petals/requirements.txt

torch>=1.12
bitsandbytes==0.34.0
accelerate==0.10.0
huggingface-hub==0.7.0
transformers==4.21.3
protobuf>=3.20.3,<4.0dev
git+https://github.com/learning-at-home/hivemind@be88b4280cdd87432168e1da238e532f1364078b
humanfriendly
async-timeout>=4.0.2
Quality of life changes: update readme, simplify run_server interface (#75) - run_server now accepts model name as both positional and keyword argument - changed names in README to account for interface updates - moved model conversion from README to a separate wiki page - updated requirements.txt 2022-09-20 00:51:57 +00:00			`torch>=1.12`
Use bitsandbytes==0.34.0, update readme (#76) * unlock bnb backward * Fix bnb version in README * Update requirements.txt 2022-09-20 10:07:34 +00:00			`bitsandbytes==0.34.0`
Add automated tests (#23) This PR will run basic tests automatically on each subsequent PR - convert a small model on every PR - run existing tests on every PR - enforce black / isort - require checks on merge - make sure tests are not flappy Co-authored-by: Alexander Borzunov <hxrussia@gmail.com> Co-authored-by: Dmitry Baranchuk <dmitrybaranchuk@gmail.com> 2022-07-15 22:59:23 +00:00			`accelerate==0.10.0`
			`huggingface-hub==0.7.0`
Update dependency versions (#71) * update dependency versions * install bitsandbytes cpuonly from pip * remove deprecated API from task pool * clearer startup logs Co-authored-by: Tim Dettmers <dettmers@cs.washington.edu> 2022-09-13 00:51:15 +00:00			`transformers==4.21.3`
Try to fix protobuf versions once again (#95) The goals of these changes are: - Make Petals work in Colab right after just doing `pip install -r requirements.txt` - Make tests work independently of the protobuf package version chosen while installing dependencies 2022-11-28 08:18:03 +00:00			`protobuf>=3.20.3,<4.0dev`
Improve server's logging (#96) Log all RPC calls with block indices and shortened peer IDs, print attention cache stats. 2022-11-29 04:45:50 +00:00			`git+https://github.com/learning-at-home/hivemind@be88b4280cdd87432168e1da238e532f1364078b`
Let users specify sequence length instead of assuming 2048 (#52) - Maximum length is now provided in `.inference_session(max_length=100)` - previously, we would always assume max length = 2048 - added a generic way to forward *kwargs to inference session - for compatibility with #47 - Note to @borzunov : it does not* pass them arbitrarily, but instead checks for kwarg names at the bottom level - run_server can be started with a custom max_length for inference - renamed --cache_size_bytes to --attention_cache_bytes (to avoid collision with --cache_dir) - --attn_cache_bytes can now support humane file sizes (e.g. 300MB instead of 314572800) - made some server-side errors more human-readable to user (e.g. when max length is exceeded) Co-authored-by: Aleksandr Borzunov <borzunov.alexander@gmail.com> Co-authored-by: Alexander Borzunov <hxrussia@gmail.com> 2022-08-29 18:04:37 +00:00			`humanfriendly`
Add various server timeouts, lower --max_batch_size and --inference_max_length defaults (#97) 2022-11-29 06:00:47 +00:00			`async-timeout>=4.0.2`