Skip to content

Commit

Permalink
Config files samples
Browse files Browse the repository at this point in the history
  • Loading branch information
vince62s committed Nov 22, 2018
1 parent 260ae7c commit 20fb278
Show file tree
Hide file tree
Showing 3 changed files with 118 additions and 0 deletions.
31 changes: 31 additions & 0 deletions config/config-rnn-summarization.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,31 @@
data: data/cnndm/CNNDM
save_model: models/cnndm
save_checkpoint_steps: 10000
keep_checkpoint: 10
seed: 3435
train_steps: 100000
valid_steps: 10000
report_every: 100

encoder_type: brnn
word_vec_size: 128
rnn_size: 512
layers: 1

optim: adagrad
learning_rate: 0.15
adagrad_accumulator_init: 0.1
max_grad_norm: 2

batch_size: 16
dropout: 0.0

copy_attn: 'true'
global_attention: mlp
reuse_copy_attn: 'true'
bridge: 'true'

world_size: 2
gpu_ranks:
- 0
- 1
42 changes: 42 additions & 0 deletions config/config-transformer-base-1GPU.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,42 @@
data: exp/dataset.de-en
save_model: exp/model.de-en
save_checkpoint_steps: 10000
keep_checkpoint: 10
seed: 3435
train_steps: 500000
valid_steps: 10000
warmup_steps: 8000
report_every: 100

decoder_type: transformer
encoder_type: transformer
word_vec_size: 512
rnn_size: 512
layers: 6
transformer_ff: 2048
heads: 8

accum_count: 8
optim: adam
adam_beta1: 0.9
adam_beta2: 0.998
decay_method: noam
learning_rate: 2.0
max_grad_norm: 0.0

batch_size: 4096
batch_type: tokens
normalization: tokens
dropout: 0.1
label_smoothing: 0.1

max_generator_batches: 2

param_init: 0.0
param_init_glorot: 'true'
position_encoding: 'true'

world_size: 1
gpu_ranks:
- 0

45 changes: 45 additions & 0 deletions config/config-transformer-base-4GPU.yml
Original file line number Diff line number Diff line change
@@ -0,0 +1,45 @@
data: exp/dataset.de-en
save_model: exp/model.de-en
save_checkpoint_steps: 10000
keep_checkpoint: 10
seed: 3435
train_steps: 200000
valid_steps: 10000
warmup_steps: 8000
report_every: 100

decoder_type: transformer
encoder_type: transformer
word_vec_size: 512
rnn_size: 512
layers: 6
transformer_ff: 2048
heads: 8

accum_count: 2
optim: adam
adam_beta1: 0.9
adam_beta2: 0.998
decay_method: noam
learning_rate: 2.0
max_grad_norm: 0.0

batch_size: 4096
batch_type: tokens
normalization: tokens
dropout: 0.1
label_smoothing: 0.1

max_generator_batches: 2

param_init: 0.0
param_init_glorot: 'true'
position_encoding: 'true'

world_size: 4
gpu_ranks:
- 0
- 1
- 2
- 3

0 comments on commit 20fb278

Please sign in to comment.