Issue using qwen2.5 VL with official wf template

### Custom Node Testing

- [x] I have tried disabling custom nodes and the issue persists (see [how to disable custom nodes](https://docs.comfy.org/troubleshooting/custom-node-issues#step-1%3A-test-with-all-custom-nodes-disabled) if you need help)

### Expected Behavior

With the official template llm_qwen3_text_gen , you can't use the load image node built in , the model doesn't see the image provided.

### Actual Behavior

The model say it see nothing (qwen3) or generate gibberish qwen2.5 VL

### Steps to Reproduce

Use the official template

### Debug Logs

```powershell
[ERROR] !!! Exception during processing !!! 'Qwen25_7BVLI_Config' object has no attribute 'stop_tokens'
[ERROR] Traceback (most recent call last):
File "/home/zeuss194/COMFY/ComfyUI/execution.py", line 536, in execute
output_data, output_ui, has_subgraph, has_pending_tasks = await get_output_data(prompt_id, unique_id, obj, input_data_all, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/zeuss194/COMFY/ComfyUI/execution.py", line 336, in get_output_data
return_values = await _async_map_node_over_list(prompt_id, unique_id, obj, input_data_all, obj.FUNCTION, allow_interrupt=True, execution_block_cb=execution_block_cb, pre_execute_cb=pre_execute_cb, v3_data=v3_data)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/zeuss194/COMFY/ComfyUI/execution.py", line 310, in _async_map_node_over_list
await process_inputs(input_dict, i)
File "/home/zeuss194/COMFY/ComfyUI/execution.py", line 298, in process_inputs
result = f(**inputs)
File "/home/zeuss194/COMFY/ComfyUI/comfy_api/internal/__init__.py", line 149, in wrapped_func
return method(locked_class, **inputs)
File "/home/zeuss194/COMFY/ComfyUI/comfy_api/latest/_io.py", line 1833, in EXECUTE_NORMALIZED
to_return = cls.execute(*args, **kwargs)
File "/home/zeuss194/COMFY/ComfyUI/comfy_extras/nodes_textgen.py", line 63, in execute
generated_ids = clip.generate(
tokens,
...<8 lines>...
seed=seed
)
File "/home/zeuss194/COMFY/ComfyUI/comfy/sd.py", line 460, in generate
return self.cond_stage_model.generate(tokens, do_sample=do_sample, max_length=max_length, temperature=temperature, top_k=top_k, top_p=top_p, min_p=min_p, repetition_penalty=repetition_penalty, seed=seed, presence_penalty=presence_penalty)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/zeuss194/COMFY/ComfyUI/comfy/sd1_clip.py", line 744, in generate
return getattr(self, self.clip).generate(tokens, do_sample=do_sample, max_length=max_length, temperature=temperature, top_k=top_k, top_p=top_p, min_p=min_p, repetition_penalty=repetition_penalty, seed=seed, presence_penalty=presence_penalty)
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/zeuss194/COMFY/ComfyUI/comfy/sd1_clip.py", line 318, in generate
return self.transformer.generate(embeds, do_sample, max_length, temperature, top_k, top_p, min_p, repetition_penalty, seed, presence_penalty=presence_penalty)
~~~~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "/home/zeuss194/COMFY/ComfyUI/comfy/text_encoders/llama.py", line 867, in generate
stop_tokens = self.model.config.stop_tokens
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
AttributeError: 'Qwen25_7BVLI_Config' object has no attribute 'stop_tokens'
```

### Other

The two core bugs:

image= kwarg never reached the tokenizer

lm_head.weight never loaded (ComfyUI renames the key but config didn't create model.lm_head)