From 0bc9a870c0cd198819c49783c2bb4bdee81ed70a Mon Sep 17 00:00:00 2001 From: flybird11111 <1829166702@qq.com> Date: Fri, 23 Aug 2024 13:47:13 +0800 Subject: [PATCH 1/3] Update train_dpo.py --- .../ColossalChat/examples/training_scripts/train_dpo.py | 6 ------ 1 file changed, 6 deletions(-) diff --git a/applications/ColossalChat/examples/training_scripts/train_dpo.py b/applications/ColossalChat/examples/training_scripts/train_dpo.py index b8de6396f..3b324ee78 100755 --- a/applications/ColossalChat/examples/training_scripts/train_dpo.py +++ b/applications/ColossalChat/examples/training_scripts/train_dpo.py @@ -279,10 +279,7 @@ def train(args): beta=args.beta, gamma=args.gamma, length_normalization=args.length_normalization, -<<<<<<< HEAD -======= apply_loss_mask=not args.disable_loss_mask, ->>>>>>> main ) trainer.fit( @@ -351,10 +348,7 @@ if __name__ == "__main__": default=False, help="Disable the reference model (enabled by default)", ) -<<<<<<< HEAD -======= parser.add_argument("--disable_loss_mask", default=False, action="store_true") ->>>>>>> main parser.add_argument("--mixed_precision", type=str, default="fp16", choices=["fp16", "bf16"], help="Mixed precision") parser.add_argument("--lora_config", type=str, default=None, help="low-rank adaptation config file path") parser.add_argument("--save_interval", type=int, default=1000, help="number of step between two checkpoints") From 3b0df30362668c6e3abaa4edbf5ddb621155808d Mon Sep 17 00:00:00 2001 From: "pre-commit-ci[bot]" <66853113+pre-commit-ci[bot]@users.noreply.github.com> Date: Fri, 23 Aug 2024 05:48:11 +0000 Subject: [PATCH 2/3] [pre-commit.ci] auto fixes from pre-commit.com hooks for more information, see https://pre-commit.ci --- colossalai/booster/plugin/low_level_zero_plugin.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/colossalai/booster/plugin/low_level_zero_plugin.py b/colossalai/booster/plugin/low_level_zero_plugin.py index 9a7534dde..42bb49bc9 100644 --- a/colossalai/booster/plugin/low_level_zero_plugin.py +++ b/colossalai/booster/plugin/low_level_zero_plugin.py @@ -349,7 +349,7 @@ class LowLevelZeroPlugin(DPPluginBase): verbose: bool = False, cast_inputs: bool = True, fp8_communication: bool = False, - use_fp8: bool = False + use_fp8: bool = False, ) -> None: super().__init__() assert stage in (1, 2), f"LowLevelZeroPlugin only supports stage 1/2 training" From 9e767643dd28428f6f1f7e95be2f0a66cf4b2558 Mon Sep 17 00:00:00 2001 From: flybird11111 <1829166702@qq.com> Date: Fri, 23 Aug 2024 13:49:53 +0800 Subject: [PATCH 3/3] Update low_level_zero_plugin.py --- colossalai/booster/plugin/low_level_zero_plugin.py | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/colossalai/booster/plugin/low_level_zero_plugin.py b/colossalai/booster/plugin/low_level_zero_plugin.py index 42bb49bc9..9a7534dde 100644 --- a/colossalai/booster/plugin/low_level_zero_plugin.py +++ b/colossalai/booster/plugin/low_level_zero_plugin.py @@ -349,7 +349,7 @@ class LowLevelZeroPlugin(DPPluginBase): verbose: bool = False, cast_inputs: bool = True, fp8_communication: bool = False, - use_fp8: bool = False, + use_fp8: bool = False ) -> None: super().__init__() assert stage in (1, 2), f"LowLevelZeroPlugin only supports stage 1/2 training"