Fix indentation bug in grpo_train.py + update requirements.txt

requirements.txt

@@ -2,9 +2,15 @@ fastapi>=0.110.0
 uvicorn[standard]>=0.29.0
 pydantic>=2.0
 openenv-core>=0.2.3
-transformers
-datasets
-trl
-peft
+transformers>=4.44.0
+datasets>=2.20.0
+trl>=0.11.0
+peft>=0.11.0
 httpx
 matplotlib
+accelerate>=0.33.0
+bitsandbytes>=0.43.0
+huggingface_hub[cli]>=0.24.0
+hf_transfer>=0.1.8
+numpy<2
+torch>=2.0.0

training/grpo_train.py

@@ -38,8 +38,8 @@ def _load_model_and_tokenizer(model_name: str, use_unsloth: bool = False):
     if use_unsloth:
         try:
             from unsloth import FastLanguageModel
-            print("Loading with Unsloth in 4-bit + LoRA...")
-                        model, tokenizer = FastLanguageModel.from_pretrained(
+                        print("Loading with Unsloth in 4-bit + LoRA...")
+            model, tokenizer = FastLanguageModel.from_pretrained(
                 model_name=model_name,
                 max_seq_length=2048,
                 load_in_4bit=True,