WIP: Adding support for Local Models #26
Draft
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
…dling - Add HuggingFaceChatClient for running models from Hugging Face Hub - Introduce LocalModelClient base class for common local model functionality - Update README with detailed local model configuration options - Add huggingface_hub dependency to requirements.txt - Improve environment variable handling for local models
Implement GGUF model discovery functionality that searches for models in a specified directory. When GGUF_MODEL_DIR is set, the system will automatically find and load GGUF models by name without requiring full paths. Update documentation to reflect new GGUF model discovery feature and clarify optional environment variables.
…lization - Move GGUF model search function below imports and conditional imports - Expand AVAILABLE_LLMS list with additional model variants - Rename chat namespace to _openai_compatible_chat for clarity - Simplify Hugging Face model parsing logic and update env var names
Added section comments to group and clarify the different model providers in the AVAILABLE_LLMS list for better readability and maintainability.
Add support for multiple new LLM providers (anthropic, bedrock, vertex_ai, deepseek, openrouter, ollama, localai, llama-cpp) and implement automatic GGUF model discovery with configurable local models directory. Include path detection logic and comprehensive provider-specific model name formatting.
Add support for --model_dir parameter to allow GGUF model discovery in a specified directory when using the 'local' provider. This enables interactive model selection when no specific model is provided and improves error handling for missing models or directories.
The example usage was uncommented and updated with a more specific prompt to demonstrate the LLM client functionality. This helps verify the client setup is working as expected.
Update response content access to use direct 'content' attribute instead of nested structure Add model path construction for llama-2 model
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Doing some of this myself, and as an experiment I'm also giving part of this task to the AI itself. So far it seems a bit sloppy and more like pseudo-code, but I'll give it more of a kick in the coming week(s)