| from __gin__ import dynamic_registration | |
| import tasks | |
| 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/xl.gin" | |
| include "t5x/configs/runs/finetune.gin" | |
| MIXTURE_OR_TASK_NAME = %gin.REQUIRED | |
| TASK_FEATURE_LENGTHS = {"inputs": 512, "targets": 2} | |
| INITIAL_CHECKPOINT_PATH = %gin.REQUIRED #"gs://t5-data/pretrained_models/t5x/mt5_base/checkpoint_1000000" | |
| TRAIN_STEPS = %gin.REQUIRED #1_010_000 # 1000000 pre-trained steps + 10000 fine-tuning steps. | |
| USE_CACHED_TASKS = False | |
| DROPOUT_RATE = 0.1 | |
| RANDOM_SEED = 0 | |
| #Fixing a small error | |
| infer_eval/utils.DatasetConfig: | |
| task_feature_lengths = %TASK_FEATURE_LENGTHS | |
| #Saving every 1000 steps | |
| utils.SaveCheckpointConfig: | |
| period = 1000 | |
| # Pere: Only necessary if we load a t5 model. We can start with an t5x model here | |
| # `LOSS_NORMALIZING_FACTOR`: When fine-tuning a model that was pre-trained | |
| # using Mesh Tensorflow (e.g. the public T5 / mT5 / ByT5 models), this should be | |
| # set to `pretraining batch_size` * `target_token_length`. For T5 and T5.1.1: | |
| # `2048 * 114`. For mT5: `1024 * 229`. For ByT5: `1024 * 189`. | |
| # LOSS_NORMALIZING_FACTOR = 234496 | |
| # Might have to ba changed based on architecture | |
| # partitioning.PjitPartitioner.num_partitions = 1 | |