[Bug]: 在WSL中运行环境，加载LLaMA+No LoRA，对话 Tokenizer class LLaMATokenizer does not exist or is not currently imported.

[yinguohang]

这个bug是否已存在现有issue了？

• 我确认没有已有issue，且已阅读常见问题。

错误表现

报错
Tokenizer class LLaMATokenizer does not exist or is not currently imported.

复现操作

1. 在WSL中完成部署
2. 在网页中切换到llama-7b-hf + No LoRA
3. 在对话框中输入你好

错误日志

2023-04-14 10:55:12,728 [INFO] [models.py:524] 现在请为 llama-7b-hf 选择LoRA模型
2023-04-14 10:55:12,731 [INFO] [models.py:543] 现在请为 llama-7b-hf 选择LoRA模型
2023-04-14 10:55:14,844 [INFO] [models.py:531] 正在加载LLaMA模型: llama-7b-hf + No LoRA
2023-04-14 10:55:16,772 [WARNING] [hf_decoder_model.py:192] llama does not support RAM optimized load. Automatically use original load instead.
Traceback (most recent call last):
  File "/home/ygh/.local/lib/python3.10/site-packages/gradio/routes.py", line 401, in run_predict
    output = await app.get_blocks().process_api(
  File "/home/ygh/.local/lib/python3.10/site-packages/gradio/blocks.py", line 1302, in process_api
    result = await self.call_function(
  File "/home/ygh/.local/lib/python3.10/site-packages/gradio/blocks.py", line 1025, in call_function
    prediction = await anyio.to_thread.run_sync(
  File "/home/ygh/.local/lib/python3.10/site-packages/anyio/to_thread.py", line 31, in run_sync
    return await get_asynclib().run_sync_in_worker_thread(
  File "/home/ygh/.local/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 937, in run_sync_in_worker_thread
    return await future
  File "/home/ygh/.local/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 867, in run
    result = context.run(func, *args)
  File "/mnt/d/ChuanhuChatGPT/modules/utils.py", line 43, in billing_info
    return current_model.billing_info()
AttributeError: 'NoneType' object has no attribute 'billing_info'
Traceback (most recent call last):
  File "/home/ygh/.local/lib/python3.10/site-packages/gradio/routes.py", line 401, in run_predict
    output = await app.get_blocks().process_api(
  File "/home/ygh/.local/lib/python3.10/site-packages/gradio/blocks.py", line 1302, in process_api
    result = await self.call_function(
  File "/home/ygh/.local/lib/python3.10/site-packages/gradio/blocks.py", line 1039, in call_function
    prediction = await anyio.to_thread.run_sync(
  File "/home/ygh/.local/lib/python3.10/site-packages/anyio/to_thread.py", line 31, in run_sync
    return await get_asynclib().run_sync_in_worker_thread(
  File "/home/ygh/.local/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 937, in run_sync_in_worker_thread
    return await future
  File "/home/ygh/.local/lib/python3.10/site-packages/anyio/_backends/_asyncio.py", line 867, in run
    result = context.run(func, *args)
  File "/home/ygh/.local/lib/python3.10/site-packages/gradio/utils.py", line 491, in async_iteration
    return next(iterator)
  File "/mnt/d/ChuanhuChatGPT/modules/utils.py", line 38, in predict
    iter = current_model.predict(*args)
AttributeError: 'NoneType' object has no attribute 'predict'
Loading checkpoint shards: 100%|████████████████████████████████████████████████████████| 33/33 [00:09<00:00,  3.51it/s]
2023-04-14 10:55:54,471 [ERROR] [models.py:545] Tokenizer class LLaMATokenizer does not exist or is not currently imported.

运行环境

• OS: Windows WSL Ubuntu-22.04
• Browser: Chrome
• Gradio version: 3.26.0
• Python version: 3.10.6

帮助解决

• 我愿意协助解决！

补充说明

感觉可能和这个issue有关，尝试了pip install git+https://github.com/huggingface/transformers，但是问题仍然存在

[GaiZhenbiao]

请查看transformers的这个issue：huggingface/transformers#22222 (comment)
host llama模型的那个仓库的配置文件有点问题……

[yinguohang]

搞定了！
手动修改~/.cache/huggingface/hub/models--decapoda-research--llama-7b-hf/snapshots/5f98eefcc80e437ef68d457ad7bf167c2c6/tokenizer_config.json中的 LLaMATokenizer 到 LlamaTokenizer