new server

interference.sh

@@ -2,7 +2,7 @@ INFER_OUTPUT_DIR="output"  # directory to write infer output
 T5X_DIR="../../t5x"  # directory where the t5x is cloned, e.g., ${HOME}"/t5x".
 TFDS_DATA_DIR="gs://nb-t5x/corpus_multi_sentencefix_mt5/"
 CHECKPOINT_PATH="gs://nb-t5x/corpus_multi_sentencefix_mt5/checkpoint_1100000"
-PROJECT_DIR=${HOME}"/mymodel/multi_sentencefix_mt5"
+PROJECT_DIR=${HOME}"/models/multi_sentencefix_mt5"
 export PYTHONPATH=${PROJECT_DIR}
 
 python3 ${T5X_DIR}/t5x/infer.py \

tasks.py

@@ -14,8 +14,8 @@ tsv_path = {
         "test": "gs://nb-t5x/corpus/test/test.tsv"
 }
 
-
-vocabulary = t5.data.ByteVocabulary()
+vocabulary = seqio.SentencePieceVocabulary(
+                'gs://t5-data/vocabs/mc4.250000.100extra/sentencepiece.model', extra_ids=0)
 
 DEFAULT_OUTPUT_FEATURES = {
     "inputs":

train.sh

@@ -1,4 +1,4 @@
-PROJECT_DIR=${HOME}"/mymodel/multi-sentencefix-mt5"
+PROJECT_DIR=${HOME}"/models/multi-sentencefix-mt5"
 T5X_DIR="../../t5x"  # directory where the t5x is cloned.
 TFDS_DATA_DIR="gs://nb-t5x/corpus_multi_sentencefix_mt5"
 MODEL_DIR="gs://nb-t5x/model_multi_sentencefix_mt5"