fix: parse tool_call arguments from JSON string to dict for Qwen3.5 template

train_qwen35_27b.py

@@ -54,9 +54,33 @@ model = FastLanguageModel.get_peft_model(
 # --- Dataset ---
 ds = load_dataset("json", data_files=DATA, split="train")
 
+def fix_tool_calls(messages):
+    """Parse tool_call arguments from JSON strings to dicts for Qwen3.5 template."""
+    import json as _json
+    fixed = []
+    for msg in messages:
+        msg = dict(msg)
+        if msg.get("tool_calls"):
+            new_tcs = []
+            for tc in msg["tool_calls"]:
+                tc = dict(tc)
+                if "function" in tc:
+                    fn = dict(tc["function"])
+                    if isinstance(fn.get("arguments"), str):
+                        try:
+                            fn["arguments"] = _json.loads(fn["arguments"])
+                        except (ValueError, TypeError):
+                            fn["arguments"] = {"raw": fn["arguments"]}
+                    tc["function"] = fn
+                new_tcs.append(tc)
+            msg["tool_calls"] = new_tcs
+        fixed.append(msg)
+    return fixed
+
 def to_chatml(ex):
+    messages = fix_tool_calls(ex["messages"])
     text = tokenizer.apply_chat_template(
-        ex["messages"], tokenize=False, add_generation_prompt=False
+        messages, tokenize=False, add_generation_prompt=False
     )
     return {"text": text}