Pedro21613 commited on
Commit
cdba320
Β·
verified Β·
1 Parent(s): f117384

Upload TRAINING_STATUS.md with huggingface_hub

Browse files
Files changed (1) hide show
  1. TRAINING_STATUS.md +1 -1
TRAINING_STATUS.md CHANGED
@@ -2,7 +2,6 @@
2
 
3
  Kernel: https://www.kaggle.com/code/pedroberbisfreire/ps-design-2-0-treino-50k-2x-gpu
4
 
5
- - 09-27 19:19 UTC β€” dados prontos: train=57128 eval=498 (32 min de kernel)
6
  - 09-27 19:19 UTC β€” preflight do collator OK (batch misto curto+longo validado)
7
  - 09-27 19:20 UTC β€” smoke DDP OK (env extra={}) em 33 min β€” iniciando treino completo
8
  - 09-27 19:20 UTC β€” πŸ‹οΈ treino completo iniciado β€” 1786 steps planejados (2 GPUs)
@@ -47,3 +46,4 @@ Kernel: https://www.kaggle.com/code/pedroberbisfreire/ps-design-2-0-treino-50k-2
47
  - 09-27 22:43 UTC β€” step 1450/1786 | loss=0.17166428565979003 | elapsed=3.38h | ETA=0.78h
48
  - 09-27 22:50 UTC β€” step 1500/1786 | loss=0.16980835199356079 | elapsed=3.50h | ETA=0.67h
49
  - 09-27 22:51 UTC β€” step 1500/1786 | loss=? | elapsed=3.51h | ETA=0.67h
 
 
2
 
3
  Kernel: https://www.kaggle.com/code/pedroberbisfreire/ps-design-2-0-treino-50k-2x-gpu
4
 
 
5
  - 09-27 19:19 UTC β€” preflight do collator OK (batch misto curto+longo validado)
6
  - 09-27 19:20 UTC β€” smoke DDP OK (env extra={}) em 33 min β€” iniciando treino completo
7
  - 09-27 19:20 UTC β€” πŸ‹οΈ treino completo iniciado β€” 1786 steps planejados (2 GPUs)
 
46
  - 09-27 22:43 UTC β€” step 1450/1786 | loss=0.17166428565979003 | elapsed=3.38h | ETA=0.78h
47
  - 09-27 22:50 UTC β€” step 1500/1786 | loss=0.16980835199356079 | elapsed=3.50h | ETA=0.67h
48
  - 09-27 22:51 UTC β€” step 1500/1786 | loss=? | elapsed=3.51h | ETA=0.67h
49
+ - 09-27 22:51 UTC β€” πŸ’Ύ checkpoint-1500 salvo (backup do adapter no HF)