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/llama-2/gptq-lora.yml
+++ b/examples/llama-2/gptq-lora.yml
@@ -15,7 +15,7 @@ hf_use_auth_token: true
 datasets:
  - path: mhenrichsen/alpaca_2k_test
    type: alpaca
-dataset_prepared_path: last_run_prepared
+dataset_prepared_path:
 val_set_size: 0.01
 adapter: lora
 lora_model_dir:
--- a/examples/llama-2/lora.yml
+++ b/examples/llama-2/lora.yml
@@ -11,7 +11,7 @@ strict: false
 datasets:
  - path: mhenrichsen/alpaca_2k_test
    type: alpaca
-dataset_prepared_path: last_run_prepared
+dataset_prepared_path:
 val_set_size: 0.01
 output_dir: ./lora-out

--- a/examples/llama-2/qlora.yml
+++ b/examples/llama-2/qlora.yml
@@ -11,7 +11,7 @@ strict: false
 datasets:
  - path: mhenrichsen/alpaca_2k_test
    type: alpaca
-dataset_prepared_path: last_run_prepared
+dataset_prepared_path:
 val_set_size: 0.01
 output_dir: ./qlora-out

--- a/examples/llama-2/relora.yml
+++ b/examples/llama-2/relora.yml
@@ -11,7 +11,7 @@ strict: false
 datasets:
  - path: teknium/GPT4-LLM-Cleaned
    type: alpaca
-dataset_prepared_path: last_run_prepared
+dataset_prepared_path:
 val_set_size: 0.01
 output_dir: ./relora-out

--- a/examples/llama-2/tiny-llama.yml
+++ b/examples/llama-2/tiny-llama.yml
@@ -12,7 +12,7 @@ strict: false
 datasets:
  - path: mhenrichsen/alpaca_2k_test
    type: alpaca
-dataset_prepared_path: last_run_prepared
+dataset_prepared_path:
 val_set_size: 0.01
 output_dir: ./lora-out