Quant it to your favorite precision. Train on it, etc.



Oh boy! Its another copy of schnell with the double blocks from dev! So what's the difference? I didn't change anything besides the blocks required to make the model fast.
It's intended to be trained on top of, quantized and merged. No other layers were touched. No FP8's, No NF4s. Nothing else included. Just model.
Tested at full precision and the text capabilities were mostly intact.
Quant it to your favorite precision. Train on it, etc.
1. Dieses Modell dient nur Lernzwecken. Urheber- und Auslegungsrechte liegen beim Originalautor.
2. Bist du der Originalautor eines Modells, kontaktiere uns bitte zur Authentifizierung über unsere offiziellen Kanäle. Wir schützen die Rechte aller Schöpfer. Hier klicken, um es zu verifizieren.
