Merged
Conversation
Implemented high-throughput vLLM inference backend with OpenAI-compatible API support. Features: - VLLMBackend class with full BaseBackend interface implementation - VLLMBackendConfig for server connection configuration - Native function/tool calling support with automatic fallback - Comprehensive test suite (16 test cases) - Helper script for starting vLLM server - Complete documentation with examples and troubleshooting Files added: - python/backends/vllm_backend.py: Main backend implementation - python/run_vllm_server.py: vLLM server startup script - tests/test_vllm_backend.py: Full test coverage - docs/vllm_backend.md: Usage guide and documentation The backend supports: - Text generation with customizable parameters - OpenAI-style function calling - Multiple model architectures (Llama, Mistral, Qwen) - GPU acceleration and tensor parallelism - Automatic connection health checks Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
Set up and tested llama.cpp backend for local Windows development with CPU inference. Features: - Export BackendConfig from backends module for easier imports - Comprehensive test script with 4 test scenarios - Complete Windows setup documentation - Verified working with Llama-2-7B-Chat Q4_K_M model Files added/modified: - python/backends/__init__.py: Export BackendConfig - python/test_llama_backend.py: Test script for llama.cpp backend - docs/llama_cpp_windows_setup.md: Complete setup guide Test Results: - Basic text generation: Working (9 tokens/sec) - Tool calling: Working with JSON parsing - Temperature variations: Working (0.1, 0.7, 1.0) - Multi-turn conversations: Working The llama.cpp backend provides: - CPU-only inference (no CUDA required) - Low memory usage with quantized models - Fast local development on Windows - Compatible with GGUF format models Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
Added infrastructure for GPU-accelerated inference with CUDA support. Features: - Extended BackendConfig with n_gpu_layers parameter - Updated llama_cpp_backend to support GPU offloading - GPU test script to compare CPU vs GPU performance - Comprehensive GPU setup documentation Changes: - python/backends/base.py: Added n_gpu_layers parameter (0=CPU, -1=all GPU) - python/backends/llama_cpp_backend.py: Implemented GPU layer offloading logic - python/test_llama_gpu.py: Performance comparison script - docs/llama_cpp_gpu_setup.md: Complete GPU setup guide GPU Configuration: - n_gpu_layers=0: CPU only (current default) - n_gpu_layers=20: Hybrid CPU/GPU (20 layers on GPU) - n_gpu_layers=-1: Full GPU offload (recommended for RTX 3090) Expected Performance: - RTX 3090: ~100+ tokens/sec (10-15x speedup over CPU) - Current CPU: ~9 tokens/sec baseline Note: Requires CUDA Toolkit installation for GPU acceleration. See docs/llama_cpp_gpu_setup.md for setup instructions. Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude <noreply@anthropic.com>
…ols to let the LLM make observations and execute actions
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
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.
Description
Brief description of what this PR does.
works on #6
Type of Change
Changes Made
Testing
pytest tests/)Manual Testing:
Describe how you tested this change manually.
Performance Impact
Documentation
Checklist
Screenshots (if applicable)
Add screenshots to help reviewers understand your changes.
Additional Notes
Any additional information reviewers should know.