Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

FIX: Failing AdaLoRA GPU test #2349

Merged
merged 1 commit into from
Jan 27, 2025
Merged

FIX: Failing AdaLoRA GPU test #2349

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
30 changes: 19 additions & 11 deletions tests/test_gpu_examples.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -370,9 +370,9 @@ def test_4bit_adalora_causalLM(self):
peft_config = AdaLoraConfig(
init_r=6,
target_r=4,
tinit=50,
tfinal=100,
total_step=200,
tinit=2,
tfinal=2,
total_step=6,
deltaT=5,
beta1=0.3,
beta2=0.3,
Expand Down Expand Up @@ -404,7 +404,7 @@ def on_optimizer_step(self, args, state, control, **kwargs):
per_device_train_batch_size=4,
gradient_accumulation_steps=4,
warmup_steps=2,
max_steps=3,
max_steps=6,
learning_rate=2e-4,
fp16=True,
logging_steps=1,
Expand Down Expand Up @@ -443,9 +443,9 @@ def test_8bit_adalora_causalLM(self):
peft_config = AdaLoraConfig(
init_r=6,
target_r=4,
tinit=50,
tfinal=100,
total_step=200,
tinit=2,
tfinal=2,
total_step=6,
deltaT=5,
beta1=0.3,
beta2=0.3,
Expand Down Expand Up @@ -477,7 +477,7 @@ def on_optimizer_step(self, args, state, control, **kwargs):
per_device_train_batch_size=4,
gradient_accumulation_steps=4,
warmup_steps=2,
max_steps=3,
max_steps=6,
learning_rate=2e-4,
fp16=True,
logging_steps=1,
Expand Down Expand Up @@ -1494,8 +1494,9 @@ def test_adalora_causalLM(self):
peft_config = AdaLoraConfig(
init_r=6,
target_r=4,
tinit=50,
tfinal=100,
tinit=2,
tfinal=2,
total_step=6,
deltaT=5,
beta1=0.3,
beta2=0.3,
Expand All @@ -1513,6 +1514,12 @@ def test_adalora_causalLM(self):
batch = tokenizer(data["train"][:3]["quote"], return_tensors="pt", padding=True)
self._check_inference_finite(model, batch)

class OptimizerStepCallback(TrainerCallback):
def on_optimizer_step(self, args, state, control, **kwargs):
model.update_and_allocate(state.global_step)

step_callback = OptimizerStepCallback()

with tempfile.TemporaryDirectory() as tmp_dir:
trainer = Trainer(
model=model,
Expand All @@ -1521,14 +1528,15 @@ def test_adalora_causalLM(self):
per_device_train_batch_size=4,
gradient_accumulation_steps=4,
warmup_steps=2,
max_steps=3,
max_steps=6,
learning_rate=2e-4,
fp16=True,
logging_steps=1,
output_dir=tmp_dir,
),
data_collator=DataCollatorForLanguageModeling(self.tokenizer, mlm=False),
)
trainer.add_callback(step_callback)
model.config.use_cache = False
trainer.train()

Expand Down
Loading