[Usage] Tokenization Mismatch

### Describe the issue

Issue: after pretraining Phi-3 with/without VIP, I am getting tokenization mismatch. 

Command:
```
bash scripts/finetune_vip_llava_phi3_stage2.sh 
```

Log: 
```
WARNING: tokenization mismatch: 185 vs. 195. (ignored)
WARNING: tokenization mismatch: 191 vs. 200. (ignored)
WARNING: tokenization mismatch: 158 vs. 162. (ignored)
```

