List your company

Fine-tuning on FineWeb (test)

1.93Test LossSISA
1.79922.68213.5654.4479Feb 15, 2025

Evaluation results

MethodLinks
SISAModel=GPT2(335M), Epoch=152025.021.9384.84
AdamWModel=GPT2(335M), Epoch=152025.021.9584.19
SISAModel=GPT2(335M), Epoch=52025.022.0485.24
SISAModel=GPT2(124M), Epoch=152025.022.1784.82
AdamWModel=GPT2(124M), Epoch=152025.022.384.73
SISAModel=GPT2(335M), Epoch=22025.022.3184.76
SISAModel=GPT2(124M), Epoch=52025.022.7384.36
AdamWModel=GPT2(335M), Epoch=52025.022.7384.5
AdamWModel=GPT2(124M), Epoch=52025.023.4584.3
AdamWModel=GPT2(335M), Epoch=22025.023.6283.17
SISAModel=GPT2(124M), Epoch=22025.023.6983.15
AdamWModel=GPT2(124M), Epoch=22025.025.282.16