final fixes

huggingface · Jan 17, 2024 · cc3b430 · cc3b430
1 parent f7cb79b
commit cc3b430
Show file tree

Hide file tree

Showing 2 changed files with 4 additions and 3 deletions.
diff --git a/Makefile b/Makefile
@@ -1,4 +1,4 @@
-.PHONY: test precommit benchmark_core benchmark_aux common_tests slow_tests test_examples
+.PHONY: test precommit benchmark_core benchmark_aux common_tests slow_tests test_examples tests_gpu
 
 check_dirs := examples tests trl
 
@@ -17,7 +17,7 @@ benchmark_core:
 benchmark_aux:
 	bash ./benchmark/benchmark_aux.sh
 
-tests_common_gpu:
+tests_gpu:
 	python -m pytest tests/test_* $(if $(IS_GITHUB_CI),--report-log "common_tests.log",)
 
 slow_tests:

diff --git a/examples/scripts/sft.py b/examples/scripts/sft.py
@@ -45,7 +45,7 @@ class ScriptArguments:
     batch_size: Optional[int] = field(default=64, metadata={"help": "the batch size"})
     seq_length: Optional[int] = field(default=512, metadata={"help": "Input sequence length"})
     gradient_accumulation_steps: Optional[int] = field(
-        default=16, metadata={"help": "the number of gradient accumulation steps"}
+        default=1, metadata={"help": "the number of gradient accumulation steps"}
     )
     load_in_8bit: Optional[bool] = field(default=False, metadata={"help": "load the model in 8 bits precision"})
     load_in_4bit: Optional[bool] = field(default=False, metadata={"help": "load the model in 4 bits precision"})
@@ -117,6 +117,7 @@ class ScriptArguments:
 training_args = TrainingArguments(
     output_dir=script_args.output_dir,
     per_device_train_batch_size=script_args.batch_size,
+    gradient_accumulation_steps=script_args.gradient_accumulation_steps,
     learning_rate=script_args.learning_rate,
     logging_steps=script_args.logging_steps,
     num_train_epochs=script_args.num_train_epochs,