Initial

app.py

@@ -4,14 +4,14 @@ import threading
 
 # Third-party imports
 import gradio as gr
-from peft import PeftModel
 from transformers import AutoTokenizer, AutoModelForCausalLM, TextIteratorStreamer
+from peft import PeftModel
 
 HF_TOKEN = os.getenv("HF_TOKEN")
 
 tokenizer = AutoTokenizer.from_pretrained(
     "bunyaminergen/Qwen2.5-Coder-1.5B-Instruct-Reasoning",
-    use_auth_token=HF_TOKEN,
+    token=HF_TOKEN,
     trust_remote_code=True
 )
 
@@ -19,12 +19,16 @@ base_model = AutoModelForCausalLM.from_pretrained(
     "Qwen/Qwen2.5-Coder-1.5B-Instruct",
     device_map="auto",
     torch_dtype="auto",
-    use_auth_token=HF_TOKEN
+    token=HF_TOKEN
 )
+
+base_model.resize_token_embeddings(len(tokenizer))
+
+# 4️⃣ PEFT adapter’ı yükle
 model = PeftModel.from_pretrained(
     base_model,
     "bunyaminergen/Qwen2.5-Coder-1.5B-Instruct-Reasoning",
-    use_auth_token=HF_TOKEN
+    token=HF_TOKEN
 )
 model.eval()