Ba2han's picture
Update README.md
45870a2 verified
|
raw
history blame
708 Bytes
metadata
license: mit
language:
  - tr
library_name: transformers

Pretrained on roughly 1.6B (mostly Turkish) tokens from HF and "high quality" scraped data using 1 RTX 3090. The training will continue. The model already can be (sort of) fine-tuned for instruction.


HF kaynaklı ve scrape edilen yaklaşık 1.6 Milyar (çoğunlukla Türkçe) token ile 1 RTX 3090 kullanılarak eğitilmiştir. Model şimdiden talimatlar için fine-tune edilebiliyor:

image/png

max_length=256, top_k=20, min_p=0.1, repetition_penalty=1.1, temperature=0.1, seed=22366 / TR_4k_LoRA