refactor examples to accommodate Lightning-AI/pytorch-lightning#18105

speediedan · May 6, 2024 · 29d355e · 29d355e
1 parent 41ba761
commit 29d355e
Show file tree

Hide file tree

Showing 3 changed files with 23 additions and 5 deletions.
diff --git a/src/fts_examples/stable/fts_superglue.py b/src/fts_examples/stable/fts_superglue.py
@@ -117,15 +117,23 @@ def __init__(
         super().__init__()
         task_name = task_name if task_name in TASK_NUM_LABELS.keys() else DEFAULT_TASK
         self.text_fields = self.TASK_TEXT_FIELD_MAP[task_name]
+        self.init_hparams = {
+            "model_name_or_path": model_name_or_path,
+            "task_name": task_name,
+            "max_seq_length": max_seq_length,
+            "train_batch_size": train_batch_size,
+            "eval_batch_size": eval_batch_size,
+            "dataloader_kwargs": dataloader_kwargs,
+            "tokenizers_parallelism": tokenizers_parallelism,
+        }
+        self.save_hyperparameters(self.init_hparams)
         self.dataloader_kwargs = {
             "num_workers": dataloader_kwargs.get("num_workers", 0),
             "pin_memory": dataloader_kwargs.get("pin_memory", False),
         }
-        self.save_hyperparameters()
         os.environ["TOKENIZERS_PARALLELISM"] = "true" if self.hparams.tokenizers_parallelism else "false"
-        self.tokenizer = AutoTokenizer.from_pretrained(
-            self.hparams.model_name_or_path, use_fast=True, local_files_only=False
-        )
+        self.tokenizer = AutoTokenizer.from_pretrained(self.hparams.model_name_or_path, use_fast=True,
+                                                       local_files_only=False)
 
     def prepare_data(self):
         """Load the SuperGLUE dataset."""

diff --git a/src/fts_examples/stable/ipynb_src/fts_superglue_nb.py b/src/fts_examples/stable/ipynb_src/fts_superglue_nb.py
@@ -250,11 +250,20 @@ def __init__(
         super().__init__()
         task_name = task_name if task_name in TASK_NUM_LABELS.keys() else DEFAULT_TASK
         self.text_fields = self.TASK_TEXT_FIELD_MAP[task_name]
+        self.init_hparams = {
+            "model_name_or_path": model_name_or_path,
+            "task_name": task_name,
+            "max_seq_length": max_seq_length,
+            "train_batch_size": train_batch_size,
+            "eval_batch_size": eval_batch_size,
+            "dataloader_kwargs": dataloader_kwargs,
+            "tokenizers_parallelism": tokenizers_parallelism,
+        }
+        self.save_hyperparameters(self.init_hparams)
         self.dataloader_kwargs = {
             "num_workers": dataloader_kwargs.get("num_workers", 0),
             "pin_memory": dataloader_kwargs.get("pin_memory", False),
         }
-        self.save_hyperparameters()
         os.environ["TOKENIZERS_PARALLELISM"] = "true" if self.hparams.tokenizers_parallelism else "false"
         self.tokenizer = AutoTokenizer.from_pretrained(
             self.hparams.model_name_or_path, use_fast=True, local_files_only=False

diff --git a/src/fts_examples/stable/test_examples.py b/src/fts_examples/stable/test_examples.py
@@ -36,6 +36,7 @@
     "does not have many workers",
     "is smaller than the logging interval",
     "sentencepiece tokenizer that you are converting",
+    "`resume_download` is deprecated",  # required because of upstream usage as of 2.2.2
     "distutils Version classes are deprecated",  # still required as of PyTorch/Lightning 2.2
     "Please use torch.utils._pytree.register_pytree_node",  # temp allow deprecated behavior of transformers
     "We are importing from `pydantic",  # temp pydantic import migration warning