Add benchmark to compare multi-stream vs single-stream encoding

[NicolasHug]

Basic benchmark aimed at validating that the new Encoder class isn't slower than the existing VideoEncoder and AudioEncoder.

Results on my laptop: all good.

~/dev/torchcodec-cuda (benchhhhhh*) » python benchmarks/encoders/bench_multi_stream_vs_single_stream.py

==================================================
Duration: 3s
==================================================

--- Video encoding (cpu) ---
VideoEncoder:    med = 328.95ms +- 11.82
StreamingEncoder:med = 331.38ms +- 11.46

--- Video encoding (cuda) ---
VideoEncoder:    med = 266.34ms +- 0.83
StreamingEncoder:med = 266.79ms +- 1.41

--- Audio encoding (cpu) ---
AudioEncoder:    med = 0.85ms +- 0.34
StreamingEncoder:med = 0.75ms +- 0.07

==================================================
Duration: 10s
==================================================

--- Video encoding (cpu) ---
VideoEncoder:    med = 1030.14ms +- 8.88
StreamingEncoder:med = 1034.21ms +- 26.75

--- Video encoding (cuda) ---
VideoEncoder:    med = 266.59ms +- 1.90
StreamingEncoder:med = 266.84ms +- 0.44

--- Audio encoding (cpu) ---
AudioEncoder:    med = 2.07ms +- 0.14
StreamingEncoder:med = 2.13ms +- 0.14

[pytorch-bot]

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/meta-pytorch/torchcodec/1432

• 📄 Preview Python docs built from this PR

Note: Links to docs will display an error until the docs builds have been completed.

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

• Run pull request jobs on OSDC runners in shadow mode

This comment was automatically generated by Dr. CI and updates every 15 minutes.