fix: multiple bug fixes - remove unused imports, fix typo, fix dataset paths, add ai command fallback

train-on-this-server.sh

@@ -12,8 +12,12 @@ echo ""
 
 # Unload AI model to free VRAM
 echo "=== Unloading AI model ==="
+if command -v ai &> /dev/null; then
     ai none
     sleep 2
+else
+    echo "Warning: 'ai' command not found, skipping AI unload"
+fi
 
 # Check Python
 if ! command -v python3 &> /dev/null; then

training/configs/llama2-7b-lora.yaml

@@ -50,7 +50,7 @@ mixed_precision: bf16
 plugin: deepspeed
 huggingface_hub:
   token: null
-deeepspeed_config:
+deepspeed_config:
   zero_optimization:
     stage: 3
     offload_optimizer:

training/scripts/train.py

@@ -22,15 +22,13 @@ def train(config_path):
     from transformers import (
         AutoModelForCausalLM,
         AutoTokenizer,
-        BitsAndBytesConfig,
         TrainingArguments,
     )
     from peft import (
-        prepare_model_for_kbit_training,
         LoraConfig,
         get_peft_model,
     )
-    from trl import SFTTrainer, SFTConfig
+    from trl import SFTTrainer
     
     print(f"Loading model: {config['base_model']}")
     
@@ -61,12 +59,18 @@ def train(config_path):
     
     # Load dataset
     from datasets import load_dataset
+    import os
+    
+    # Get dataset paths relative to script location
+    script_dir = Path(__file__).parent.parent
+    train_path = str(script_dir / ".." / ".." / "training" / "data" / "train.jsonl")
+    test_path = str(script_dir / ".." / ".." / "training" / "data" / "test.jsonl")
     
     dataset = load_dataset(
         "json",
         data_files={
-            "train": config["dataset"][0]["path"].replace("../", ""),
-            "test": config["dataset"][0]["path"].replace("../", "").replace("combined_20k.jsonl", "test.jsonl"),
+            "train": train_path,
+            "test": test_path,
         },
     )
     
@@ -99,6 +103,7 @@ def train(config_path):
         train_dataset=dataset["train"],
         eval_dataset=dataset["test"],
         args=training_args,
+        dataset_text_field="text",
         max_seq_length=config["train_params"]["max_seq_length"],
     )