Make a another package

bigscience-workshop · Nov 8, 2021 · e17d454 · e17d454
1 parent 964340d
commit e17d454
Show file tree

Hide file tree

Showing 3 changed files with 3 additions and 4 deletions.
diff --git a/t5x/configs/dataset/pile/task.py → t5x/configs/dataset/pile/pile/task.py b/t5x/configs/dataset/pile/task.py → t5x/configs/dataset/pile/pile/task.py
diff --git a/t5x/configs/dataset/pile/run_cache_tasks_main.sh b/t5x/configs/dataset/pile/run_cache_tasks_main.sh
@@ -1,8 +1,7 @@
 # Need to install seqio
 # gcloud auth application-default login
 
-
-MODULE_IMPORT=t5x.configs.dataset.pile.task
+MODULE_IMPORT=pile.task
 TASK_NAME=pile_t2t_span_corruption
 JOB_NAME=pilet2tspancorruption # the name must consist of only the characters [-a-z0-9], starting with a letter and ending with a letter or number
 BUCKET=gs://bigscience/pile/$TASK_NAME # Don't know is cache needs to be task specific or not ...
@@ -13,4 +12,4 @@ seqio_cache_tasks \
  --module_import=$MODULE_IMPORT \
  --tasks=${TASK_NAME} \
  --output_cache_dir=${BUCKET}/cache \
- --pipeline_options="--runner=DataflowRunner,--project=$PROJECT,--region=$REGION,--job_name=$JOB_NAME,--staging_location=$BUCKET/binaries,--temp_location=$BUCKET/tmp,--setup_file=$PWD/t5x/configs/dataset/pile/setup.py"
+ --pipeline_options="--runner=DataflowRunner,--project=$PROJECT,--region=$REGION,--job_name=$JOB_NAME,--staging_location=$BUCKET/binaries,--temp_location=$BUCKET/tmp,--setup_file=$PWD/setup.py"
diff --git a/t5x/configs/dataset/pile/setup.py b/t5x/configs/dataset/pile/setup.py
@@ -112,7 +112,7 @@ def run(self):
 ]
 
 setuptools.setup(
-    name='cache_pile',
+    name='pile',
     version='0.0.1',
     description='Cache pile set workflow package.',
     install_requires=REQUIRED_PACKAGES,