AdaMLLab/smolkalam-arabic-conversational-sft download history

AdaMLLab/smolkalam-arabic-conversational-sft is a text generation dataset on the Hugging Face Hub. In the last 30 days it was downloaded 942 times (231 in the last 7 days), and 2,482 times in total. It ranks #20,209 among datasets by monthly downloads.

SmolKalam SmolKalam is a quality-filtered Arabic SFT dataset of 1,790,478 examples (~2.45B tokens), built as an ensemble translation of SmolTalk2. It covers multi-turn dialogue (23% of rows), reasoning traces (19% carry <think>), tool and function calling (4.4%), and long context, cat

Models trained on smolkalam-arabic-conversational-sft

1 models list it as training data.

• mastouri/ALLaM-7B-Arabic-SFT – downloads in 30 days

Open AdaMLLab/smolkalam-arabic-conversational-sft on Hugging Face

Sister project: Paper Pulse, the upvote history of every Hugging Face Daily Paper.