remove save_strategy

docs/source/llm_finetuning.mdx

@@ -75,7 +75,7 @@ usage: autotrain <command> [<args>] llm [-h] [--train] [--deploy] [--inference]
                                         [--weight_decay WEIGHT_DECAY] [--max_grad_norm MAX_GRAD_NORM] [--add_eos_token] [--block_size BLOCK_SIZE]
                                         [--peft] [--lora_r LORA_R] [--lora_alpha LORA_ALPHA] [--lora_dropout LORA_DROPOUT]
                                         [--logging_steps LOGGING_STEPS] [--evaluation_strategy {epoch,steps,no}]
-                                        [--save_total_limit SAVE_TOTAL_LIMIT] [--save_strategy {epoch,steps}] [--auto_find_batch_size]
+                                        [--save_total_limit SAVE_TOTAL_LIMIT] [--auto_find_batch_size]
                                         [--mixed_precision {fp16,bf16,None}] [--quantization {int4,int8,None}] [--model_max_length MODEL_MAX_LENGTH]
                                         [--max_prompt_length MAX_PROMPT_LENGTH] [--max_completion_length MAX_COMPLETION_LENGTH]
                                         [--trainer {default,dpo,sft,orpo,reward}] [--target_modules TARGET_MODULES] [--merge_adapter]
@@ -162,8 +162,6 @@ options:
   --save_total_limit SAVE_TOTAL_LIMIT, --save-total-limit SAVE_TOTAL_LIMIT
                         Limit the total number of saved model checkpoints to manage disk usage effectively. Default is to save only the latest
                         checkpoint
-  --save_strategy {epoch,steps}, --save-strategy {epoch,steps}
-                        Define the checkpoint saving strategy, with 'epoch' as the default, saving checkpoints at the end of each training epoch.
   --auto_find_batch_size, --auto-find-batch-size
                         Automatically determine the optimal batch size based on system capabilities to maximize efficiency.
   --mixed_precision {fp16,bf16,None}, --mixed-precision {fp16,bf16,None}

docs/source/seq2seq.mdx

@@ -45,7 +45,7 @@ usage: autotrain <command> [<args>] seq2seq [-h] [--train] [--deploy] [--inferen
                                             [--optimizer OPTIMIZER] [--scheduler SCHEDULER] [--weight-decay WEIGHT_DECAY]
                                             [--max-grad-norm MAX_GRAD_NORM] [--logging-steps LOGGING_STEPS]
                                             [--evaluation-strategy EVALUATION_STRATEGY] [--save-total-limit SAVE_TOTAL_LIMIT]
-                                            [--save-strategy SAVE_STRATEGY] [--auto-find-batch-size] [--mixed-precision {fp16,bf16,None}] [--peft]
+                                            [--auto-find-batch-size] [--mixed-precision {fp16,bf16,None}] [--peft]
                                             [--quantization {int8,None}] [--lora-r LORA_R] [--lora-alpha LORA_ALPHA] [--lora-dropout LORA_DROPOUT]
                                             [--target-modules TARGET_MODULES]
 
@@ -118,9 +118,6 @@ options:
   --save-total-limit SAVE_TOTAL_LIMIT
                         Limit the total number of model checkpoints to save. Helps manage disk space by retaining only the most recent checkpoints.
                         Default is to save only the latest one.
-  --save-strategy SAVE_STRATEGY
-                        Determine the strategy for saving model checkpoints. Possible values are 'no', 'steps', 'epoch'. 'epoch' saves a checkpoint
-                        at the end of each epoch by default.
   --auto-find-batch-size
                         Enable automatic batch size determination based on your hardware capabilities. When set, it tries to find the largest batch
                         size that fits in memory.

src/autotrain/__init__.py

@@ -41,4 +41,4 @@
 
 
 logger = Logger().get_logger()
-__version__ = "0.7.71.dev0"
+__version__ = "0.7.72.dev0"

src/autotrain/app.py

@@ -96,7 +96,6 @@
     epochs=3,
     padding="right",
     chat_template="none",
-    save_strategy="no",
     max_completion_length=128,
 ).model_dump()
 
@@ -111,7 +110,6 @@
 PARAMS["seq2seq"] = Seq2SeqParams(
     mixed_precision="fp16",
     target_modules="all-linear",
-    save_strategy="no",
     log="tensorboard",
 ).model_dump()
 PARAMS["tabular"] = TabularParams(
@@ -257,7 +255,6 @@ async def fetch_params(task: str, param_type: str):
                         "logging_steps",
                         "evaluation_strategy",
                         "save_total_limit",
-                        "save_strategy",
                         "auto_find_batch_size",
                         "warmup_ratio",
                         "weight_decay",
@@ -281,7 +278,6 @@ async def fetch_params(task: str, param_type: str):
                 "logging_steps",
                 "auto_find_batch_size",
                 "save_total_limit",
-                "save_strategy",
                 "evaluation_strategy",
             ]
             task_params = {k: v for k, v in task_params.items() if k not in more_hidden_params}
@@ -294,7 +290,6 @@ async def fetch_params(task: str, param_type: str):
                 "logging_steps",
                 "auto_find_batch_size",
                 "save_total_limit",
-                "save_strategy",
                 "evaluation_strategy",
             ]
             task_params = {k: v for k, v in task_params.items() if k not in more_hidden_params}
@@ -307,7 +302,6 @@ async def fetch_params(task: str, param_type: str):
                 "logging_steps",
                 "auto_find_batch_size",
                 "save_total_limit",
-                "save_strategy",
                 "evaluation_strategy",
                 "quantization",
                 "lora_r",
@@ -325,7 +319,6 @@ async def fetch_params(task: str, param_type: str):
                 "logging_steps",
                 "auto_find_batch_size",
                 "save_total_limit",
-                "save_strategy",
                 "evaluation_strategy",
             ]
             task_params = {k: v for k, v in task_params.items() if k not in more_hidden_params}

src/autotrain/cli/run_llm.py

@@ -155,15 +155,6 @@ def register_subcommand(parser: ArgumentParser):
                 "default": 1,
                 "alias": ["--save-total-limit"],
             },
-            {
-                "arg": "--save_strategy",
-                "help": "Define the checkpoint saving strategy, with 'epoch' as the default, saving checkpoints at the end of each training epoch. ",
-                "required": False,
-                "type": str,
-                "default": "epoch",
-                "alias": ["--save-strategy"],
-                "choices": ["epoch", "steps"],
-            },
             {
                 "arg": "--auto_find_batch_size",
                 "help": "Automatically determine the optimal batch size based on system capabilities to maximize efficiency.",

src/autotrain/cli/run_seq2seq.py

@@ -207,11 +207,6 @@ def __init__(self, args):
         else:
             raise ValueError("Must specify --train, --deploy or --inference")
 
-        if len(self.args.target_modules.strip()) == 0:
-            self.args.target_modules = []
-        else:
-            self.args.target_modules = self.args.target_modules.split(",")
-
     def run(self):
         logger.info("Running Seq2Seq Classification")
         if self.args.train:

src/autotrain/trainers/clm/__main__.py

@@ -228,7 +228,7 @@ def train(config):
         evaluation_strategy=config.evaluation_strategy if config.valid_split is not None else "no",
         logging_steps=logging_steps,
         save_total_limit=config.save_total_limit,
-        save_strategy=config.save_strategy,
+        save_strategy=config.evaluation_strategy if config.valid_split is not None else "no",
         gradient_accumulation_steps=config.gradient_accumulation,
         report_to=config.log,
         auto_find_batch_size=config.auto_find_batch_size,

src/autotrain/trainers/clm/params.py

@@ -26,7 +26,6 @@ class LLMTrainingParams(AutoTrainParams):
     logging_steps: int = Field(-1, title="Logging steps")
     evaluation_strategy: str = Field("epoch", title="Evaluation strategy")
     save_total_limit: int = Field(1, title="Save total limit")
-    save_strategy: str = Field("no", title="Save strategy")
     auto_find_batch_size: bool = Field(False, title="Auto find batch size")
     mixed_precision: Optional[str] = Field(None, title="fp16, bf16, or None")
     lr: float = Field(3e-5, title="Learning rate")

src/autotrain/trainers/image_classification/__main__.py

@@ -129,7 +129,7 @@ def train(config):
         evaluation_strategy=config.evaluation_strategy if config.valid_split is not None else "no",
         logging_steps=logging_steps,
         save_total_limit=config.save_total_limit,
-        save_strategy=config.save_strategy,
+        save_strategy=config.evaluation_strategy if config.valid_split is not None else "no",
         gradient_accumulation_steps=config.gradient_accumulation,
         report_to=config.log,
         auto_find_batch_size=config.auto_find_batch_size,

src/autotrain/trainers/image_classification/params.py

@@ -26,7 +26,6 @@ class ImageClassificationParams(AutoTrainParams):
     auto_find_batch_size: bool = Field(False, title="Auto find batch size")
     mixed_precision: Optional[str] = Field(None, title="fp16, bf16, or None")
     save_total_limit: int = Field(1, title="Save total limit")
-    save_strategy: str = Field("epoch", title="Save strategy")
     token: Optional[str] = Field(None, title="Hub Token")
     push_to_hub: bool = Field(False, title="Push to hub")
     evaluation_strategy: str = Field("epoch", title="Evaluation strategy")

src/autotrain/trainers/seq2seq/__main__.py

@@ -102,7 +102,7 @@ def train(config):
         evaluation_strategy=config.evaluation_strategy if config.valid_split is not None else "no",
         logging_steps=logging_steps,
         save_total_limit=config.save_total_limit,
-        save_strategy=config.save_strategy,
+        save_strategy=config.evaluation_strategy if config.valid_split is not None else "no",
         gradient_accumulation_steps=config.gradient_accumulation,
         report_to=config.log,
         auto_find_batch_size=config.auto_find_batch_size,
@@ -179,6 +179,8 @@ def train(config):
         target_modules = config.target_modules.split(",") if config.target_modules is not None else None
         if target_modules:
             target_modules = [module.strip() for module in target_modules]
+        if len(target_modules) == 1 and target_modules[0] == "all-linear":
+            target_modules = "all-linear"
         lora_config = LoraConfig(
             r=config.lora_r,
             lora_alpha=config.lora_alpha,

src/autotrain/trainers/seq2seq/params.py

@@ -33,7 +33,6 @@ class Seq2SeqParams(AutoTrainParams):
     auto_find_batch_size: bool = Field(False, title="Auto find batch size")
     mixed_precision: Optional[str] = Field(None, title="fp16, bf16, or None")
     save_total_limit: int = Field(1, title="Save total limit")
-    save_strategy: str = Field("no", title="Save strategy")
     token: Optional[str] = Field(None, title="Hub Token")
     push_to_hub: bool = Field(False, title="Push to hub")
     peft: bool = Field(False, title="Use PEFT")

src/autotrain/trainers/text_classification/__main__.py

@@ -133,7 +133,7 @@ def train(config):
         evaluation_strategy=config.evaluation_strategy if config.valid_split is not None else "no",
         logging_steps=logging_steps,
         save_total_limit=config.save_total_limit,
-        save_strategy=config.save_strategy,
+        save_strategy=config.evaluation_strategy if config.valid_split is not None else "no",
         gradient_accumulation_steps=config.gradient_accumulation,
         report_to=config.log,
         auto_find_batch_size=config.auto_find_batch_size,

src/autotrain/trainers/text_classification/params.py

@@ -28,7 +28,6 @@ class TextClassificationParams(AutoTrainParams):
     auto_find_batch_size: bool = Field(False, title="Auto find batch size")
     mixed_precision: Optional[str] = Field(None, title="fp16, bf16, or None")
     save_total_limit: int = Field(1, title="Save total limit")
-    save_strategy: str = Field("epoch", title="Save strategy")
     token: Optional[str] = Field(None, title="Hub Token")
     push_to_hub: bool = Field(False, title="Push to hub")
     evaluation_strategy: str = Field("epoch", title="Evaluation strategy")

src/autotrain/trainers/token_classification/__main__.py

@@ -131,7 +131,7 @@ def train(config):
         evaluation_strategy=config.evaluation_strategy if config.valid_split is not None else "no",
         logging_steps=logging_steps,
         save_total_limit=config.save_total_limit,
-        save_strategy=config.save_strategy,
+        save_strategy=config.evaluation_strategy if config.valid_split is not None else "no",
         gradient_accumulation_steps=config.gradient_accumulation,
         report_to=config.log,
         auto_find_batch_size=config.auto_find_batch_size,

src/autotrain/trainers/token_classification/params.py

@@ -28,7 +28,6 @@ class TokenClassificationParams(AutoTrainParams):
     auto_find_batch_size: bool = Field(False, title="Auto find batch size")
     mixed_precision: Optional[str] = Field(None, title="fp16, bf16, or None")
     save_total_limit: int = Field(1, title="Save total limit")
-    save_strategy: str = Field("epoch", title="Save strategy")
     token: Optional[str] = Field(None, title="Hub Token")
     push_to_hub: bool = Field(False, title="Push to hub")
     evaluation_strategy: str = Field("epoch", title="Evaluation strategy")