Closed parfeniukink closed 3 months ago
click
main
OpenAIBackend
openai.NotFound
SerializableFileType
SerializableFileExtension
str
rate_type_to_load_gen_mode
RATE_TYPE_TO_LOAD_GEN_MODE_MAPPER
rate_type_to_profile_mode
RATE_TYPE_TO_PROFILE_MODE_MAPPER
--num-seconds
--max-seconds
--num-requests
--max-requests
path
GUIDELLM
backend.openai
_base_url
base_url
OpenAIBackend.make_request
GenerativeResponse
output_tokens
self._token_count
SerializableFileExtensions
Summary
click
CLI interface is tested with a bunch of unit testsmain
function validation is added.OpenAIBackend
initializer parameters are optimizedopenai.NotFound
available models error is handledSerializableFileType
renamed toSerializableFileExtension
SerializableFileExtension
now inheritsstr
to simplify usage, since this Enum class is mostly used to work with strings.rate_type_to_load_gen_mode
renamed toRATE_TYPE_TO_LOAD_GEN_MODE_MAPPER
rate_type_to_profile_mode
renamed toRATE_TYPE_TO_PROFILE_MODE_MAPPER
--num-seconds
->--max-seconds
--num-requests
->--max-requests
path
removed from CLI arguments since it is not usedGUIDELLM
prefix is fixedbackend.openai
has_base_url
renamed to thebase_url
OpenAIBackend.make_request
, theGenerativeResponse
always countsoutput_tokens
withself._token_count
SerializableFileExtensions
is replaced with pure Python strings