Benjamin-png/swahili-gpt-71m

Upload README.md with huggingface_hub

Upload README.md with huggingface_hub

Upload pretrain_loss_curve.png with huggingface_hub

Update weights: best val 3.858 (iter 120k)

Update status: val 3.86 @ iter 120k

Update weights: best val 3.909 (iter 80k)

Update status: val 3.91 @ iter 80k

Update weights: best val 3.982 (iter 42k)

Update status: val 3.98 @ iter 42k

Add banner to model card

Add banner image

Move manifesto after roadmap; add Colab demo link

Manifesto: tokenizer meaning-vs-efficiency, English-dominance data, open-vs-closed case

Add config.json so Hub counts downloads

Update weights: best checkpoint val 4.064 (iter 36k)

Update status: val 4.06 @ iter 36k

Fix usage: auto-download modeling_kiswahili.py (inference.py)

Fix usage: auto-download modeling_kiswahili.py (README.md)

Add swahili-gpt-71m (early checkpoint, val 4.27)

initial commit