From 4dafbe595a44a7b4490ecb1bb608403db5e6ef05 Mon Sep 17 00:00:00 2001 From: Kashif Rasul Date: Tue, 15 Oct 2024 09:53:38 +0200 Subject: [PATCH 1/2] failing tests --- tests/test_online_dpo_trainer.py | 2 ++ 1 file changed, 2 insertions(+) diff --git a/tests/test_online_dpo_trainer.py b/tests/test_online_dpo_trainer.py index acabbbafa4..6d6bca14ed 100644 --- a/tests/test_online_dpo_trainer.py +++ b/tests/test_online_dpo_trainer.py @@ -70,6 +70,7 @@ def test_training_with_ref_model(self): max_steps=3, learning_rate=5.0e-7, eval_strategy="steps", + eval_steps=1, report_to="none", ) dummy_dataset = load_dataset("trl-internal-testing/zen", "standard_prompt_only") @@ -118,6 +119,7 @@ def test_training_with_peft(self): max_steps=3, learning_rate=5.0e-7, eval_strategy="steps", + eval_steps=1, report_to="none", ) dummy_dataset = load_dataset("trl-internal-testing/zen", "standard_prompt_only") From 659224823ac7a86a378ad5fb4212c7a9fb23a2e3 Mon Sep 17 00:00:00 2001 From: Kashif Rasul Date: Tue, 15 Oct 2024 09:56:15 +0200 Subject: [PATCH 2/2] more failing tests --- tests/test_nash_md_trainer.py | 2 ++ tests/test_xpo_trainer.py | 2 ++ 2 files changed, 4 insertions(+) diff --git a/tests/test_nash_md_trainer.py b/tests/test_nash_md_trainer.py index aff4d9e2cd..43c246c8f7 100644 --- a/tests/test_nash_md_trainer.py +++ b/tests/test_nash_md_trainer.py @@ -76,6 +76,7 @@ def test_training_with_peft(self): max_steps=3, learning_rate=5.0e-7, eval_strategy="steps", + eval_steps=1, report_to="none", ) dummy_dataset = load_dataset("trl-internal-testing/zen", "standard_prompt_only") @@ -105,6 +106,7 @@ def test_training_with_peft_and_ref_model(self): max_steps=3, learning_rate=5.0e-7, eval_strategy="steps", + eval_steps=1, report_to="none", ) dummy_dataset = load_dataset("trl-internal-testing/zen", "standard_prompt_only") diff --git a/tests/test_xpo_trainer.py b/tests/test_xpo_trainer.py index 734d30ba4d..fd770426b5 100644 --- a/tests/test_xpo_trainer.py +++ b/tests/test_xpo_trainer.py @@ -47,6 +47,7 @@ def test_xpo_trainer_training(self, config_name): gradient_accumulation_steps=1, learning_rate=9e-1, eval_strategy="steps", + eval_steps=1, report_to="none", ) dummy_dataset = load_dataset("trl-internal-testing/zen", config_name) @@ -76,6 +77,7 @@ def test_training_with_peft(self): max_steps=3, learning_rate=5.0e-7, eval_strategy="steps", + eval_steps=1, report_to="none", ) dummy_dataset = load_dataset("trl-internal-testing/zen", "standard_prompt_only")