 Tobias Lütke
|
eb1b77c8cb
Deploy fine-tuned GRPO model as default query expansion (#67)
|
há 3 meses atrás |
|
Tobi Lutke
|
8572c2fd94
Deploy fine-tuned GRPO model as default for query expansion
|
há 3 meses atrás |
|
Tobi Lutke
|
5ab78d00a2
Add HF Jobs scripts, temporal query examples, and training results
|
há 3 meses atrás |
|
Tobi Lutke
|
354744af53
Finetune 2.0: consolidate and simplify the entire training pipeline
|
há 3 meses atrás |
|
jdvmi00
|
64c6e6c2e3
fix: rename collectionId to collectionName in searchVec for proper filtering (#61)
|
há 3 meses atrás |
|
Freeman Jiang
|
bfb0eebc3e
fix: use sequential embedding on CPU-only systems to avoid race condition (#54)
|
há 3 meses atrás |
|
Copilot
|
053252ca24
Add Windows path utilities with cross-platform test coverage (#51)
|
há 3 meses atrás |
|
sh54
|
ba7391832d
Add org-mode title extraction support (#50)
|
há 3 meses atrás |
|
sh54
|
65c0f89560
Enable SQLite extension loading in devshell (#48)
|
há 3 meses atrás |
|
Tobi Lutke
|
9b3a209a97
Fix GRPO training: apply chat template to prompts
|
há 3 meses atrás |
|
Tobi Lutke
|
3ea85eff50
Make TUI model list dynamic from HuggingFace Hub
|
há 3 meses atrás |
|
Tobi Lutke
|
891f3262cf
Fix GRPO reward function to handle think blocks and end tokens
|
há 3 meses atrás |
|
Tobi Lutke
|
66bb8ed963
Remove beads reference from CLAUDE.md
|
há 3 meses atrás |
|
Tobi Lutke
|
2267986302
Remove beads issue tracking
|
há 3 meses atrás |
|
Tobi Lutke
|
8a1c4cdab0
Add 1.7B and 4B GRPO training and GGUF conversion scripts
|
há 4 meses atrás |
|
Tobi Lutke
|
b9b1b39a76
Update README with separate model repos
|
há 4 meses atrás |
|
Tobi Lutke
|
312c281109
Update README for unified model repository structure
|
há 4 meses atrás |
|
Tobi Lutke
|
2648512b7c
Fix TUI to load GRPO models with SFT base first
|
há 4 meses atrás |
|
Tobi Lutke
|
f96766cce8
Fix GRPO model loading to use SFT base first
|
há 4 meses atrás |
|
Tobi Lutke
|
f6a6716c44
Refactor evals into separate run and score scripts
|
há 4 meses atrás |
|
Tobi Lutke
|
857a85ab58
Clean up evaluation files
|
há 4 meses atrás |
|
Tobi Lutke
|
dc8f5a2335
Strict format validation: every line must be lex:/vec:/hyde:
|
há 4 meses atrás |
|
Tobi Lutke
|
2ad507a86e
Add chat template leakage detection to reward function
|
há 4 meses atrás |
|
Tobi Lutke
|
6062dc769f
Add named entity extraction to GRPO reward function
|
há 4 meses atrás |
|
Tobi Lutke
|
32706a720f
Refactor finetune folder: train/rl scripts with YAML configs
|
há 4 meses atrás |
|
Tobi Lutke
|
d32e13c172
Add HuggingFace login and comprehensive scoring to GRPO v2 training
|
há 4 meses atrás |
|
Tobi Lutke
|
c35dbd6cbd
Add comprehensive scoring system for query expansion
|
há 4 meses atrás |
|
Tobi Lutke
|
994a094546
Update README with final evaluation results
|
há 4 meses atrás |
|
Tobi Lutke
|
0353994e7d
Fix GRPO training script for TRL API compatibility
|
há 4 meses atrás |
|
Tobi Lutke
|
7cca164dd9
Add query expansion model finetuning infrastructure
|
há 4 meses atrás |
