tianxiaoban

tianxiaoban's activity

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

New activity in Ex0bit/Qwen3.6-27B-PRISM-PRO-DQ 2 months ago

Why is using native MTP speculative decoding slower than the no-spec baseline?

7

#4 opened 2 months ago by

tianxiaoban

Why is using native MTP speculative decoding slower than the no-spec baseline?

7

#4 opened 2 months ago by

tianxiaoban

Why is using native MTP speculative decoding slower than the no-spec baseline?

7

#4 opened 2 months ago by

tianxiaoban