File size: 15,433 Bytes
6cf8c96
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
import mesh_tensorflow.optimize
import mesh_tensorflow.transformer.dataset
import mesh_tensorflow.transformer.learning_rate_schedules
import mesh_tensorflow.transformer.t2t_vocabulary
import mesh_tensorflow.transformer.transformer
import mesh_tensorflow.transformer.transformer_layers
import mesh_tensorflow.transformer.utils
import t5.models.mesh_transformer

# Macros:
# ==============================================================================
d_ff = 3072
d_kv = 64
d_model = 768
dropout_rate = 0.0
inputs_length = 512
mean_noise_span_length = 3.0
MIXTURE_NAME = 'c4_v220_unsupervised'
noise_density = 0.15
num_heads = 12
num_layers = 12

# Parameters for adafactor_decay_rate_pow:
# ==============================================================================
adafactor_decay_rate_pow.offset = 0

# Parameters for AdafactorOptimizer:
# ==============================================================================
AdafactorOptimizer.beta1 = 0.0
AdafactorOptimizer.clipping_threshold = 1.0
AdafactorOptimizer.decay_rate = None
AdafactorOptimizer.epsilon1 = 1e-30
AdafactorOptimizer.epsilon2 = 0.001
AdafactorOptimizer.factored = True
AdafactorOptimizer.min_dim_size_to_factor = 128
AdafactorOptimizer.multiply_by_parameter_scale = True

# Parameters for Bitransformer:
# ==============================================================================
Bitransformer.shared_embedding = True

# Parameters for denoise:
# ==============================================================================
denoise.inputs_fn = @preprocessors.noise_span_to_unique_sentinel
denoise.noise_density = %noise_density
denoise.noise_mask_fn = @preprocessors.random_spans_noise_mask
denoise.targets_fn = @preprocessors.nonnoise_span_to_unique_sentinel

# Parameters for decoder/DenseReluDense:
# ==============================================================================
decoder/DenseReluDense.activation = 'relu'
decoder/DenseReluDense.dropout_rate = %dropout_rate
decoder/DenseReluDense.hidden_size = %d_ff
decoder/DenseReluDense.use_bias = False

# Parameters for encoder/DenseReluDense:
# ==============================================================================
encoder/DenseReluDense.activation = 'relu'
encoder/DenseReluDense.dropout_rate = %dropout_rate
encoder/DenseReluDense.hidden_size = %d_ff
encoder/DenseReluDense.use_bias = False

# Parameters for enc_dec_attention:
# ==============================================================================
# None.

# Parameters for enc_dec_attention_bias:
# ==============================================================================
# None.

# Parameters for decoder/EncDecAttention:
# ==============================================================================
decoder/EncDecAttention.relative_attention_type = None

# Parameters for get_variable_dtype:
# ==============================================================================
get_variable_dtype.activation_dtype = 'bfloat16'

# Parameters for get_vocab_embedding_cls:
# ==============================================================================
# None.

# Parameters for get_vocabulary:
# ==============================================================================
get_vocabulary.mixture_or_task_name = %MIXTURE_NAME

# Parameters for decoder/LayerStack:
# ==============================================================================
decoder/LayerStack.dropout_rate = None
decoder/LayerStack.norm_epsilon = None
decoder/LayerStack.recompute_grads = False
decoder/LayerStack.sublayers_final = \
[@transformer.sublayer_rms_norm, @transformer.sublayer_dropout]
decoder/LayerStack.sublayers_initial = [@transformer.sublayer_dropout]
decoder/LayerStack.sublayers_per_layer = \
[@transformer.sublayer_rms_norm,
@transformer.sublayer_call_layer,
@transformer.sublayer_dropout,
@transformer.sublayer_residual]

# Parameters for encoder/LayerStack:
# ==============================================================================
encoder/LayerStack.dropout_rate = None
encoder/LayerStack.norm_epsilon = None
encoder/LayerStack.recompute_grads = False
encoder/LayerStack.sublayers_final = \
[@transformer.sublayer_rms_norm, @transformer.sublayer_dropout]
encoder/LayerStack.sublayers_initial = [@transformer.sublayer_dropout]
encoder/LayerStack.sublayers_per_layer = \
[@transformer.sublayer_rms_norm,
@transformer.sublayer_call_layer,
@transformer.sublayer_dropout,
@transformer.sublayer_residual]

# Parameters for learning_rate_schedule_noam:
# ==============================================================================
learning_rate_schedule_noam.linear_decay_fraction = 0.0
learning_rate_schedule_noam.multiplier = 1.0
learning_rate_schedule_noam.offset = 0
learning_rate_schedule_noam.warmup_steps = 10000

# Parameters for make_bitransformer:
# ==============================================================================
make_bitransformer.decoder_name = 'decoder'
make_bitransformer.encoder_name = 'decoder'

# Parameters for decoder/make_layer_stack:
# ==============================================================================
decoder/make_layer_stack.block_scope = True
decoder/make_layer_stack.layers = \
[('self_attention',
@mesh_tensorflow.transformer.transformer_layers.SelfAttention),
('enc_dec_attention',
@mesh_tensorflow.transformer.transformer_layers.EncDecAttention),
('dense_relu_dense',
@mesh_tensorflow.transformer.transformer_layers.DenseReluDense)]
decoder/make_layer_stack.num_layers = %num_layers

# Parameters for encoder/make_layer_stack:
# ==============================================================================
encoder/make_layer_stack.block_scope = True
encoder/make_layer_stack.layers = \
[('self_attention',
@mesh_tensorflow.transformer.transformer_layers.SelfAttention),
('dense_relu_dense',
@mesh_tensorflow.transformer.transformer_layers.DenseReluDense)]
encoder/make_layer_stack.num_layers = %num_layers

# Parameters for mesh_train_dataset_fn:
# ==============================================================================
mesh_train_dataset_fn.mixture_or_task_name = %MIXTURE_NAME
mesh_train_dataset_fn.pack = True
mesh_train_dataset_fn.seed = None
mesh_train_dataset_fn.shuffle = True
mesh_train_dataset_fn.use_cached = 1

# Parameters for noise_span_to_unique_sentinel:
# ==============================================================================
# None.

# Parameters for nonnoise_span_to_unique_sentinel:
# ==============================================================================
# None.

# Parameters for pack_dataset:
# ==============================================================================
pack_dataset.use_custom_ops = True

# Parameters for pack_or_pad:
# ==============================================================================
# None.

# Parameters for random_spans_helper:
# ==============================================================================
random_spans_helper.extra_tokens_per_span_inputs = 1
random_spans_helper.extra_tokens_per_span_targets = 1
random_spans_helper.inputs_length = %inputs_length
random_spans_helper.mean_noise_span_length = %mean_noise_span_length
random_spans_helper.noise_density = %noise_density
random_spans_helper.verbose = False

# Parameters for random_spans_noise_mask:
# ==============================================================================
random_spans_noise_mask.mean_noise_span_length = %mean_noise_span_length

# Parameters for random_spans_tokens_length:
# ==============================================================================
# None.

# Parameters for reduce_concat_tokens:
# ==============================================================================
reduce_concat_tokens.batch_size = 128
reduce_concat_tokens.feature_key = 'targets'

# Parameters for rewrite_stack_variables:
# ==============================================================================
rewrite_stack_variables.max_combined_variable_size = 536870912

# Parameters for run:
# ==============================================================================
run.autostack = True
run.batch_size = ('tokens_per_batch', 65536)
run.checkpoint_input_pipeline = False
run.dataset_split = 'train'
run.ensemble_inputs = None
run.eval_checkpoint_step = None
run.eval_dataset_fn = None
run.eval_summary_dir = None
run.export_checkpoint_step = None
run.export_path = ''
run.init_checkpoint = None
run.iterations_per_loop = 100
run.keep_checkpoint_max = None
run.layout_rules = \
'ensemble:ensemble,batch:batch,d_ff:model,heads:model,vocab:model,experts:batch'
run.learning_rate_schedule = @learning_rate_schedules.learning_rate_schedule_noam
run.mesh_devices = None
run.mesh_shape = @mesh_tensorflow.transformer.utils.tpu_mesh_shape()
run.mode = 'train'
run.model_type = 'bitransformer'
run.optimizer = @optimize.AdafactorOptimizer
run.output_eval_examples = True
run.perplexity_eval_steps = 100
run.predict_fn = None
run.save_checkpoints_steps = 5000
run.seen_data_init_step = 0
run.sequence_length = {'inputs': 512, 'targets': 128}
run.skip_seen_data = False
run.total_run_steps = None
run.train_dataset_fn = @t5.models.mesh_transformer.mesh_train_dataset_fn
run.train_steps = 524288
run.variable_filter = None

# Parameters for select_random_chunk:
# ==============================================================================
select_random_chunk.additional_feature_keys = None
select_random_chunk.additional_passthrough_keys = None
select_random_chunk.feature_key = 'targets'
select_random_chunk.max_length = 65536
select_random_chunk.uniform_random_start = False

# Parameters for decoder/SelfAttention:
# ==============================================================================
decoder/SelfAttention.attention_func = None
decoder/SelfAttention.attention_kwargs = None
decoder/SelfAttention.combine_dims = True
decoder/SelfAttention.dropout_rate = %dropout_rate
decoder/SelfAttention.fold_scaling_into_initializer = True
decoder/SelfAttention.keep_query_heads_dims = False
decoder/SelfAttention.key_value_size = %d_kv
decoder/SelfAttention.num_heads = %num_heads
decoder/SelfAttention.num_memory_heads = 0
decoder/SelfAttention.relative_attention_num_buckets = 32
decoder/SelfAttention.relative_attention_type = 'bias_shared'
decoder/SelfAttention.shared_kv = False

# Parameters for encoder/SelfAttention:
# ==============================================================================
encoder/SelfAttention.attention_func = None
encoder/SelfAttention.attention_kwargs = None
encoder/SelfAttention.combine_dims = True
encoder/SelfAttention.dropout_rate = %dropout_rate
encoder/SelfAttention.fold_scaling_into_initializer = True
encoder/SelfAttention.keep_query_heads_dims = False
encoder/SelfAttention.key_value_size = %d_kv
encoder/SelfAttention.num_heads = %num_heads
encoder/SelfAttention.num_memory_heads = 0
encoder/SelfAttention.relative_attention_num_buckets = 32
encoder/SelfAttention.relative_attention_type = 'bias_shared'
encoder/SelfAttention.shared_kv = False

# Parameters for serialize_num_microbatches:
# ==============================================================================
serialize_num_microbatches.tokens_per_microbatch_per_replica = 8192

# Parameters for SimdMeshImpl:
# ==============================================================================
SimdMeshImpl.allreduce_in_bfloat16_max_group_size = 8

# Parameters for split_tokens:
# ==============================================================================
split_tokens.additional_feature_keys = None
split_tokens.feature_key = 'targets'
split_tokens.max_tokens_per_segment = @preprocessors.random_spans_tokens_length()
split_tokens.min_tokens_per_segment = None
split_tokens.passthrough_feature_keys = None

# Parameters for sublayer_call_layer:
# ==============================================================================
# None.

# Parameters for sublayer_dropout:
# ==============================================================================
sublayer_dropout.dropout_rate = %dropout_rate

# Parameters for sublayer_mask_padding:
# ==============================================================================
# None.

# Parameters for sublayer_residual:
# ==============================================================================
# None.

# Parameters for sublayer_rms_norm:
# ==============================================================================
sublayer_rms_norm.epsilon = 1e-06
sublayer_rms_norm.name = 'rms_norm'

# Parameters for tpu_estimator_model_fn:
# ==============================================================================
tpu_estimator_model_fn.hierarchical_tiling_spec = None
tpu_estimator_model_fn.init_variable_filter = ''
tpu_estimator_model_fn.model_info_file = ''
tpu_estimator_model_fn.outer_batch_size = 1
tpu_estimator_model_fn.tpu_summaries = False

# Parameters for tpu_mesh_shape:
# ==============================================================================
tpu_mesh_shape.ensemble_parallelism = None
tpu_mesh_shape.model_parallelism = 1
tpu_mesh_shape.tpu_topology = '4x4'

# Parameters for unit_scaling_convention:
# ==============================================================================
unit_scaling_convention.value = False

# Parameters for decoder/Unitransformer:
# ==============================================================================
decoder/Unitransformer.d_model = %d_model
decoder/Unitransformer.ensemble = None
decoder/Unitransformer.input_full_attention = False
decoder/Unitransformer.label_smoothing = 0.0
decoder/Unitransformer.loss_denominator = None
decoder/Unitransformer.loss_fn = None
decoder/Unitransformer.loss_on_targets_only = False
decoder/Unitransformer.max_length = 512
decoder/Unitransformer.positional_embedding = False
decoder/Unitransformer.shared_embedding_and_softmax_weights = True
decoder/Unitransformer.sinusoid_positional_embedding = False
decoder/Unitransformer.token_dropout_rate = 0.0
decoder/Unitransformer.vocab_divisor = 128
decoder/Unitransformer.z_loss = 0.0001

# Parameters for encoder/Unitransformer:
# ==============================================================================
encoder/Unitransformer.d_model = %d_model
encoder/Unitransformer.ensemble = None
encoder/Unitransformer.input_full_attention = False
encoder/Unitransformer.label_smoothing = 0.0
encoder/Unitransformer.loss_denominator = None
encoder/Unitransformer.loss_fn = None
encoder/Unitransformer.loss_on_targets_only = False
encoder/Unitransformer.max_length = 512
encoder/Unitransformer.positional_embedding = False
encoder/Unitransformer.shared_embedding_and_softmax_weights = True
encoder/Unitransformer.sinusoid_positional_embedding = False
encoder/Unitransformer.token_dropout_rate = 0.0
encoder/Unitransformer.vocab_divisor = 128
encoder/Unitransformer.z_loss = 0.0001

# Parameters for unsupervised:
# ==============================================================================
unsupervised.preprocessors = \
[@preprocessors.select_random_chunk,
@preprocessors.reduce_concat_tokens,
@preprocessors.split_tokens,
@preprocessors.denoise]

# Parameters for VarianceScalingInitializer:
# ==============================================================================
VarianceScalingInitializer.distribution = 'normal'
VarianceScalingInitializer.mode = 'fan_in'
VarianceScalingInitializer.scale = 1.0

# Parameters for VocabEmbedding:
# ==============================================================================
VocabEmbedding.scale_variable_like_classifier_weights = False