update configs

huggingface · May 23, 2024 · 5e0160c · 5e0160c
1 parent 027f4b0
commit 5e0160c
Show file tree

Hide file tree

Showing 9 changed files with 20 additions and 16 deletions.
diff --git a/configs/llm_finetuning/gpt2_sft.yml b/configs/llm_finetuning/gpt2_sft.yml
@@ -1,4 +1,4 @@
-task: llm
+task: llm-sft
 base_model: openai-community/gpt2
 project_name: autotrain-gpt2-finetuned-guanaco
 log: tensorboard
@@ -13,7 +13,6 @@ data:
     text_column: text
 
 params:
-  trainer: sft
   block_size: 1024
   model_max_length: 2048
   max_prompt_length: 512

diff --git a/configs/llm_finetuning/llama3-70b-orpo-v1.yml b/configs/llm_finetuning/llama3-70b-orpo-v1.yml
@@ -1,4 +1,4 @@
-task: llm
+task: llm-orpo
 base_model: meta-llama/Meta-Llama-3-70B-Instruct
 project_name: autotrain-llama3-70b-orpo-v1
 log: tensorboard
@@ -15,7 +15,6 @@ data:
     prompt_text_column: prompt
 
 params:
-  trainer: orpo
   block_size: 2048
   model_max_length: 8192
   max_prompt_length: 1024

diff --git a/configs/llm_finetuning/llama3-70b-sft.yml b/configs/llm_finetuning/llama3-70b-sft.yml
@@ -1,4 +1,4 @@
-task: llm
+task: llm-sft
 base_model: meta-llama/Meta-Llama-3-70B-Instruct
 project_name: autotrain-llama3-70b-math-v1
 log: tensorboard
@@ -13,7 +13,6 @@ data:
     text_column: text
 
 params:
-  trainer: sft
   block_size: 2048
   model_max_length: 8192
   epochs: 2

diff --git a/configs/llm_finetuning/llama3-8b-dpo-qlora.yml b/configs/llm_finetuning/llama3-8b-dpo-qlora.yml
@@ -1,4 +1,4 @@
-task: llm
+task: llm-dpo
 base_model: meta-llama/Meta-Llama-3-8B-Instruct
 project_name: autotrain-llama3-8b-dpo-qlora
 log: tensorboard
@@ -15,7 +15,6 @@ data:
     prompt_text_column: prompt
 
 params:
-  trainer: dpo
   block_size: 1024
   model_max_length: 2048
   max_prompt_length: 512

diff --git a/configs/llm_finetuning/llama3-8b-orpo-space.yml b/configs/llm_finetuning/llama3-8b-orpo-space.yml
@@ -1,4 +1,4 @@
-task: llm
+task: llm-orpo
 base_model: meta-llama/Meta-Llama-3-8B-Instruct
 project_name: autotrain-llama3-8b-orpo-t1
 log: tensorboard
@@ -15,7 +15,6 @@ data:
     prompt_text_column: prompt
 
 params:
-  trainer: orpo
   block_size: 1024
   model_max_length: 8192
   max_prompt_length: 512

diff --git a/configs/llm_finetuning/llama3-8b-orpo.yml b/configs/llm_finetuning/llama3-8b-orpo.yml
@@ -1,4 +1,4 @@
-task: llm
+task: llm-orpo
 base_model: meta-llama/Meta-Llama-3-8B-Instruct
 project_name: autotrain-llama3-8b-orpo
 log: tensorboard
@@ -15,7 +15,6 @@ data:
     prompt_text_column: prompt
 
 params:
-  trainer: orpo
   block_size: 1024
   model_max_length: 8192
   max_prompt_length: 512

diff --git a/src/autotrain/__init__.py b/src/autotrain/__init__.py
@@ -41,7 +41,7 @@
 warnings.filterwarnings("ignore", category=UserWarning, module="huggingface_hub")
 
 logger = Logger().get_logger()
-__version__ = "0.7.106"
+__version__ = "0.7.108"
 
 
 def is_colab():

diff --git a/src/autotrain/app/params.py b/src/autotrain/app/params.py
@@ -283,9 +283,10 @@ def _munge_params_img_obj_det(self):
         _params["log"] = "tensorboard"
         if not self.using_hub_dataset:
             _params["image_column"] = "autotrain_image"
-            _params["objects_column"] = "autotrain_label"
+            _params["objects_column"] = "autotrain_objects"
             _params["valid_split"] = "validation"
         else:
+
             _params["image_column"] = self.column_mapping.get("image" if not self.api else "image_column", "image")
             _params["objects_column"] = self.column_mapping.get(
                 "objects" if not self.api else "objects_column", "objects"

diff --git a/src/autotrain/parser.py b/src/autotrain/parser.py
@@ -70,8 +70,11 @@ def __post_init__(self):
         }
         self.task_aliases = {
             "llm": "lm_training",
-            "llm_training": "lm_training",
-            "llm_finetuning": "lm_training",
+            "llm-sft": "lm_training",
+            "llm-orpo": "lm_training",
+            "llm-generic": "lm_training",
+            "llm-dpo": "lm_training",
+            "llm-reward": "lm_training",
             "dreambooth": "dreambooth",
             "image_binary_classification": "image_multi_class_classification",
             "image-binary-classification": "image_multi_class_classification",
@@ -123,6 +126,12 @@ def _parse_config(self):
 
         if self.task == "lm_training":
             params["chat_template"] = self.config["data"]["chat_template"]
+            if "-" in self.config["task"]:
+                params["trainer"] = self.config["task"].split("-")[1]
+                if params["trainer"] == "generic":
+                    params["trainer"] = "default"
+                if params["trainer"] not in ["sft", "orpo", "dpo", "reward", "default"]:
+                    raise ValueError("Invalid LLM training task")
 
         if self.task != "dreambooth":
             for k, v in self.config["data"]["column_mapping"].items():