long

batch_nynorsk_balanced_small_long.sh

@@ -0,0 +1,12 @@
+PROJECT_DIR=${HOME}"/models/t5-nynorsk-oversetter"
+export PYTHONPATH=${PROJECT_DIR}
+INITIAL_CHECKPOINT_PATH=\"gs://nb-t5x-us-central2/norwegian_NCC_plus_English_pluss200k_balanced_bokmaal_nynorsk_t5x_small/checkpoint_1700000\"
+TRAIN_STEPS=1705000
+
+python3 ../../t5x/t5x/train.py --gin_search_paths="./" --gin.TRAIN_STEPS=${TRAIN_STEPS} --gin_file="finetune_translate_small.gin" --gin.INITIAL_CHECKPOINT_PATH=${INITIAL_CHECKPOINT_PATH}  --gin.MIXTURE_OR_TASK_NAME=\"translate_long\" --gin.MODEL_DIR=\"gs://nb-t5x-us-central2/finetuned/nynorsk_balanced_small_long_v1\"
+
+#python3 ../../t5x/t5x/train.py --gin_search_paths="./" --gin.TRAIN_STEPS=${TRAIN_STEPS} --gin_file="finetune_translate_small.gin" --gin.INITIAL_CHECKPOINT_PATH=${INITIAL_CHECKPOINT_PATH}  --gin.MIXTURE_OR_TASK_NAME=\"translate_long\" --gin.MODEL_DIR=\"gs://nb-t5x-us-central2/finetuned/nynorsk_balanced_small_long_v2\" &&
+#python3 ../../t5x/t5x/train.py --gin_search_paths="./" --gin.TRAIN_STEPS=${TRAIN_STEPS} --gin_file="finetune_translate_small.gin" --gin.INITIAL_CHECKPOINT_PATH=${INITIAL_CHECKPOINT_PATH}  --gin.MIXTURE_OR_TASK_NAME=\"translate_long\" --gin.MODEL_DIR=\"gs://nb-t5x-us-central2/finetuned/nynorsk_balanced_small_long_v3\" &&
+#python3 ../../t5x/t5x/train.py --gin_search_paths="./" --gin.TRAIN_STEPS=${TRAIN_STEPS} --gin_file="finetune_translate_small.gin" --gin.INITIAL_CHECKPOINT_PATH=${INITIAL_CHECKPOINT_PATH}  --gin.MIXTURE_OR_TASK_NAME=\"translate_long\" --gin.MODEL_DIR=\"gs://nb-t5x-us-central2/finetuned/nynorsk_balanced_small_long_v4\" &&
+#python3 ../../t5x/t5x/train.py --gin_search_paths="./" --gin.TRAIN_STEPS=${TRAIN_STEPS} --gin_file="finetune_translate_small.gin" --gin.INITIAL_CHECKPOINT_PATH=${INITIAL_CHECKPOINT_PATH}  --gin.MIXTURE_OR_TASK_NAME=\"translate_long\" --gin.MODEL_DIR=\"gs://nb-t5x-us-central2/finetuned/nynorsk_balanced_small_long_v5\"
+

tasks.py

@@ -25,6 +25,12 @@ tsv_translate_path = {
         "test": "gs://nb-t5x-us-central2/corpus_bokmal_nynorsk/test.tsv"
 }
 
+tsv_translate_long_path = {
+        "train": "gs://nb-t5x-us-central2/corpus_bokmal_nynorsk/train_long.tsv",
+        "validation": "gs://nb-t5x-us-central2/corpus_bokmal_nynorsk/dev_long.tsv",
+        "test": "gs://nb-t5x-us-central2/corpus_bokmal_nynorsk/test_long.tsv"
+}
+
 tsv_sentiment_path = {
         "train": "gs://notram-public/finetune_datasets/norec_sentiment/train.tsv",
         "validation": "gs://notram-public/finetune_datasets/norec_sentiment/dev.tsv",
@@ -182,3 +188,20 @@ seqio.TaskRegistry.add(
     output_features=DEFAULT_OUTPUT_FEATURES,
 )   
 
+seqio.TaskRegistry.add(
+    "translate_long",
+    source=seqio.TextLineDataSource(
+        split_to_filepattern=tsv_translate_long_path,
+        #num_input_examples=num_nq_examples
+        ),
+    preprocessors=[
+      functools.partial(
+          t5.data.preprocessors.parse_tsv,
+          field_names=["source","target"]),
+      categorise_preprocessor,
+      seqio.preprocessors.tokenize_and_append_eos,
+    ],
+    metric_fns=[metrics.accuracy,my_metrics.f1_macro,metrics.bleu],
+    output_features=DEFAULT_OUTPUT_FEATURES,
+)   
+