Is this page helpful?

NVIDIA NMT NIM Microservice API References#

The NMT NIM microservice exposes a gRPC API based on protocol buffers for text-to-text and streaming speech translation. The RivaTranslation service supports synchronous translation between language pairs, language-pair discovery, streaming speech-to-text translation, and streaming speech-to-speech translation.

Top

riva/proto/health.proto#

HealthCheckRequest#

Field	Type	Label	Description
service	string

HealthCheckResponse#

Field	Type	Label	Description
status	HealthCheckResponse.ServingStatus

HealthCheckResponse.ServingStatus#

Name	Number	Description
UNKNOWN	0
SERVING	1
NOT_SERVING	2

Health#

Method Name	Request Type	Response Type	Description
Check	HealthCheckRequest	HealthCheckResponse
Watch	HealthCheckRequest	HealthCheckResponse stream

Top

riva/proto/riva_common.proto#

RequestId#

Specifies the request ID of the request.

Field	Type	Label	Description
value	string

Top

riva/proto/riva_nmt.proto#

AvailableLanguageRequest#

Returns a map of model names to its source and target language pairs. Can specify a specific model name to retrieve only its language pairs.

Field	Type	Label	Description
model	string		Supported values: “s2s_model”, “s2t_model”, and name of the deployed t2t model. If empty, returns all available models and languages.

AvailableLanguageResponse#

Language pairs are the sets of src to tgt languages available per model. languages contains all the model_name -> Language pair

Field	Type	Label	Description
languages	AvailableLanguageResponse.LanguagesEntry	repeated

AvailableLanguageResponse.LanguagePair#

Field	Type	Label	Description
src_lang	string	repeated
tgt_lang	string	repeated

AvailableLanguageResponse.LanguagesEntry#

Field	Type	Label	Description
key	string
value	AvailableLanguageResponse.LanguagePair

TranslateTextRequest#

request for synchronous translation of each text in texts. Available languages can be queried using ListSupportLanguagePairs RPC. source and target languages must be specified, are currently two character ISO codes, this will likely change to BCP-47 inline with other Riva Services for GA.

Field	Type	Label	Description
texts	string	repeated
model	string
source_language	string
target_language	string
dnt_phrases	string	repeated	A list of words or phrases that are not to be translated or to be custom translated by the pipeline. Words to be custom translated should be specified as “<word>##<custom_translation>” and words not be translated should be specified as “<word>”.
max_len_variation	string		Controls the maximum allowed variation in token count between the source text and translated output.
id	nvidia.riva.RequestId		The ID to be associated with the request. If provided, this will be returned in the corresponding response.

StreamingTranslateSpeechToTextRequest#

Streams a configuration message followed by source audio chunks for speech-to-text translation.

Field	Type	Label	Description
config	StreamingTranslateSpeechToTextConfig	oneof	Configuration for the ASR and translation stages.
audio_content	bytes	oneof	A chunk of source audio.
id	nvidia.riva.RequestId		The ID associated with the request.

StreamingTranslateSpeechToSpeechRequest#

Streams a configuration message followed by source audio chunks for an ASR, NMT, and TTS translation pipeline.

Field	Type	Label	Description
config	StreamingTranslateSpeechToSpeechConfig	oneof	Configuration for the ASR, translation, and TTS stages.
audio_content	bytes	oneof	A chunk of source audio.
id	nvidia.riva.RequestId		The ID associated with the request.

TranslateTextResponse#

Translations are returned as text:language pairs. These are 1:1 for the passed in ‘texts’ from the request.

Field	Type	Label	Description
translations	Translation	repeated
id	nvidia.riva.RequestId		The ID associated with the request

StreamingTranslateSpeechToTextResponse#

Returns translated streaming recognition results.

Field	Type	Label	Description
results	nvidia.riva.asr.StreamingRecognitionResult	repeated	Streaming recognition results containing translated text.
id	nvidia.riva.RequestId		The ID associated with the request.

StreamingTranslateSpeechToSpeechResponse#

Returns a synthesized audio chunk. The final response contains an empty audio buffer to mark the end of the stream.

Field	Type	Label	Description
speech	nvidia.riva.tts.SynthesizeSpeechResponse		Synthesized speech from the TTS stage.
id	nvidia.riva.RequestId		The ID associated with the request.

Translation#

contains a single translation, collecting into the translate text response Includes the target language code, since with multi lingual models there are multiple possibilities.

Field	Type	Label	Description
text	string
language	string

TranslationConfig#

Field	Type	Label	Description
source_language_code	string		BCP-47 “en-US”
target_language_code	string
model_name	string
dnt_phrases	string	repeated	A list of words or phrases that will not be translated by the pipeline. This list can include special words or phrases, for example, names, acronyms or any phrases desired to be excluded from translation. These words or phrases will be present as-is in the translated output.

SynthesizeSpeechConfig#

Configures the synthesized audio produced by a speech-to-speech translation pipeline.

Field	Type	Description
encoding	AudioEncoding	Encoding for the synthesized audio.
sample_rate_hz	int32	Sample rate of the synthesized audio in hertz.
voice_name	string	TTS voice name.
language_code	string	BCP-47 target language code, for example, `es-US`.
prosody_rate	string	Prosody rate.
prosody_pitch	string	Prosody pitch.
prosody_volume	string	Prosody volume.

StreamingTranslateSpeechToTextConfig#

Configures the ASR and translation stages of a streaming speech-to-text translation pipeline.

Field	Type	Label	Description
asr_config	nvidia.riva.asr.StreamingRecognitionConfig		Streaming ASR configuration.
translation_config	TranslationConfig		Translation configuration.

StreamingTranslateSpeechToSpeechConfig#

Configures the ASR, NMT, and TTS stages of a streaming speech-to-speech translation pipeline.

Field	Type	Description
asr_config	nvidia.riva.asr.StreamingRecognitionConfig	Streaming ASR configuration.
tts_config	SynthesizeSpeechConfig	Speech synthesis configuration.
translation_config	TranslationConfig	Translation configuration.

RivaTranslation#

RivaTranslation provides RPCs for text and speech translation between languages.

Method Name	Request Type	Response Type	Description
TranslateText	TranslateTextRequest	TranslateTextResponse	Translate text to text, from a source to a target language. Currently source and target language fields is required, along with the model name. Multiple texts can be passed per request up to the given batch size for the model, which is set at translation pipeline creation time.
ListSupportedLanguagePairs	AvailableLanguageRequest	AvailableLanguageResponse	Lists the available language pairs and models names to be used for TranslateText
StreamingTranslateSpeechToText	StreamingTranslateSpeechToTextRequest stream	StreamingTranslateSpeechToTextResponse stream	Transcribes source audio and streams translated text.
StreamingTranslateSpeechToSpeech	StreamingTranslateSpeechToSpeechRequest stream	StreamingTranslateSpeechToSpeechResponse stream	Transcribes and translates source audio and streams synthesized speech in the target language.

Scalar Value Types#

.proto Type	Notes	C++	Java	Python	Go	C#	PHP	Ruby
double		double	double	float	float64	double	float	Float
float		float	float	float	float32	float	float	Float
int32	Uses variable-length encoding. Inefficient for encoding negative numbers – if your field is likely to have negative values, use sint32 instead.	int32	int	int	int32	int	integer	Bignum or Fixnum (as required)
int64	Uses variable-length encoding. Inefficient for encoding negative numbers – if your field is likely to have negative values, use sint64 instead.	int64	long	int/long	int64	long	integer/string	Bignum
uint32	Uses variable-length encoding.	uint32	int	int/long	uint32	uint	integer	Bignum or Fixnum (as required)
uint64	Uses variable-length encoding.	uint64	long	int/long	uint64	ulong	integer/string	Bignum or Fixnum (as required)
sint32	Uses variable-length encoding. Signed int value. These more efficiently encode negative numbers than regular int32s.	int32	int	int	int32	int	integer	Bignum or Fixnum (as required)
sint64	Uses variable-length encoding. Signed int value. These more efficiently encode negative numbers than regular int64s.	int64	long	int/long	int64	long	integer/string	Bignum
fixed32	Always four bytes. More efficient than uint32 if values are often greater than 2^28.	uint32	int	int	uint32	uint	integer	Bignum or Fixnum (as required)
fixed64	Always eight bytes. More efficient than uint64 if values are often greater than 2^56.	uint64	long	int/long	uint64	ulong	integer/string	Bignum
sfixed32	Always four bytes.	int32	int	int	int32	int	integer	Bignum or Fixnum (as required)
sfixed64	Always eight bytes.	int64	long	int/long	int64	long	integer/string	Bignum
bool		bool	boolean	boolean	bool	bool	boolean	TrueClass/FalseClass
string	A string must always contain UTF-8 encoded or 7-bit ASCII text.	string	String	str/unicode	string	string	string	String (UTF-8)
bytes	Can contain any arbitrary sequence of bytes.	string	ByteString	str	[]byte	ByteString	string	String (ASCII-8BIT)