text-generation-inference/integration-tests
Nicolas Patry 5df8059037
Auto max prefill (#2797)
* Attempt at automatic max batch prefill.

* Taking into account number of shards.

* Adding more cards.

* Adding A100 + H100

* Adding a few more cards.

* Logprobs cost too much.

* h100 better name, and keep factor of 2

* Damn inflated sparse tflops.

* Typo in h100.

* Updated the flops calculation (checked with fvcore).

* chunking by default.

* Fix prefix caching for chat completion since we removed logprobs.

* More tests.

* Dropping all the prefill logprobs.

* Add a flag that enables users to get logprobs back.

* Repairing prompt token counting.

* Fixing a few tests.

* Remove some scaffolding.

* Attempting to reduces the issues (workarounds for now).
2024-12-06 05:52:00 +01:00
..
images Pali gemma modeling (#1895) 2024-05-16 06:58:47 +02:00
models Auto max prefill (#2797) 2024-12-06 05:52:00 +01:00
conftest.py Monkey patching as a desperate measure. (#2704) 2024-10-28 11:25:13 +01:00
poetry.lock Prefix test - Different kind of load test to trigger prefix test bugs. (#2490) 2024-09-11 18:10:40 +02:00
pyproject.toml nix: add black and isort to the closure (#2619) 2024-10-09 11:08:02 +02:00
pytest.ini chore: add pre-commit (#1569) 2024-02-16 11:58:58 +01:00
requirements.txt Prefix test - Different kind of load test to trigger prefix test bugs. (#2490) 2024-09-11 18:10:40 +02:00