package tei.v1

Get desktop application:
View/edit binary Protocol Buffers messages

• rpc Embed (EmbedRequest, EmbedResponse)

  tei.proto:12

• rpc EmbedStream (stream EmbedRequest, stream EmbedResponse)

  tei.proto:13

• rpc EmbedSparse (EmbedSparseRequest, EmbedSparseResponse)

  tei.proto:14

• rpc EmbedSparseStream (stream EmbedSparseRequest, stream EmbedSparseResponse)

  tei.proto:15

• rpc EmbedAll (EmbedAllRequest, EmbedAllResponse)

  tei.proto:16

• rpc EmbedAllStream (stream EmbedAllRequest, stream EmbedAllResponse)

  tei.proto:17

• rpc Info (InfoRequest, InfoResponse)

  tei.proto:6

  message InfoRequest

  tei.proto:39

  (message has no fields)

  message InfoResponse

  tei.proto:47

  • string version = 1

  • optional string sha = 2

  • optional string docker_label = 3

  • string model_id = 4

  • optional string model_sha = 5

  • string model_dtype = 6

  • ModelType model_type = 7

  • uint32 max_concurrent_requests = 8

  • uint32 max_input_length = 9

  • uint32 max_batch_tokens = 10

  • optional uint32 max_batch_requests = 11

  • uint32 max_client_batch_size = 12

  • uint32 tokenization_workers = 13

• rpc Predict (PredictRequest, PredictResponse)

  tei.proto:21

• rpc PredictPair (PredictPairRequest, PredictResponse)

  tei.proto:22

• rpc PredictStream (stream PredictRequest, stream PredictResponse)

  tei.proto:23

• rpc PredictPairStream (stream PredictPairRequest, stream PredictResponse)

  tei.proto:24

• rpc Rerank (RerankRequest, RerankResponse)

  tei.proto:28

  message RerankRequest

  tei.proto:147

  • string query = 1

  • repeated string texts = 2

  • bool truncate = 3

  • bool raw_scores = 4

  • bool return_text = 5

  • TruncationDirection truncation_direction = 6

• rpc RerankStream (stream RerankStreamRequest, RerankResponse)

  tei.proto:29

  message RerankStreamRequest

  tei.proto:156

  • string query = 1

  • string text = 2

  • bool truncate = 3

  • bool raw_scores = 4

    The server will only consider the first value

  • bool return_text = 5

    The server will only consider the first value

  • TruncationDirection truncation_direction = 6

• rpc Tokenize (EncodeRequest, EncodeResponse)

  tei.proto:33

• rpc TokenizeStream (stream EncodeRequest, stream EncodeResponse)

  tei.proto:34

• rpc Decode (DecodeRequest, DecodeResponse)

  tei.proto:35

• rpc DecodeStream (stream DecodeRequest, stream DecodeResponse)

  tei.proto:36

Used as request type in: Tokenize.Decode, Tokenize.DecodeStream

• repeated uint32 ids = 1

• bool skip_special_tokens = 2

Used as response type in: Tokenize.Decode, Tokenize.DecodeStream

• string text = 1

Used as request type in: Embed.EmbedAll, Embed.EmbedAllStream

• string inputs = 1

• bool truncate = 2

• TruncationDirection truncation_direction = 3

• optional string prompt_name = 4

Used as response type in: Embed.EmbedAll, Embed.EmbedAllStream

• repeated TokenEmbedding token_embeddings = 1

• optional Metadata metadata = 2

Used as request type in: Embed.Embed, Embed.EmbedStream

• string inputs = 1

• bool truncate = 2

• bool normalize = 3

• TruncationDirection truncation_direction = 4

• optional string prompt_name = 5

Used as response type in: Embed.Embed, Embed.EmbedStream

• repeated float embeddings = 1

• optional Metadata metadata = 2

Used as request type in: Embed.EmbedSparse, Embed.EmbedSparseStream

• string inputs = 1

• bool truncate = 2

• TruncationDirection truncation_direction = 3

• optional string prompt_name = 4

Used as response type in: Embed.EmbedSparse, Embed.EmbedSparseStream

• repeated SparseValue sparse_embeddings = 1

• optional Metadata metadata = 2

Used as request type in: Tokenize.Tokenize, Tokenize.TokenizeStream

• string inputs = 1

• bool add_special_tokens = 2

• optional string prompt_name = 3

Used as response type in: Tokenize.Tokenize, Tokenize.TokenizeStream

• repeated SimpleToken tokens = 1

Used in: EmbedAllResponse, EmbedResponse, EmbedSparseResponse, PredictResponse, RerankResponse

• uint32 compute_chars = 1

• uint32 compute_tokens = 2

• uint64 total_time_ns = 3

• uint64 tokenization_time_ns = 4

• uint64 queue_time_ns = 5

• uint64 inference_time_ns = 6

Used in: InfoResponse

• MODEL_TYPE_EMBEDDING = 0

• MODEL_TYPE_CLASSIFIER = 1

• MODEL_TYPE_RERANKER = 2

Used as request type in: Predict.PredictPair, Predict.PredictPairStream

• repeated string inputs = 1

• bool truncate = 2

• bool raw_scores = 3

• TruncationDirection truncation_direction = 4

Used as request type in: Predict.Predict, Predict.PredictStream

• string inputs = 1

• bool truncate = 2

• bool raw_scores = 3

• TruncationDirection truncation_direction = 4

Used as response type in: Predict.Predict, Predict.PredictPair, Predict.PredictPairStream, Predict.PredictStream

• repeated Prediction predictions = 1

• optional Metadata metadata = 2

Used in: PredictResponse

• float score = 1

• string label = 2

Used in: RerankResponse

• uint32 index = 1

• optional string text = 2

• float score = 3

Used as response type in: Rerank.Rerank, Rerank.RerankStream

• repeated Rank ranks = 1

• optional Metadata metadata = 2

Used in: EncodeResponse

• uint32 id = 1

• string text = 2

• bool special = 3

• optional uint32 start = 4

• optional uint32 stop = 5

Used in: EmbedSparseResponse

• uint32 index = 1

• float value = 2

Used in: EmbedAllResponse

• repeated float embeddings = 1

Used in: EmbedAllRequest, EmbedRequest, EmbedSparseRequest, PredictPairRequest, PredictRequest, RerankRequest, RerankStreamRequest

• TRUNCATION_DIRECTION_RIGHT = 0

• TRUNCATION_DIRECTION_LEFT = 1