Commit graph

13 commits

Author SHA1 Message Date
jinli.yl
08b771b6c4 feat(mem-agent): introduce version 4 memory agents and tools 2026-01-18 15:50:46 +08:00
jinli.yl
e6ad682ede feat(core): add ReMe V3 implementation with optional MCP client and enhanced filtering 2026-01-17 01:15:49 +08:00
jinli.yl
327dc58f70 feat(bench): add HaluMem dataset statistics analyzer and update LLM concurrency control
- Add new analyze_dataset_stats.py script for comprehensive HaluMem dataset analysis
- Include statistics for user sessions, dialogues, content lengths and chunk distributions
- Replace rate limiting with concurrency control in BaseLLM using semaphore mechanism
- Update configuration to use max_concurrency instead of max_rps and rps_window
- Modify dialogue formatting to include only user messages in evaluation
- Add percentile calculations and detailed content size distribution metrics
- Implement session splitting logic based on character length thresholds
- Provide per-user statistics and summary tables for dataset analysis
- Refactor BaseLLM to use internal _chat_impl and _stream_chat_impl methods
- Remove rate limiting locks and timestamps from LLM initialization
- Add command-line interface for dataset statistics analysis tool
2026-01-14 10:14:35 +08:00
jinli.yl
52a4b66d59 refactor(bench): update halumem evaluation with concurrent session processing
- Increased max_rps from 9 to 20 in default config
- Removed tqdm progress bar for session processing
- Implemented asyncio semaphore for session concurrency control
- Added parallel processing of sessions within each user
- Updated logging to show session count per user
- Changed user processing from concurrent to sequential
- Added completion tracking with progress indicators
- Modified output formatting for better readability
2026-01-14 00:13:55 +08:00
jinli.yl
f0573a09d9 feat(bench): add progress bar for session processing in halumem evaluation
- Integrated tqdm library for progress tracking
- Added progress bar display for user sessions processing
- Implemented session-by-session progress updates with custom description
- Maintained existing session processing logic while adding visual feedback
- Preserved all original functionality including logging and data handling
2026-01-14 00:02:26 +08:00
jinli.yl
b8124fe31a feat(benchmark): add HaluMem baseline evaluation and analysis tools 2026-01-13 23:59:55 +08:00
jinli.yl
59c0702759 feat(halumem): update memory operations to include agent messages and success flags 2026-01-13 14:02:29 +08:00
jinli.yl
581482eeb0 feat(mem_agent): introduce version 2 memory agents and tools
- Add ReMeRetrieverV2 and ReMeSummarizerV2 components
- Implement new memory tools including AddMemoryDrafts, RetrieveMemories, UpdateMemories
- Create PersonalSummarizerV2 with three-step workflow for memory management
- Add simplified evaluation script for ReMe on HaluMem benchmark
- Update base memory agent with enhanced logging capabilities
- Introduce new prompt formats and evaluation methods for question answering
- Modify LLM utilities and message formatting with markdown header stripping option
2026-01-13 01:54:20 +08:00
jinli.yl
5a0c1309f1 feat(bench): add statistics computation script for benchmark evaluation 2026-01-11 16:17:51 +08:00
jinli.yl
e302ec72cb feat(llm): add rate limiting and model override support to BaseLLM 2026-01-11 11:24:20 +08:00
jinli.yl
2e2d96a53e feat(vector-store): add delete_all method and improve evaluation pipeline 2026-01-10 22:47:56 +08:00
jinli.yl
9f114cb9b4 feat(benchmark): add HaluMem evaluation suite with ReMe integration 2026-01-10 22:11:20 +08:00
jinli.yl
0b7843f557 feat(core): implement memory node tracking and embedding text truncation 2026-01-09 17:55:11 +08:00