[Docs] ggml: add README for the llava-1.6 execution command

[dm4]

No description provided.

[juntao]

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

---

Commit c8de85cf96a5f29a671c537d3d094e1deba6878b

Key Changes:

1. Added a new section "Execute (llava-1.6)" in the README.md file under the llava directory.
2. Provided instructions on how to execute the WASM with the wasmedge using the named model feature to preload a large model for the llava-1.6 model.
3. Included a snippet with example commands for executing the llava-1.6 model with specific parameters.

Potential Problems:

1. Incomplete User and Assistant Dialogue: The additional user and assistant dialogue included in the patch might be unnecessary for a README file. It seems like these conversations are AI-generated responses related to image recognition, which may not be directly relevant to the documentation. Ensure the relevance of content added to the README.

2. Context Size Discrepancy: In the new llava-1.6 execution part, the note mentions setting the context size to at least 4096, but in the provided commands, the parameter is named ctx_size (underscore instead of dash). This inconsistency should be checked and corrected for clarity and correctness.

3. Consistency with Existing Formatting: Verify if the new section aligns with the existing formatting and style of the README file to maintain consistency across the document.

4. Clarity in Instructions: Review whether the instructions provided for executing the llava-1.6 model are clear and understandable. Ensure that users can easily follow the steps and configurations mentioned without ambiguity.

5. Testing Required: It would be beneficial to test the new llava-1.6 execution instructions to ensure accuracy and effectiveness before finalizing and merging the changes.

Overall, while the patch adds valuable content for executing the llava-1.6 model, attention should be given to the completeness, correctness, and relevance of the added information in the README file.