Search results

[Question] Clarification on Inference Strategy for VQA Benchmarks (Standard Sampling vs. Streaming KV Cache)

Status: Open.

#15 In mit-han-lab/streaming-vlm;

· lern-to-write opened on Dec 22, 2025

About training

Status: Open.

#14 In mit-han-lab/streaming-vlm;

· jun0wanan opened on Dec 10, 2025

The result seems that the model is only recalling train episodes, but not explaining the real time frames

Status: Open.

#13 In mit-han-lab/streaming-vlm;

· Enlion91 opened on Dec 9, 2025

Question about the Inf-Stream-Eval

Status: Open.

#12 In mit-han-lab/streaming-vlm;

· xieyc99 opened on Dec 8, 2025

cannot install block_sparse_attn from PyPi

Status: Open.

#11 In mit-han-lab/streaming-vlm;

· Nidhan03 opened on Nov 29, 2025

transformers==4.51.3 is not a correct version

Status: Open.

#10 In mit-han-lab/streaming-vlm;

· Dr-Corgi opened on Nov 26, 2025

Missing High Quality Annealing Data

Status: Open.

#9 In mit-han-lab/streaming-vlm;

· yingift opened on Nov 6, 2025

Missing file eval_all_metadata.jsonl

Status: Open.

#8 In mit-han-lab/streaming-vlm;

· zx386 opened on Oct 27, 2025

demo page is down

Status: Open.

#7 In mit-han-lab/streaming-vlm;

· kabugge opened on Oct 22, 2025

Data Processing Steps

Status: Open.

#6 In mit-han-lab/streaming-vlm;

· xiami2019 opened on Oct 19, 2025

Possibility of running it on the CPU?

Status: Open.

#5 In mit-han-lab/streaming-vlm;

· ramkumarkoppu opened on Oct 14, 2025

How does the model know the names of the various characters in the video?

Status: Open.

#4 In mit-han-lab/streaming-vlm;

· King-king424 opened on Oct 14, 2025