huggingface/text-generation-inference

Pull Requests Commits

fix(router): fix api-inference deployment (#31)

OlivierDehaene committed 2 years ago

Verified 5c01e254

fix(docker): fix api-inference deployment (#30)

OlivierDehaene committed 2 years ago

Verified ab2ad91d

feat(docker): Make the image compatible with api-inference (#29)

OlivierDehaene committed 2 years ago

Verified f9d0ec37

fix(server): Fix position ids (#28)

OlivierDehaene committed 3 years ago

Verified 1f570d18

feat(server): Support SantaCoder (#26)

OlivierDehaene committed 3 years ago

Verified 15511edc

fix(router): Obey max batch size (#23)

njhill committed 3 years ago

Verified f7ac3949

fix(server): Minor refactorization using new_zeros (#24)

njhill committed 3 years ago

Verified e6d3eb5d

feat(launcher): Log server stdout (#19)

OlivierDehaene committed 3 years ago

Verified fcc2c5fc

fix(server): Use cleanup_tokenization_spaces=False for lossless decoding (#13)

Narsil committed 3 years ago

Verified b94f3021

feat(router): Add const parameters to validation logic (#15)

njhill committed 3 years ago

Verified 60472f9d

fix(router): Include special tokens when tokenizing (#14)

njhill committed 3 years ago

Verified 3efa5bbb

fix(server): Check for device type correctly when determining initial padding (#16)

njhill committed 3 years ago

Verified 686cc667

fix(server): Fix stop sequences (#11)

OlivierDehaene committed 3 years ago

Verified 611e21cb

feat(launcher): Add integration tests (#9)

OlivierDehaene committed 3 years ago

Verified 3e2e6240

feat: Return logprobs (#8)

OlivierDehaene committed 3 years ago

Verified 32a25306

feat: Support stop sequences (#7)

OlivierDehaene committed 3 years ago

Verified 718096f6

fix(server): Only pad to multiple of 8 on GPUs

OlivierDehaene committed 3 years ago

042180d8

feat(server): Add model tests (#6)

OlivierDehaene committed 3 years ago

Verified a2985036

fix(batching): Avoid theoretical hang in batcher loop (#5)

njhill committed 3 years ago

Verified 31d76e23

feat(server): Support Galactica (#4)

OlivierDehaene committed 3 years ago

Verified daa1d81d

fix(router): Handle tokenizer errors

OlivierDehaene committed 3 years ago

d6d5b12e

fix(readme): Typo

OlivierDehaene committed 3 years ago

feb7806c

fix(router): Fix HTTP status codes

OlivierDehaene committed 3 years ago

91f5f862

feat(rust): Update to 1.65

OlivierDehaene committed 3 years ago

6c781025

feat(server): Clarify CausalLMBatch concatenate method

OlivierDehaene committed 3 years ago

dccd5c2b

fix(server): Fix Transformers fork version

OlivierDehaene committed 3 years ago

fa43fb71

feat(server): Improved doc

OlivierDehaene committed 3 years ago

4236e41b

feat(launcher): Pass CUDA_VISIBLE_DEVICES to the shard

OlivierDehaene committed 3 years ago

cea6051e

feat(server): Support AutoModelForSeq2SeqLM

OlivierDehaene committed 3 years ago

427d7cc4

feat(server): Support generic AutoModelForCausalLM

OlivierDehaene committed 3 years ago

c5665f5c

Newer Older