Saran

saran1999's activity

• All
• Models
• Datasets
• Spaces
• Papers
• Collections
• Community
• Posts
• Upvotes
• Likes
• Articles

New activity in answerdotai/ModernBERT-base about 1 year ago

Loss = 0 and Gradient = NaN in ModernBERT Fine-Tuning for Regression

7

#63 opened about 1 year ago by

saran1999

nan or 0.0 loss when training with flash attention

16

#59 opened about 1 year ago by

roadtoagi

Loss = 0 and Gradient = NaN in ModernBERT Fine-Tuning for Regression

7

#63 opened about 1 year ago by

saran1999

nan or 0.0 loss when training with flash attention

16

#59 opened about 1 year ago by

roadtoagi

nan or 0.0 loss when training with flash attention

16

#59 opened about 1 year ago by

roadtoagi