15 days ago
-
Card: shape runs change more than one setting (similar parameter count)
-
Add test_wznowienia/B/ckpt_40.pt (full training checkpoint)
-
Add test_wznowienia/A/ckpt_40.pt (full training checkpoint)
-
Add test_wznowienia/A/ckpt_20.pt (full training checkpoint)
-
Card: SHA-256 of resume-test files
-
Add test_wznowienia/mini_R1/ckpt_20.pt (full training checkpoint)
-
Add R3/ckpt_4768.pt (full training checkpoint)
-
Card: finished runs (R1-R3, K22) with results; resume-test folder
-
Add R2/ckpt_4768.pt (full training checkpoint)
-
Add R1/ckpt_4768.pt (full training checkpoint)
-
Add run records (args, metrics, logs, val, environment) for all 400M runs, fala-1 finals K22/R1-R3, resume test, code
-
Card: state exactly which checksums were recorded before the copy
-
Add K22/ckpt_4768.pt (full training checkpoint)
-
Add model card (contents, checkpoint format, SHA-256)
-
Add VR0/ckpt_2000.pt (full training checkpoint)
-
Add VR0/ckpt_1500.pt (full training checkpoint)
-
Add VR0/ckpt_1000.pt (full training checkpoint)
-
Add VR0/ckpt_500.pt (full training checkpoint)
-
Add MU2/ckpt_1500.pt (full training checkpoint)
-
Add MU2/ckpt_1000.pt (full training checkpoint)
-
Add MU2/ckpt_500.pt (full training checkpoint)
-
Add K12/ckpt_2500.pt (full training checkpoint)
-
Add K12/ckpt_2000.pt (full training checkpoint)
-
Add K12/ckpt_1500.pt (full training checkpoint)
-
Add K12/ckpt_1000.pt (full training checkpoint)
-
Add K12/ckpt_500.pt (full training checkpoint)
-
Add GQA/ckpt_2000.pt (full training checkpoint)
-
Add GQA/ckpt_1500.pt (full training checkpoint)
-
Add GQA/ckpt_1000.pt (full training checkpoint)
-
Add GQA/ckpt_500.pt (full training checkpoint)
-
Add MU2/ckpt_2000.pt (full training checkpoint)
-
Add VR0/ckpt_2500.pt (full training checkpoint)
-
Add GQA/ckpt_2500.pt (full training checkpoint)
-
Add K12/ckpt_3000.pt (full training checkpoint)
-
initial commit
