Timmy_Foundation/timmy-config
15
0
Fork 0
Code Issues 201 Pull Requests Actions 25 Packages Projects Releases Wiki Activity

Merge PR #653: training/README.md (changed)

Browse Source
This commit is contained in:
Merge Bot
2026-04-16 05:12:09 +00:00
parent a27061b38b
commit c006d8cf4f
1 changed files with 3 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files

3
training/README.md
View File

@@ -20,6 +20,7 @@ make train-local # LoRA on Apple Silicon (MLX) — free, ~30 min on M3 Max
make train-cloud # QLoRA on cloud GPU (Axolotl) — ~$1/run on A100
make eval # Standard benchmarks via lm-eval-harness against Ollama
make vibes # Hand-picked prompts → human review (the sacred test)
make adversary-value-violations # 200 jailbreak prompts targeting value violations
make ingest # Pull heartbeat trajectories into training data
make curated # Regenerate curated exemplar dataset
make convert # Convert merged data to MLX train/valid format
@@ -46,10 +47,12 @@ training/
├── eval-tasks.yaml ← Benchmark config
├── build_curated.py ← Exemplar data authoring (the soul conversations)
├── ingest_trajectories.py ← Quality filter for heartbeat cycle data
├── run_adversary_eval.py ← Adversarial prompt runner + transcript capture
└── data/
├── curated_dataset.jsonl ← 26 gold-standard conversations (proprietary)
├── preference_pairs.jsonl ← DPO preference pairs (proprietary)
├── prompts_vibes.yaml ← Custom eval prompts
├── prompts_adversary_value_violations.yaml ← 200 jailbreak prompts for value-violation red-team eval
├── prompts_nexus_vibes.yaml ← Nexus-specific eval prompts
└── mlx_curated/ ← MLX-format train/valid splits
```