kjj0/fineweb10B-gpt2 download history
kjj0/fineweb10B-gpt2 is a dataset on the Hugging Face Hub. In the last 30 days it was downloaded 11,418 times (3,647 in the last 7 days), and 89,357 times in total. It ranks #2,681 among datasets by monthly downloads.
fineweb10B-gpt2 This repo contains the GPT-2 tokens for fineweb10B, just as would be generated by https://github.com/KellerJordan/modded-nanogpt/tree/master (or llm.c). You can download from this repo instead of re-tokenizing to save a couple hours of setup on a new machine.
Models trained on fineweb10B-gpt2
26 models list it as training data.
- tim-lawson/skip-middle-fineweb-baseline-12-layers 47 downloads in 30 days
- tim-lawson/skip-middle-fineweb-baseline-8-layers 42 downloads in 30 days
- tim-lawson/skip-middle-fineweb-baseline-6-layers 31 downloads in 30 days
- tim-lawson/skip-middle-fineweb-nocontrol-2-layers 30 downloads in 30 days
- tim-lawson/skip-middle-fineweb-baseline-10-layers 28 downloads in 30 days
- tim-lawson/skip-middle-fineweb-gated-target-end-0.3 26 downloads in 30 days
- tim-lawson/skip-middle-fineweb-nocontrol-10-layers 26 downloads in 30 days
- tim-lawson/skip-middle-fineweb-baseline-2-layers 22 downloads in 30 days
- tim-lawson/skip-middle-fineweb-nocontrol-8-layers 22 downloads in 30 days
- tim-lawson/skip-middle-fineweb-gated-target-end-0.1 21 downloads in 30 days
- tim-lawson/skip-middle-fineweb-baseline-4-layers 21 downloads in 30 days
- tim-lawson/skip-middle-fineweb-gated-target-end-0.7 19 downloads in 30 days
Spaces using fineweb10B-gpt2
- Model Pulse 8 likes