Some files are missing

[Xnhyacinth]

Excellent work! but I noticed that certain files might have been omitted during evaluation, for example, the following eval_livecc, eval_streaming_qwen, get_phrase_before_timestamp.

streaming-vlm/streaming_vlm/eval/model_compete/generate_segments.py

Line 3 in a056760

from streaming_vlm.eval.model_compete.eval_livecc import eval_livecc

[xrorrim]

Apologize for the inconvenience and thank you for your kind remind! The latest commit has fixed it.

[RL09999]

Seems files in from streaming_vlm.data.lmm_dataset import LMMDataset are missing too. ---- If I'm not wrong, there's missing folder called data under /streaming_vlm/streaming_vlm/ while streaming_vlm.inference requires it.

[xrorrim]

Fixed in latest commit!

[RL09999]

Bug Report: streaming_inference() does not return token_decoded_num required by efficiency_test.py

In efficiency_test.py, each chunk run expects streaming_inference() to return both time_result and token_decoded_num. However, the current implementation of streaming_inference() does not return token_decoded_num, which leads to a TypeError when trying to extend a list with an int.

Relevant code in efficiency_test.py:

time_result, token_decoded_num = streaming_inference(...)
token_decoded_nums.extend(token_decoded_num)  # TypeError: 'int' object is not iterable