from __gin__ import dynamic_registration
import tasks
import seqio
import optax

import __main__ as train_script
from t5.data import mixtures
from t5x import models
from t5x import partitioning
from t5x import utils

include 't5x/examples/t5/mt5/base.gin'
include "t5x/configs/runs/finetune.gin"

MIXTURE_OR_TASK_NAME = %gin.REQUIRED
TASK_FEATURE_LENGTHS = {"inputs": 256, "targets": 256}
INITIAL_CHECKPOINT_PATH = %gin.REQUIRED
LR = %gin.REQUIRED
TRAIN_STEPS = %gin.REQUIRED  # 1000000 pre-trained steps + 10000 fine-tuning steps.
USE_CACHED_TASKS = False
DROPOUT_RATE = 0.1
RANDOM_SEED = 0
BATCH_SIZE = 128
EVAL_PERIOD = 1000

#Fixing a small error
infer_eval/utils.DatasetConfig:
  task_feature_lengths = %TASK_FEATURE_LENGTHS

#Saving every 500 steps
utils.SaveCheckpointConfig:
  period = 1000
  keep = 1  # number of checkpoints to keep


#optax.adamw.weight_decay = 0.1
#OPTIMIZER = @optax.adamw
#import t5x.optimizers
#OPTIMIZER = @optax.adamw
#optax.adamw.learning_rate = %LR
#optax.adamw.weight_decay = 0.1


utils.create_learning_rate_scheduler:
  factors = 'constant'
  base_learning_rate = %LR
  warmup_steps = 1000

# Might have to ba changed based on architecture
# partitioning.PjitPartitioner.num_partitions = 1