ly17/sa2va-26b-hard-v14

Sa2VA 26B Hard SFT v14

Final checkpoint of Sa2VA 26B model trained with hard sample mining strategy (v14).

Model Details

• Model Type: Sa2VA (Segment Anything 2 Vision Assistant)
• Base Model: InternLM2-26B
• Training Strategy: Hard Sample Mining SFT
• Resolution: 512x512
• Task: Medical vessel segmentation (coronary arteries)

Usage

from transformers import AutoModel, AutoTokenizer

model = AutoModel.from_pretrained("ly17/sa2va-26b-hard-v14", trust_remote_code=True)
tokenizer = AutoTokenizer.from_pretrained("ly17/sa2va-26b-hard-v14", trust_remote_code=True)

Performance

This model achieved strong performance on vessel segmentation tasks with improved handling of difficult cases through hard sample mining.