Hi, Since the precompiled version doesn't work as mentioned in https://github.com/cgisky1980/ai00_rwkv_server/issues/15, I compiled the source code and run it, but it crashed after I submit any text in web user interface. Here is the messages in the console:

$ cargo run --release -- --model assets/models/RWKV-4-World-0.4B-v1-20230529-ctx4096.st 
    Finished release [optimized] target(s) in 1.53s
     Running `target/release/ai00_server --model assets/models/RWKV-4-World-0.4B-v1-20230529-ctx4096.st`
2023-08-01T01:24:06.717Z WARN  [wgpu_core::instance] Missing downlevel flags: DownlevelFlags(SURFACE_VIEW_FORMATS)
The underlying API or device in use does not support enough features to be a fully compliant implementation of WebGPU. A subset of the features can still be used. If you are running this program on native and not in a browser and wish to limit the features you use to the supported subset, call Adapter::downlevel_properties or Device::downlevel_properties to get a listing of the features the current platform supports.
WARNING: lavapipe is not a conformant vulkan implementation, testing use only.
2023-08-01T01:24:06.768Z INFO  [ai00_server] AdapterInfo {
    name: "llvmpipe (LLVM 12.0.0, 256 bits)",
    vendor: 65541,
    device: 0,
    device_type: Cpu,
    driver: "llvmpipe",
    driver_info: "Mesa 21.2.6 (LLVM 12.0.0)",
    backend: Vulkan,
}
2023-08-01T01:24:08.293Z INFO  [ai00_server] ModelInfo {
    num_layers: 24,
    num_emb: 1024,
    num_vocab: 65536,
}
2023-08-01T01:24:08.554Z INFO  [ai00_server] server started at http://0.0.0.0:65530
2023-08-01T01:24:21.065Z TRACE [ai00_server] Sampler {
    top_p: 0.5,
    temperature: 1.0,
    presence_penalty: 0.3,
    frequency_penalty: 0.3,
}
2023-08-01T01:24:21.065Z TRACE [ai00_server] state cache miss
2023-08-01T01:24:21.065Z TRACE [ai00_server] User: 现在的时间是2023 8月 1日 星期二 早上

Assistant: 好的我知道了！

User: 你是谁？

Assistant: Hello, I am your AI assistant. If you have any questions or instructions, please let me know!

User: test

Assistant:
Segmentation fault

1. 建议使用标准现代汉语“愉快地聊天吧！”
2. 此处可使用/models接口获取真实模型名称
3. 此处也可标准化成"Max Tokens"、"Top P"、“Presence Penalty”以及"Frequency Penalty"，体现专业性

Hi, I tried the compiled version and run ./ai00_server --model assets/models/RWKV-4-World-0.4B-v1-20230529-ctx4096.st but I get following error message:

$ ./ai00_server --model assets/models/RWKV-4-World-0.4B-v1-20230529-ctx4096.st
./ai00_server: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.33' not found (required by ./ai00_server)
./ai00_server: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.32' not found (required by ./ai00_server)
./ai00_server: /lib/x86_64-linux-gnu/libc.so.6: version `GLIBC_2.34' not found (required by ./ai00_server)

2023-08-18T04:24:01.957Z WARN [wgpu::backend::direct] Shader translation error for stage ShaderStages(COMPUTE): HLSL: Unimplemented("write_expr_math Unpack4x8unorm") 2023-08-18T04:24:01.958Z WARN [wgpu::backend::direct] Please report it to https://github.com/gfx-rs/naga 2023-08-18T04:24:01.958Z ERROR [wgpu::backend::direct] Handling wgpu errors as fatal by default thread 'main' panicked at 'wgpu error: Validation Error

Caused by: In Device::create_compute_pipeline note: label = matmul Internal error: HLSL: Unimplemented("write_expr_math Unpack4x8unorm")

', C:\Users\runneradmin.cargo\registry\src\index.crates.io-6f17d22bba15001f\wgpu-0.16.3\src\backend\direct.rs:3019:5 note: run with RUST_BACKTRACE=1 environment variable to display a backtrace

以下为openai给出的解释： 根据您提供的错误消息，看起来是在使用wgpu库进行计算着色器（Compute Shader）编译时出现了错误。

错误消息中指出了一个HLSL（High-Level Shading Language）编译错误，具体是关于未实现的功能："Unimplemented("write_expr_math Unpack4x8unorm")"。这可能表示您的计算着色器中使用了HLSL中尚未实现的操作或函数。

解决此问题的步骤如下：

1. 报告问题：根据错误消息中的提示，您可以将此问题报告给wgpu库的开发者，以便他们了解到该功能尚未实现，并可能提供修复或解决方案。您可以访问https://github.com/gfx-rs/naga并提交一个新的issue。

2. 检查计算着色器代码：检查您的计算着色器代码，特别关注使用了"Unpack4x8unorm"操作或函数的地方。如果可能，尝试使用其他可用的操作或函数替代。

3. 版本更新：确保您正在使用wgpu库的最新版本。可能已经有人报告了这个问题，并且在更新的版本中可能已经得到修复。

4. 回溯信息：根据错误消息中的提示，您可以设置环境变量RUST_BACKTRACE为1，以显示完整的回溯信息。这可能会提供更多关于错误发生位置的信息，帮助您进行故障排除。

我已经是最新驱动了，请问该怎么解决😔

For some low performance devices, using GPTQ for 4Bit quantization is very important.

Do you have any plans to add GPTQ quantification support?

Here is a GPTQ quantification project that can be run on a computer: https://github.com/3outeille/GPTQ-for-RWKV

This project seems to use Vulkan to provide hardware acceleration, which means it can be easily ported to mobile platforms. Do you currently support mobile platforms, or do you have plans to support them in the future?

增加文本转换API，把常见的docx、pdf等文档转化为txt， 方便后续 对文档对话、个人知识库等功能的开发。

可以说一下 还有那些文档格式的支持， 欢迎Pr

Add a text conversion API to convert docx and PDF documents into txt,

Facilitate the development of functions such as document chat and personal knowledge base in the future.

Needs Pr