fix import bug

Former-commit-id: 35b91ea34caade45dd51813b94da5177b852aa4c

src/llmtuner/model/__init__.py

@@ -1,3 +1,5 @@
+# Level: loader > adapter > parser, utils
+
 from llmtuner.model.loader import load_model_and_tokenizer
 from llmtuner.model.parser import get_train_args, get_infer_args, get_eval_args
-from llmtuner.model.utils import create_ref_model, create_reward_model, dispatch_model, generate_model_card
+from llmtuner.model.utils import dispatch_model, generate_model_card, load_valuehead_params

src/llmtuner/model/parser.py

@@ -82,15 +82,13 @@ def get_train_args(args: Optional[Dict[str, Any]] = None) -> _TRAIN_CLS:
         raise ValueError("Please enable `predict_with_generate` to save model predictions.")
 
     if finetuning_args.stage in ["rm", "ppo"]:
-        if finetuning_args.finetuning_type != "lora":
-            raise ValueError("RM and PPO stages can only be performed with the LoRA method.")
         if training_args.resume_from_checkpoint is not None:
             raise ValueError("RM and PPO stages do not support `resume_from_checkpoint`.")
         if training_args.load_best_model_at_end:
             raise ValueError("RM and PPO stages do not support `load_best_model_at_end`.")
 
     if finetuning_args.stage == "ppo" and not training_args.do_train:
-        raise ValueError("PPO training does not support evaluation.")
+        raise ValueError("PPO training does not support evaluation, use the SFT stage to evaluate models.")
 
     if finetuning_args.stage in ["rm", "dpo"]:
         for dataset_attr in data_args.dataset_list:
@@ -131,6 +129,9 @@ def get_train_args(args: Optional[Dict[str, Any]] = None) -> _TRAIN_CLS:
     if (not training_args.do_train) and model_args.quantization_bit is not None:
         logger.warning("Evaluating model in 4/8-bit mode may cause lower scores.")
 
+    if (not training_args.do_train) and finetuning_args.stage == "dpo" and finetuning_args.ref_model is None:
+        logger.warning("Specify `ref_model` for computing rewards at evaluation.")
+
     # postprocess training_args
     if (
         training_args.local_rank != -1

src/llmtuner/model/utils.py

@@ -1,5 +1,5 @@
 import torch
-from typing import TYPE_CHECKING, Any, Dict, List, Literal, Optional, Set, Tuple, Union
+from typing import TYPE_CHECKING, Any, Dict, List, Optional, Set, Tuple
 
 from transformers.utils import cached_file
 from transformers.trainer import WEIGHTS_NAME, SAFE_WEIGHTS_NAME
@@ -7,83 +7,15 @@ from transformers.trainer import WEIGHTS_NAME, SAFE_WEIGHTS_NAME
 from llmtuner.extras.constants import LAYERNORM_NAMES
 from llmtuner.extras.logging import get_logger
 from llmtuner.hparams import ModelArguments, FinetuningArguments
-from llmtuner.model import load_model_and_tokenizer
 
 if TYPE_CHECKING:
     from transformers.modeling_utils import PreTrainedModel
-    from trl import AutoModelForCausalLMWithValueHead
     from llmtuner.hparams import DataArguments
 
 
 logger = get_logger(__name__)
 
 
-def create_ref_model(
-    model_args: "ModelArguments",
-    finetuning_args: "FinetuningArguments",
-    stage: Literal["ppo", "dpo"]
-) -> Union["PreTrainedModel", "AutoModelForCausalLMWithValueHead"]:
-    r"""
-    Creates reference model for PPO/DPO training. Evaluation mode is not supported.
-
-    The valuehead parameter is randomly initialized since it is useless for PPO training.
-    """
-    if finetuning_args.ref_model is not None:
-        ref_model_args_dict = model_args.to_dict()
-        ref_model_args_dict.update(dict(
-            model_name_or_path=finetuning_args.ref_model,
-            checkpoint_dir=finetuning_args.ref_model_checkpoint,
-            quantization_bit=finetuning_args.ref_model_quantization_bit
-        ))
-        ref_model_args = ModelArguments(**ref_model_args_dict)
-        ref_finetuning_args = FinetuningArguments(finetuning_type="lora")
-        ref_model, _ = load_model_and_tokenizer(ref_model_args, ref_finetuning_args, is_trainable=False, stage=stage)
-        logger.info("Created reference model from {}".format(finetuning_args.ref_model))
-    else:
-        if finetuning_args.finetuning_type == "lora":
-            ref_model = None
-        else:
-            ref_model, _ = load_model_and_tokenizer(model_args, finetuning_args, is_trainable=False, stage=stage)
-            logger.info("Created reference model from the model itself.")
-
-    return ref_model
-
-
-def create_reward_model(
-    model: "AutoModelForCausalLMWithValueHead",
-    model_args: "ModelArguments",
-    finetuning_args: "FinetuningArguments"
-) -> "AutoModelForCausalLMWithValueHead":
-    r"""
-    Creates reward model for PPO training.
-    """
-    if finetuning_args.reward_model_type == "lora":
-        model.pretrained_model.load_adapter(finetuning_args.reward_model, "reward")
-        for name, param in model.named_parameters(): # https://github.com/huggingface/peft/issues/1090
-            if "default" in name:
-                param.data = param.data.to(torch.float32) # trainable params should in fp32
-        vhead_params = load_valuehead_params(model_args.checkpoint_dir[-1], model_args)
-        assert vhead_params is not None, "Reward model is not correctly loaded."
-        model.register_buffer("reward_head_weight", vhead_params["v_head.summary.weight"], persistent=False)
-        model.register_buffer("reward_head_bias", vhead_params["v_head.summary.bias"], persistent=False)
-        model.register_buffer("default_head_weight", torch.zeros_like(vhead_params["v_head.summary.weight"]), persistent=False)
-        model.register_buffer("default_head_bias", torch.zeros_like(vhead_params["v_head.summary.bias"]), persistent=False)
-        logger.info("Loaded adapter weights of reward model from {}".format(finetuning_args.reward_model))
-        return None
-    else:
-        reward_model_args_dict = model_args.to_dict()
-        reward_model_args_dict.update(dict(
-            model_name_or_path=finetuning_args.reward_model,
-            checkpoint_dir=finetuning_args.reward_model_checkpoint,
-            quantization_bit=finetuning_args.reward_model_quantization_bit
-        ))
-        reward_model_args = ModelArguments(**reward_model_args_dict)
-        reward_finetuning_args = FinetuningArguments(finetuning_type="lora")
-        reward_model, _ = load_model_and_tokenizer(reward_model_args, reward_finetuning_args, is_trainable=False, stage="ppo")
-        logger.info("Load full weights of reward model from {}".format(finetuning_args.reward_model))
-        return reward_model
-
-
 def dispatch_model(model: "PreTrainedModel") -> "PreTrainedModel":
     r"""
     Dispatches a pre-trained model to GPUs with balanced memory.

src/llmtuner/train/dpo/workflow.py

@@ -6,10 +6,10 @@ from transformers import Seq2SeqTrainingArguments
 
 from llmtuner.data import get_dataset, preprocess_dataset, split_dataset
 from llmtuner.extras.constants import IGNORE_INDEX
-from llmtuner.extras.logging import get_logger
 from llmtuner.extras.ploting import plot_loss
 from llmtuner.hparams import ModelArguments
-from llmtuner.model import create_ref_model, generate_model_card, load_model_and_tokenizer
+from llmtuner.model import generate_model_card, load_model_and_tokenizer
+from llmtuner.train.utils import create_ref_model
 from llmtuner.train.dpo.collator import DPODataCollatorWithPadding
 from llmtuner.train.dpo.trainer import CustomDPOTrainer
 
@@ -18,9 +18,6 @@ if TYPE_CHECKING:
     from llmtuner.hparams import DataArguments, FinetuningArguments
 
 
-logger = get_logger(__name__)
-
-
 def run_dpo(
     model_args: "ModelArguments",
     data_args: "DataArguments",
@@ -74,7 +71,6 @@ def run_dpo(
     if training_args.do_eval:
         metrics = trainer.evaluate(metric_key_prefix="eval")
         if id(model) == id(ref_model): # unable to compute rewards without a reference model
-            logger.warning("Specify `ref_model` for computing rewards at evaluation.")
             remove_keys = [key for key in metrics.keys() if "rewards" in key]
             for key in remove_keys:
                 metrics.pop(key)

src/llmtuner/train/ppo/workflow.py

@@ -9,9 +9,9 @@ from transformers.optimization import get_scheduler
 
 from llmtuner.data import get_dataset, preprocess_dataset
 from llmtuner.extras.callbacks import SavePeftModelCallback
-from llmtuner.extras.logging import get_logger
 from llmtuner.extras.ploting import plot_loss
-from llmtuner.model import create_ref_model, create_reward_model, load_model_and_tokenizer
+from llmtuner.model import load_model_and_tokenizer
+from llmtuner.train.utils import create_ref_model, create_reward_model
 from llmtuner.train.ppo.trainer import CustomPPOTrainer
 
 if TYPE_CHECKING:
@@ -19,9 +19,6 @@ if TYPE_CHECKING:
     from llmtuner.hparams import ModelArguments, DataArguments, FinetuningArguments, GeneratingArguments
 
 
-logger = get_logger(__name__)
-
-
 def run_ppo(
     model_args: "ModelArguments",
     data_args: "DataArguments",

src/llmtuner/train/utils.py

@@ -0,0 +1,79 @@
+import torch
+from typing import TYPE_CHECKING, Literal, Union
+
+from llmtuner.extras.logging import get_logger
+from llmtuner.hparams import ModelArguments, FinetuningArguments
+from llmtuner.model import load_model_and_tokenizer, load_valuehead_params
+
+if TYPE_CHECKING:
+    from transformers.modeling_utils import PreTrainedModel
+    from trl import AutoModelForCausalLMWithValueHead
+
+
+logger = get_logger(__name__)
+
+
+def create_ref_model(
+    model_args: "ModelArguments",
+    finetuning_args: "FinetuningArguments",
+    stage: Literal["ppo", "dpo"]
+) -> Union["PreTrainedModel", "AutoModelForCausalLMWithValueHead"]:
+    r"""
+    Creates reference model for PPO/DPO training. Evaluation mode is not supported.
+
+    The valuehead parameter is randomly initialized since it is useless for PPO training.
+    """
+    if finetuning_args.ref_model is not None:
+        ref_model_args_dict = model_args.to_dict()
+        ref_model_args_dict.update(dict(
+            model_name_or_path=finetuning_args.ref_model,
+            checkpoint_dir=finetuning_args.ref_model_checkpoint,
+            quantization_bit=finetuning_args.ref_model_quantization_bit
+        ))
+        ref_model_args = ModelArguments(**ref_model_args_dict)
+        ref_finetuning_args = FinetuningArguments(finetuning_type="lora")
+        ref_model, _ = load_model_and_tokenizer(ref_model_args, ref_finetuning_args, is_trainable=False, stage=stage)
+        logger.info("Created reference model from {}".format(finetuning_args.ref_model))
+    else:
+        if finetuning_args.finetuning_type == "lora":
+            ref_model = None
+        else:
+            ref_model, _ = load_model_and_tokenizer(model_args, finetuning_args, is_trainable=False, stage=stage)
+            logger.info("Created reference model from the model itself.")
+
+    return ref_model
+
+
+def create_reward_model(
+    model: "AutoModelForCausalLMWithValueHead",
+    model_args: "ModelArguments",
+    finetuning_args: "FinetuningArguments"
+) -> "AutoModelForCausalLMWithValueHead":
+    r"""
+    Creates reward model for PPO training.
+    """
+    if finetuning_args.reward_model_type == "lora":
+        model.pretrained_model.load_adapter(finetuning_args.reward_model, "reward")
+        for name, param in model.named_parameters(): # https://github.com/huggingface/peft/issues/1090
+            if "default" in name:
+                param.data = param.data.to(torch.float32) # trainable params should in fp32
+        vhead_params = load_valuehead_params(model_args.checkpoint_dir[-1], model_args)
+        assert vhead_params is not None, "Reward model is not correctly loaded."
+        model.register_buffer("reward_head_weight", vhead_params["v_head.summary.weight"], persistent=False)
+        model.register_buffer("reward_head_bias", vhead_params["v_head.summary.bias"], persistent=False)
+        model.register_buffer("default_head_weight", torch.zeros_like(vhead_params["v_head.summary.weight"]), persistent=False)
+        model.register_buffer("default_head_bias", torch.zeros_like(vhead_params["v_head.summary.bias"]), persistent=False)
+        logger.info("Loaded adapter weights of reward model from {}".format(finetuning_args.reward_model))
+        return None
+    else:
+        reward_model_args_dict = model_args.to_dict()
+        reward_model_args_dict.update(dict(
+            model_name_or_path=finetuning_args.reward_model,
+            checkpoint_dir=finetuning_args.reward_model_checkpoint,
+            quantization_bit=finetuning_args.reward_model_quantization_bit
+        ))
+        reward_model_args = ModelArguments(**reward_model_args_dict)
+        reward_finetuning_args = FinetuningArguments(finetuning_type="lora")
+        reward_model, _ = load_model_and_tokenizer(reward_model_args, reward_finetuning_args, is_trainable=False, stage="ppo")
+        logger.info("Load full weights of reward model from {}".format(finetuning_args.reward_model))
+        return reward_model