experiments/STT
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

native_client: option to run streaming mode

Browse Source 
Signed-off-by: Li Li <eggonlea@msn.com>
This commit is contained in:
Li Li 2019-06-17 18:30:21 -07:00
parent dce068f3cb
commit 2604da4bc2
2 changed files with 39 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

14
native_client/args.h
View File

@ -30,6 +30,8 @@ bool extended_metadata = false;
bool json_output = false; bool json_output = false;
int stream_size = 0;
void PrintHelp(const char* bin) void PrintHelp(const char* bin)
{ {
std::cout << std::cout <<
@ -45,6 +47,7 @@ void PrintHelp(const char* bin)
" -t Run in benchmark mode, output mfcc & inference time\n" " -t Run in benchmark mode, output mfcc & inference time\n"
" --extended Output string from extended metadata\n" " --extended Output string from extended metadata\n"
" --json Extended output, shows word timings as JSON\n" " --json Extended output, shows word timings as JSON\n"
" --stream size Run in stream mode, output intermediate results\n"
" --help Show help\n" " --help Show help\n"
" --version Print version and exits\n"; " --version Print version and exits\n";
DS_PrintVersions(); DS_PrintVersions();
@ -64,6 +67,7 @@ bool ProcessArgs(int argc, char** argv)
{"t", no_argument, nullptr, 't'}, {"t", no_argument, nullptr, 't'},
{"extended", no_argument, nullptr, 'e'}, {"extended", no_argument, nullptr, 'e'},
{"json", no_argument, nullptr, 'j'}, {"json", no_argument, nullptr, 'j'},
{"stream", required_argument, nullptr, 's'},
{"help", no_argument, nullptr, 'h'}, {"help", no_argument, nullptr, 'h'},
{"version", no_argument, nullptr, 'v'}, {"version", no_argument, nullptr, 'v'},
{nullptr, no_argument, nullptr, 0} {nullptr, no_argument, nullptr, 0}
@ -118,6 +122,10 @@ bool ProcessArgs(int argc, char** argv)
json_output = true; json_output = true;
break; break;
case 's':
stream_size = atoi(optarg);
break;
case 'h': // -h or --help case 'h': // -h or --help
case '?': // Unrecognized option case '?': // Unrecognized option
default: default:
@ -136,6 +144,12 @@ bool ProcessArgs(int argc, char** argv)
return false; return false;
} }
if (stream_size < 0 || stream_size % 160 != 0) {
std::cout <<
"Stream buffer size must be multiples of 160\n";
return false;
}
return true; return true;
} }

25
native_client/client.cc
View File

@ -70,6 +70,31 @@ LocalDsSTT(ModelState* aCtx, const short* aBuffer, size_t aBufferSize,
Metadata *metadata = DS_SpeechToTextWithMetadata(aCtx, aBuffer, aBufferSize, aSampleRate); Metadata *metadata = DS_SpeechToTextWithMetadata(aCtx, aBuffer, aBufferSize, aSampleRate);
res.string = JSONOutput(metadata); res.string = JSONOutput(metadata);
DS_FreeMetadata(metadata); DS_FreeMetadata(metadata);
} else if (stream_size > 0) {
StreamingState* ctx;
int status = DS_SetupStream(aCtx, 0, aSampleRate, &ctx);
if (status != DS_ERR_OK) {
res.string = strdup("");
return res;
}
size_t off = 0;
const char *last = nullptr;
while (off < aBufferSize) {
size_t cur = aBufferSize - off > stream_size ? stream_size : aBufferSize - off;
DS_FeedAudioContent(ctx, aBuffer + off, cur);
off += cur;
const char* partial = DS_IntermediateDecode(ctx);
if (last == nullptr || strcmp(last, partial)) {
printf("%s\n", partial);
last = partial;
} else {
DS_FreeString((char *) partial);
}
}
if (last != nullptr) {
DS_FreeString((char *) last);
}
res.string = DS_FinishStream(ctx);
} else { } else {
res.string = DS_SpeechToText(aCtx, aBuffer, aBufferSize, aSampleRate); res.string = DS_SpeechToText(aCtx, aBuffer, aBufferSize, aSampleRate);
} }