Z-Image-Turbo - 图生图，图像反推提示词生图 批量洗图 支持50系显卡 一键整合包下载

138***

大大，有时间的话帮看看这是什么原因造成，刚打开点击生成图像出现下面错误代码。还有救嘛？

无言***

1387573 发表于 2026-7-3 09:28
大大，有时间的话帮看看这是什么原因造成，刚打开点击生成图像出现下面错误代码。还有救嘛？
...

看看帖子里使用教程里的说明，对比下自己的硬件是否符合要求

138***

无言以对 发表于 2026-7-3 09:42
看看帖子里使用教程里的说明，对比下自己的硬件是否符合要求

大大，我又去看了下全部的教程和硬件要求，我应该符合低要求：英特尔-i5-13400，技嘉B750还是650主板给忘记了，技嘉大雕rtx 3080显卡（10GB），4代DDR 32G，1Tb的固态硬盘就给装的系统盘，还有4Tb硬盘。
下了几个模型都没问题，比如Z-IMAGE-白玉AIO,KREA2-V3,QWEN-IMAGE-EDIT-AIO-V8等。都能顺利运行，就这个出现了这个情况。

无言***

1387573 发表于 2026-7-3 18:47
大大，我又去看了下全部的教程和硬件要求，我应该符合低要求：英特尔-i5-13400，技嘉B750还是650主板给忘 ...

应该是超时了，设置了默认300s的等待时间，也就是5分钟后如果没启动就会失败。
正常情况5分钟是足够了
你先提交试试，提示失败了，再次启动试试（最好放固态，加载模型更快，毕竟你的卡有点老了）

138***

无言以对 发表于 2026-7-3 18:55
应该是超时了，设置了默认300s的等待时间，也就是5分钟后如果没启动就会失败。
正常情况5分钟是足够了
你 ...

大大，模型都放在固态硬盘，这个错误提示是点了之后2秒内就跳出来了，就添加了一张图片，其余都是默认，也是这样。下方是跳出来的信息，大大有空可以看下。
[19:29:31] [ComfyUI] got prompt
[19:29:32] [ComfyUI] `torch_dtype` is deprecated! Use `dtype` instead!
[19:29:32] [ComfyUI] !!! Exception during processing !!! FP8 quantized models is only supported on GPUs with compute capability >= 8.9 (e.g 4090/H100), actual = `8.6`
[19:29:32] [ComfyUI] Traceback (most recent call last):
[19:29:32] [ComfyUI] File "C:\AI\Z-Image-V6\ComfyUI\execution.py", line 516, in execute
[19:29:32] [ComfyUI] 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)
[19:29:32] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[19:29:32] [ComfyUI] File "C:\AI\Z-Image-V6\ComfyUI\execution.py", line 330, in get_output_data
[19:29:32] [ComfyUI] 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)
[19:29:32] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[19:29:32] [ComfyUI] File "C:\AI\Z-Image-V6\ComfyUI\execution.py", line 304, in _async_map_node_over_list
[19:29:32] [ComfyUI] await process_inputs(input_dict, i)
[19:29:32] [ComfyUI] File "C:\AI\Z-Image-V6\ComfyUI\execution.py", line 292, in process_inputs
[19:29:32] [ComfyUI] result = f(**inputs)
[19:29:32] [ComfyUI] ^^^^^^^^^^^
[19:29:32] [ComfyUI] File "C:\AI\Z-Image-V6\ComfyUI\custom_nodes\ComfyUI-QwenVL\AILab_QwenVL.py", line 935, in process
[19:29:32] [ComfyUI] return self.run(model_name, quantization, preset_prompt, custom_prompt, image, video, frame_count, max_tokens, temperature, top_p, num_beams, repetition_penalty, seed, keep_model_loaded, attention_mode, use_torch_compile, device)
[19:29:32] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[19:29:32] [ComfyUI] File "C:\AI\Z-Image-V6\ComfyUI\custom_nodes\ComfyUI-QwenVL\AILab_QwenVL.py", line 828, in run
[19:29:32] [ComfyUI] self.load_model(
[19:29:32] [ComfyUI] File "C:\AI\Z-Image-V6\ComfyUI\custom_nodes\ComfyUI-QwenVL\AILab_QwenVL.py", line 653, in load_model
[19:29:32] [ComfyUI] self.model = AutoModelForVision2Seq.from_pretrained(model_path, **load_kwargs)
[19:29:32] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[19:29:32] [ComfyUI] File "C:\AI\Z-Image-V6\deepface\Lib\site-packages\transformers\models\auto\auto_factory.py", line 604, in from_pretrained
[19:29:32] [ComfyUI] return model_class.from_pretrained(
[19:29:32] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[19:29:32] [ComfyUI] File "C:\AI\Z-Image-V6\deepface\Lib\site-packages\transformers\modeling_utils.py", line 277, in _wrapper
[19:29:32] [ComfyUI] return func(*args, **kwargs)
[19:29:32] [ComfyUI] ^^^^^^^^^^^^^^^^^^^^^
[19:29:32] [ComfyUI] File "C:\AI\Z-Image-V6\deepface\Lib\site-packages\transformers\modeling_utils.py", line 4881, in from_pretrained
[19:29:32] [ComfyUI] hf_quantizer, config, dtype, device_map = get_hf_quantizer(
[19:29:32] [ComfyUI] ^^^^^^^^^^^^^^^^^
[19:29:32] [ComfyUI] File "C:\AI\Z-Image-V6\deepface\Lib\site-packages\transformers\quantizers\auto.py", line 319, in get_hf_quantizer
[19:29:32] [ComfyUI] hf_quantizer.validate_environment(
[19:29:32] [ComfyUI] File "C:\AI\Z-Image-V6\deepface\Lib\site-packages\transformers\quantizers\quantizer_finegrained_fp8.py", line 54, in validate_environment
[19:29:32] [ComfyUI] raise ValueError(
[19:29:32] [ComfyUI] ValueError: FP8 quantized models is only supported on GPUs with compute capability >= 8.9 (e.g 4090/H100), actual = `8.6`
[19:29:32] [ComfyUI] Prompt executed in 0.29 seconds
[19:29:33] [Gradio] Traceback (most recent call last):
[19:29:33] [Gradio] File "<frozen app>", line 382, in generate_image
[19:29:33] [Gradio] File "<frozen app>", line 346, in wait_for_completion
[19:29:33] [Gradio] Exception: 工作流执行出错
[19:29:33] [Gradio] During handling of the above exception, another exception occurred:
[19:29:33] [Gradio] Traceback (most recent call last):
[19:29:33] [Gradio] File "C:\AI\Z-Image-V6\deepface\Lib\site-packages\gradio\queueing.py", line 763, in process_events
[19:29:33] [Gradio] response = await route_utils.call_process_api(
[19:29:33] [Gradio] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[19:29:33] [Gradio] File "C:\AI\Z-Image-V6\deepface\Lib\site-packages\gradio\route_utils.py", line 354, in call_process_api
[19:29:33] [Gradio] output = await app.get_blocks().process_api(
[19:29:33] [Gradio] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[19:29:33] [Gradio] File "C:\AI\Z-Image-V6\deepface\Lib\site-packages\gradio\blocks.py", line 2106, in process_api
[19:29:33] [Gradio] result = await self.call_function(
[19:29:33] [Gradio] ^^^^^^^^^^^^^^^^^^^^^^^^^
[19:29:33] [Gradio] File "C:\AI\Z-Image-V6\deepface\Lib\site-packages\gradio\blocks.py", line 1588, in call_function
[19:29:33] [Gradio] prediction = await anyio.to_thread.run_sync(  # type: ignore
[19:29:33] [Gradio] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[19:29:33] [Gradio] File "C:\AI\Z-Image-V6\deepface\Lib\site-packages\anyio\to_thread.py", line 56, in run_sync
[19:29:33] [Gradio] return await get_async_backend().run_sync_in_worker_thread(
[19:29:33] [Gradio] ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
[19:29:33] [Gradio] File "C:\AI\Z-Image-V6\deepface\Lib\site-packages\anyio\_backends\_asyncio.py", line 2485, in run_sync_in_worker_thread
[19:29:33] [Gradio] return await future
[19:29:33] [Gradio] ^^^^^^^^^^^^
[19:29:33] [Gradio] File "C:\AI\Z-Image-V6\deepface\Lib\site-packages\anyio\_backends\_asyncio.py", line 976, in run
[19:29:33] [Gradio] result = context.run(func, *args)
[19:29:33] [Gradio] ^^^^^^^^^^^^^^^^^^^^^^^^
[19:29:33] [Gradio] File "C:\AI\Z-Image-V6\deepface\Lib\site-packages\gradio\utils.py", line 1048, in wrapper
[19:29:33] [Gradio] response = f(*args, **kwargs)
[19:29:33] [Gradio] ^^^^^^^^^^^^^^^^^^
[19:29:33] [Gradio] File "<frozen app>", line 384, in generate_image
[19:29:33] [Gradio] gradio.exceptions.Error: '生成失败: 工作流执行出错'

无言***

1387573 发表于 2026-7-3 19:34
大大，模型都放在固态硬盘，这个错误提示是点了之后2秒内就跳出来了，就添加了一张图片，其余都是默认， ...

Exception during processing !!! FP8 quantized models is only supported on GPUs with compute capability >= 8.9 (e.g 4090/H100), actual = `8.6`

这里是关键原因，30系不支持原生fp8加速，但正常来说会回退到非fp8机制
你试试显卡驱动更新到最新，cuda安装12.X版试试

138***

无言以对 发表于 2026-7-3 22:54
Exception during processing !!! FP8 quantized models is only supported on GPUs with compute capabi ...

除了CUDA是13.3，其他的都符合要求。好像上次看到cuda可以共存，我再去装12.x试试看，谢谢大大。

138***

无言以对 发表于 2026-7-3 22:54
Exception during processing !!! FP8 quantized models is only supported on GPUs with compute capabi ...

刚才安装了CUDA_12.9.2,并在path中把\bin和\libnvvp两个环境变量上移到了顶部，从新开启还是这样的老问题，算了，以后再说，这个也不是很重要。

无言***

1387573 发表于 2026-7-4 09:59
刚才安装了CUDA_12.9.2,并在path中把\bin和\libnvvp两个环境变量上移到了顶部，从新开启还是这样的老问题 ...

那只能用gguf模型了，后期我更新会带上gguf

138***

无言以对 发表于 2026-7-4 10:05
那只能用gguf模型了，后期我更新会带上gguf

大大辛苦了，个人需求，而且也不是急需，等有空了慢慢弄。