ryderwishart
commented
1 year ago Update (selecting the trained model, same two inputs as above):
Sequence ids tensor([[212272, 212273]])
[MAT 1:1_a] [MAT 1:1_b] Kubi na awa zazin nkifo nanyan dutu kitin Yesu ku, nwa ti lau kilari nlire kuculu, inan leo iba dak [MAT 10:41_c] [MAT 17:21_c] [1CO 2:1_c] [ACT 21:39_b] [ROM 8:25_c] [1CO 6:9_e] Yene inan ni umong di nin salin lanzun nanite nene sa unuzu bara inung uwa ni wui ba. Meng na nwui ma yitu nin cikilUsing seven magic tokens as input:
Sequence ids tensor([[265143, 265144, 265145, 265146, 265147, 265148, 265149]])
[REV 1:1_a] [REV 1:1_b] [REV 1:1_c] [REV 1:1_d] [REV 1:1_e] [REV 1:1_f] [REV 1:1_g] Nanere wang din sesu umon imonli ulau me wang, na anari wa su kuru nin lisa usuu nin kubi kongo na umon wa punghe ba, na nan nanere ba unan nakara ba, bara na ima kewu ku unan kotu kuwucuwucue, ukuse nani, ukuse na ibaSummary
Ground truth:
Ulelere upunu mbeleng Yesu Kristi na Kutelle wa nighe, a duro achin me imon ncin dak nan nya nayiri baat. A taa iyinno inin nan nyan ntuu ngono kadura me udu nkiti kuchin mye Yohanna.Prediction:
Nanere wang din sesu umon imonli ulau me wang, na anari wa su kuru nin lisa usuu nin kubi kongo na umon wa punghe ba, na nan nanere ba unan nakara ba, bara na ima kewu ku unan kotu kuwucuwucue, ukuse nani, ukuse na iba
@JEdward7777
Using 7 magic tokens, here's some sample output: In:
generate_text( "[MAT 1:1_a][MAT 1:1_b]", 100, os.path.join( trained_model_folder, selected_model_dropdown.value ) )Out:
In:
generate_text( "[REV 1:1_a][REV 1:1_b][REV 1:1_c][REV 1:1_d][REV 1:1_e][REV 1:1_f][REV 1:1_g]", 100, os.path.join( trained_model_folder, selected_model_dropdown.value ) )Out: