prepared dataset caching, other misc fixes (#665)

* prepared dataset caching, other misc fixes * also don't load from disk cache unless explicit
2023-10-02 21:07:24 -04:00
parent f4868d733c
commit e50a64e85e
32 changed files with 35 additions and 34 deletions
--- a/examples/xgen-7b/xgen-7b-8k-qlora.yml
+++ b/examples/xgen-7b/xgen-7b-8k-qlora.yml
@@ -16,7 +16,7 @@ datasets:
    data_files:
      - openassistant_best_replies_train.jsonl
    type: "completion"
-dataset_prepared_path: last_run_prepared
+dataset_prepared_path:
 val_set_size: 0.01
 # enable QLoRA
 adapter: qlora