Aug 10 00:40 Delete states 00:33 V6.7: upload scripts/test_bbpe_refit_serial.py (10.3KB) — FASE1+FASE2 training results 00:33 V6.7: upload scripts/fix_bbpe_refit_oom.py (19.6KB) — FASE1+FASE2 training results 00:33 V6.7: upload scripts/run_v67_fase2_only.py (8.2KB) — FASE1+FASE2 training results 00:33 V6.7: upload scripts/run_v67_train_quick.py (15.7KB) — FASE1+FASE2 training results 00:33 V6.7: upload src/bigru_t/utils/oom_guard.py (21.1KB) — FASE1+FASE2 training results 00:33 V6.7: upload src/bigru_t/data/streaming_datasets.py (37.9KB) — FASE1+FASE2 training results 00:33 V6.7: upload reports/v67_fase2_metrics.json (10.0KB) — FASE1+FASE2 training results 00:33 V6.7: upload states/v6_5_v2_conhecimento_partial_d1.pt (1.9KB) — FASE1+FASE2 training results Aug 9 23:54 Delete reports/archive 23:33 V6.7: scrub literal HF_TOKEN from upload script (security) 23:33 V6.7: archive old reports (v6_4, v6_5_7ds_attn v1/v2/v3) 23:33 V6.7: deprecated → scripts/deprecated/upload_v6_5_7ds_attn_v3.py 23:33 V6.7: deprecated → scripts/deprecated/upload_v6_5_7ds_attn_v2.py 23:33 V6.7: deprecated → scripts/deprecated/upload_v6_5_7ds_attn.py 23:33 V6.7: deprecated → scripts/deprecated/upload_v6_5_7ds.py 23:33 V6.7: deprecated → scripts/deprecated/train_v6_5_7ds_attn_v3.py 23:33 V6.7: deprecated → scripts/deprecated/train_v6_5_7ds_attn_v2.py 23:33 V6.7: deprecated → scripts/deprecated/train_v6_5_7ds_attn.py 23:33 V6.7: deprecated → scripts/deprecated/train_v6_5_7ds.py 23:33 V6.7: deprecated → scripts/deprecated/train_v6_4.py 23:33 V6.7: upload scripts/som_auto_adjust_runner.py (som_auto_adjust_runner + train script V6.7) 23:33 V6.7: upload scripts/train_v6_5_v2.py (som_auto_adjust_runner + train script V6.7) 23:32 V6.7: upload src/bigru_t/model/kohonen_learning_system.py (BBPE serial mode + OomGuard V7) 23:32 V6.7: upload src/bigru_t/tokenizer/bbpe_tokenizer.py (BBPE serial mode + OomGuard V7) 23:32 Delete v6_5_v2_user_questions.json with huggingface_hub 23:32 Delete v6_5_v2_report.json with huggingface_hub 23:32 Delete v6_5_v2_predict_fix_eval.json with huggingface_hub 23:32 Delete v6_5_v2_phases_eval.json with huggingface_hub 23:32 Delete v6_5_v2_attention_eval.json with huggingface_hub 22:26 Delete v6_5_v2_model_states.pt 21:43 V6.5-V2-dynamic: upload batch (scripts + state + model) [80 files] 21:03 Delete states 20:30 V7 FINAL: FASE1+FASE2 completed (10000 samples, 20 punishments) V7 fixes:
- AdvancedSOMAugmenter.fit(): torch.no_grad() wrapper (OOM fix + .item() fix)
- V7Fase2Integrator.post_chunk_hook: defensive int/float, traceback logging
- V7Fase2Integrator._compute_adaptive_dpo: kls.neuron_label fix (was kls.som.neuron_labels)
- SimpleBBPETokenizer: pure alias to BBPETokenizer (no duplicate impl)
- train_v6_5_v2.py: SKIP_FASE1 support for FASE2 resume
Results:
- FASE1: 8000 samples (8 datasets, Kohonen only)
- FASE2: 2000 samples (BrunoN-Dev/corpus-ptbr-v1, punishment active)
- V7 augmenter: 10 fit() calls, 0 errors
- V7 DPO: 20 triggers, 0 errors
- SOM: dead 92.6%->56%, TE 0.87->0.39, VE 0.98
Cleanup:
- Deleted 1 misplaced/old files from HF root
- Uploaded 18 files to correct paths 19:08 V7: dpo.py + data_augmentation.py + oom_guard.py aprimorados Módulos aprimorados:
- training/dpo.py: AdaptiveDPOLoss (nn.Module, log_beta otimizável, atualização dual KL-aware, clamps [0.01, 2.0])
- training/data_augmentation.py: AdvancedSOMAugmenter (manifold-aware, grid 4D canônico (4,4,4,4)=256, α₀=0.5, σ₀=2.0, ruído gaussiano adaptativo por QE local, interpolação convexa 0.7*real + 0.3*BMU)
- utils/oom_guard.py: OomGuard V7 (memory leak detection, logic failure detection, crash dump JSON, signal handlers, excepthook, ring buffer de ops, cgroup limit detection)
Arquivos removidos (fora de src/bigru_t/):
- 5 JSONs antigos na raiz: v6_5_v2_*.json (substituídos por reports/)
Parâmetros canônicos mantidos:
- HIDDEN_DIM=1024, VOCAB_SIZE=16384, N_HYPOTHESES=16, MAX_N_HYPOTHESES=32, HYP_TRAIN_STEPS=30, HYP_HIDDEN_DIM=256
- Grid SOM (4,4,4,4)=256, Buffer=256 (alinhamento 1:1)
- α₀=0.5 (∈ [0.5, 1.0] Kohonen rough training)
- σ₀=2.0 (= max(4,4,4,4)/2 = metade da maior dimensão)
Timestamp: 2026-08-09T19:08:19.776243 18:18 V6.5-V4-canonical-256: buffer=256 + grid=(4,4,4,4)=256 + BBPE serial fix Mudanças principais:
- Buffer canônico 256 (alinhado ao grid 256, elimina OOM de V6.5-V3)
- BBPE serial in-process para corpus < 200 textos (evita OOM por ProcessPoolExecutor)
- FASE1: 7000 samples processados (7/8 datasets)
- FASE2: 2000 samples com punição ativa (meta atingida)
- 23 punishment events, 11 hypotheses trainings, 11 delta applications
- Peak RSS: 1933MB (dentro do limite 4GB cgroup)
- Paths remapeados: TODOS módulos em src/bigru_t/**
- Sem duplicidade: deprecated/ NÃO enviados
Timestamp: 2026-08-09T18:18:20.529072 17:49 V6.5-V3-hyp-synergy: SynergyHypothesisEnsemble (4 pilares: diversidade k-means++, cooperação, competição, ortogonalidade) + testes de validação (7 files) 17:31 Delete model_states 17:31 Delete v6_5_v2_model_states.pt 17:24 V6.5-V2-parallel-reorganize: ParallelSOM4D + ReorganizingSOM4D + OOM-killer guards + unified state (75 files) 17:22 V6.5-V2-dynamic: upload batch (scripts + state + model) [79 files] 16:55 Delete v6_5_v2_model_states.pt 16:50 V6.5-V2-dynamic: upload batch (scripts + state + model) [79 files] 16:08 Delete model_final 16:07 Delete model_final/config.json 16:06 Delete model_states/v6_5_v2_model_states.pt 15:50 V6.5-V2-parallel-reorganize: ParallelSOM4D + ReorganizingSOM4D + OOM-killer guards + unified state (75 files) 15:23 Delete v6_5_v2_model_states.pt 14:25 V6.5-V2-dynamic: upload batch (scripts + state + model) [79 files] 13:47 V6.5-V2-unified: cleanup misplaced root files + old scripts [18 files deleted] — removes duplicates from repo root (they live in src/bigru_t/...) and outdated scripts in scripts/ (now in deprecados/) 
