Nick Hill 2efce05dc3

[Fix] Avoid pickling entire LLMEngine for Ray workers (#3207) ...

Co-authored-by: Antoni Baum <antoni.baum@protonmail.com>

2024-03-06 00:17:20 +00:00

..

core

Store eos_token_id in Sequence for easy access (#3166)

2024-03-05 15:35:43 -08:00

engine

[Fix] Avoid pickling entire LLMEngine for Ray workers (#3207)

2024-03-06 00:17:20 +00:00

entrypoints

Push logprob generation to LLMEngine (#3065)

2024-03-04 19:54:06 +00:00

lora

[Neuron] Support inference with transformers-neuronx (#2569)

2024-02-28 09:34:34 -08:00

model_executor

Store eos_token_id in Sequence for easy access (#3166)

2024-03-05 15:35:43 -08:00

transformers_utils

Support starcoder2 architecture (#3089)

2024-02-29 00:51:48 -08:00

worker

[ROCm] enable cupy in order to enable cudagraph mode for AMD GPUs (#3123)

2024-03-04 18:14:53 -08:00

__init__.py

Bump up to v0.3.3 (#3129)

2024-03-01 12:58:06 -08:00

block.py

Add Automatic Prefix Caching (#2762)

2024-03-02 00:50:01 -08:00

config.py

Push logprob generation to LLMEngine (#3065)

2024-03-04 19:54:06 +00:00

logger.py

Make vLLM logging formatting optional (#2877)

2024-02-20 14:38:55 -08:00

outputs.py

Store eos_token_id in Sequence for easy access (#3166)

2024-03-05 15:35:43 -08:00

py.typed

Add py.typed so consumers of vLLM can get type checking (#1509)

2023-10-30 14:50:47 -07:00

sampling_params.py

[Fix] Don't deep-copy LogitsProcessors when copying SamplingParams (#3099)

2024-02-29 19:20:42 +00:00

sequence.py

Store eos_token_id in Sequence for easy access (#3166)

2024-03-05 15:35:43 -08:00

test_utils.py

Use CuPy for CUDA graphs (#2811)

2024-02-13 11:32:06 -08:00

utils.py

[Minor fix] The domain dns.google may cause a socket.gaierror exception (#3176)

2024-03-04 19:17:12 +00:00