode 100644
index 0000000..7f0bbb7
--- /dev/null
+++ b/training/experiments/image_transformer.yaml
@@ -0,0 +1,70 @@
+network:
+ type: ImageTransformer
+ args:
+ input_shape: None
+ output_shape: None
+ encoder:
+ type: None
+ args: None
+ mapping: sentence_piece
+ num_decoder_layers: 4
+ hidden_dim: 256
+ num_heads: 4
+ expansion_dim: 1024
+ dropout_rate: 0.1
+ transformer_activation: glu
+
+model:
+ type: LitTransformerModel
+ args:
+ optimizer:
+ type: MADGRAD
+ args:
+ lr: 1.0e-2
+ momentum: 0.9
+ weight_decay: 0
+ eps: 1.0e-6
+ lr_scheduler:
+ type: CosineAnnealingLR
+ args:
+ T_max: 512
+ criterion:
+ type: CrossEntropyLoss
+ args:
+ weight: None
+ ignore_index: -100
+ reduction: mean
+
+ monitor: val_loss
+ mapping: sentence_piece
+
+data:
+ type: IAMExtendedParagraphs
+ args:
+ batch_size: 16
+ num_workers: 12
+ train_fraction: 0.8
+ augment: true
+
+callbacks:
+ - type: ModelCheckpoint
+ args:
+ monitor: val_loss
+ mode: min
+ - type: EarlyStopping
+ args:
+ monitor: val_loss
+ mode: min
+ patience: 10
+
+trainer:
+ args:
+ stochastic_weight_avg: true
+ auto_scale_batch_size: power
+ gradient_clip_val: 0
+ fast_dev_run: false
+ gpus: 1
+ precision: 16
+ max_epocs: 512
+ terminate_on_nan: true
+ weights_summary: true