[Example] ggml: add llava-base64-stream example

[dm4]

• Merge after [WASI-NN] ggml: add inline base64 prompt support for llava  WasmEdge/WasmEdge#3248

[juntao]

Hello, I am a code review bot on flows.network. Here are my reviews of code commits in this PR.

---

Potential Issues and Errors:

1. The inclusion of large base64 image data in the source file could increase its size unnecessarily, suggesting a more efficient alternative like loading from an external resource.
2. User input requirements during program execution may hinder automation and batch processing.
3. Insufficient error handling and informative messages could impact user experience.
4. Dependencies on environment variables for configuring options might restrict deployment flexibility.
5. Lack of proper unit tests and error handling for unexpected scenarios.

Most Important Findings:

1. Redundant commands across multiple job runs could be consolidated to improve efficiency and maintainability.
2. Incomplete setup and preloading steps in the "Build llava-base64-stream" job may lead to execution issues.
3. Unspecified value for the $NGL environment variable could cause unexpected behavior during execution.

Overall Summary:

The Pull Request introduces a new example to the wasmedge-ggml project, encompassing code additions, configuration updates, and workflow enhancements. While the changes aim to streamline workflows and enhance functionality, several potential issues and errors have been identified. Addressing these concerns, such as optimizing data handling, improving user interaction, enhancing error handling, and refining job configurations, is crucial for ensuring the effectiveness and reliability of the new example. Additionally, consolidating redundant commands and ensuring the completeness of job configurations are essential steps towards achieving a cohesive and error-free contribution. Further refinement and attention to detail are necessary to align the changes with best practices and user expectations.

Details

Commit 258acdcca66e78c6323e4f53376a4cb6481fd32f

Key Changes:

1. Added a new example named "llava-base64-stream" to the wasmedge-ggml project.
2. Added a Cargo.toml file with dependencies and project information.
3. Added a README.md with instructions and details about the new example.
4. Created a new Rust source file, main.rs, implementing functionality related to base64 encoding and stream processing.
5. Included functions for reading input, setting options from environment variables, setting data to context, and getting data from the context.

Potential Problems:

1. The source code seems to include a large base64 encoded image data in the image_base64_bytes variable. This may bloat the source file and could be handled differently (e.g., loading from a separate resource).
2. The program requires user input during execution, which may not be suitable for all deployment scenarios (e.g., automated testing, batch processing).
3. Error handling in the code could be enhanced to provide more informative messages to users in case of failures.
4. The use of environment variables for setting options may limit the flexibility of the program in certain environments or deployment setups.
5. Lack of unit tests or error handling for scenarios where unexpected input or errors may occur.

Commit 1d15aa3fd951984e7922e27417b08ae6da648524

Key Changes:

1. Merged the "m1" job into the matrix for different runners.
2. Added a new job named "Build llava-base64-stream" under the "llava" workflow.
3. Updated existing job configurations to include the necessary environment variables.

Potential Problems:

1. The "test -f ~/.wasmedge/env && source ~/.wasmedge/env" command is repeated in multiple job runs. It might be more efficient to extract this into a separate step to avoid redundancy.
2. In the new job "Build llava-base64-stream," the script appears to be incomplete. It lacks the necessary setup and preloading steps that are present in other job configurations.
3. The environment variable $NGL is used but its value is not specified or set anywhere in the provided context, which could lead to issues during execution.

Overall, the changes seem to aim at streamlining the existing workflows and consolidating job configurations, but some aspects need further refinement for completeness and clarity.