}
/**
- * @brief Initialize the video info structure
- * @param v_info video info structure to be initialized
+ * @brief Initialize the tensor info structure
+ * @param info tensor info structure to be initialized
*/
void
-gst_tensor_video_info_init (GstTensorVideoInfo * v_info)
+gst_tensor_info_init (GstTensorInfo * info)
{
- g_return_if_fail (v_info != NULL);
+ guint i;
- /**
- * Refer: https://gstreamer.freedesktop.org/documentation/design/mediatype-video-raw.html
- */
- v_info->format = GST_VIDEO_FORMAT_UNKNOWN;
- v_info->w = 0;
- v_info->h = 0;
- v_info->fn = -1;
- v_info->fd = -1;
- v_info->frames = 0;
-}
+ g_return_if_fail (info != NULL);
-/**
- * @brief Initialize the audio info structure
- * @param a_info audio info structure to be initialized
- */
-void
-gst_tensor_audio_info_init (GstTensorAudioInfo * a_info)
-{
- g_return_if_fail (a_info != NULL);
-
- /**
- * Refer: https://gstreamer.freedesktop.org/documentation/design/mediatype-audio-raw.html
- */
- a_info->format = GST_AUDIO_FORMAT_UNKNOWN;
- a_info->ch = 0;
- a_info->rate = 0;
- a_info->frames = 0;
-}
+ info->type = _NNS_END;
-/**
- * @brief Initialize the text info structure
- * @param t_info text info structure to be initialized
- */
-void
-gst_tensor_text_info_init (GstTensorTextInfo * t_info)
-{
- g_return_if_fail (t_info != NULL);
-
- /**
- * Refer: https://gstreamer.freedesktop.org/documentation/design/mediatype-text-raw.html
- */
- t_info->format = 0;
- t_info->size = 0;
- t_info->frames = 0;
+ for (i = 0; i < NNS_TENSOR_RANK_LIMIT; i++) {
+ info->dimension[i] = 0;
+ }
}
/**
- * @brief Set video info to configure tensor
- * @param v_info video info structure to be filled
- * @param structure caps structure
- * @note Fill frames in GstTensorVideoInfo after calling this function.
+ * @brief Check the tensor info is valid
+ * @param info tensor info structure
+ * @return TRUE if info is valid
*/
-void
-gst_tensor_video_info_from_structure (GstTensorVideoInfo * v_info,
- const GstStructure * structure)
+gboolean
+gst_tensor_info_validate (const GstTensorInfo * info)
{
- const gchar *format;
+ guint i;
- g_return_if_fail (v_info != NULL);
- g_return_if_fail (structure != NULL);
+ g_return_val_if_fail (info != NULL, FALSE);
- gst_tensor_video_info_init (v_info);
+ if (info->type == _NNS_END) {
+ return FALSE;
+ }
- format = gst_structure_get_string (structure, "format");
- if (format) {
- v_info->format = gst_video_format_from_string (format);
+ for (i = 0; i < NNS_TENSOR_RANK_LIMIT; i++) {
+ if (info->dimension[i] == 0) {
+ return FALSE;
+ }
}
- gst_structure_get_int (structure, "width", &v_info->w);
- gst_structure_get_int (structure, "height", &v_info->h);
- gst_structure_get_fraction (structure, "framerate", &v_info->fn, &v_info->fd);
+ return TRUE;
}
/**
- * @brief Set audio info to configure tensor
- * @param a_info audio info structure to be filled
- * @param structure caps structure
- * @note Fill frames in GstTensorAudioInfo after calling this function.
+ * @brief Compare tensor info
+ * @param TRUE if equal
*/
-void
-gst_tensor_audio_info_from_structure (GstTensorAudioInfo * a_info,
- const GstStructure * structure)
+gboolean
+gst_tensor_info_is_equal (const GstTensorInfo * i1, const GstTensorInfo * i2)
{
- const gchar *format;
+ guint i;
- g_return_if_fail (a_info != NULL);
- g_return_if_fail (structure != NULL);
+ g_return_val_if_fail (i1 != NULL, FALSE);
+ g_return_val_if_fail (i2 != NULL, FALSE);
- gst_tensor_audio_info_init (a_info);
+ if (i1->type != i2->type) {
+ return FALSE;
+ }
- format = gst_structure_get_string (structure, "format");
- if (format) {
- a_info->format = gst_audio_format_from_string (format);
+ for (i = 0; i < NNS_TENSOR_RANK_LIMIT; i++) {
+ if (i1->dimension[i] != i2->dimension[i]) {
+ return FALSE;
+ }
}
- gst_structure_get_int (structure, "channels", &a_info->ch);
- gst_structure_get_int (structure, "rate", &a_info->rate);
+ /** matched all */
+ return TRUE;
}
/**
- * @brief Set text info to configure tensor
- * @param t_info text info structure to be filled
- * @param structure caps structure
- * @note Fill size and frames in GstTensorTextInfo after calling this function.
+ * @brief Initialize the tensors info structure
+ * @param info tensors info structure to be initialized
*/
void
-gst_tensor_text_info_from_structure (GstTensorTextInfo * t_info,
- const GstStructure * structure)
+gst_tensors_info_init (GstTensorsInfo * info)
{
- const gchar *format;
+ guint i;
- g_return_if_fail (t_info != NULL);
- g_return_if_fail (structure != NULL);
+ g_return_if_fail (info != NULL);
- gst_tensor_text_info_init (t_info);
+ info->num_tensors = 0;
- format = gst_structure_get_string (structure, "format");
- if (format) {
- if (g_str_equal (format, "utf8")) {
- t_info->format = 1;
- }
+ for (i = 0; i < NNS_TENSOR_SIZE_LIMIT; i++) {
+ gst_tensor_info_init (&info->info[i]);
}
}
/**
- * @brief Set the video info structure from tensor config
- * @param v_info video info structure to be filled
- * @param config tensor config structure to be interpreted
- * @note We cannot get the exact media info from tensor config, you have to check media info after calling this function.
- * @return TRUE if no error
+ * @brief Check the tensors info is valid
+ * @param info tensors info structure
+ * @return TRUE if info is valid
*/
gboolean
-gst_tensor_video_info_from_config (GstTensorVideoInfo * v_info,
- const GstTensorConfig * config)
+gst_tensors_info_validate (const GstTensorsInfo * info)
{
- g_return_val_if_fail (config != NULL, FALSE);
- g_return_val_if_fail (v_info != NULL, FALSE);
+ guint i;
- gst_tensor_video_info_init (v_info);
+ g_return_val_if_fail (info != NULL, FALSE);
- /**
- * Set video format with the dimension (color-space).
- */
- switch (config->dimension[0]) {
- case 1:
- v_info->format = GST_VIDEO_FORMAT_GRAY8;
- break;
- case 3:
- v_info->format = GST_VIDEO_FORMAT_RGB;
- break;
- case 4:
- v_info->format = GST_VIDEO_FORMAT_BGRx;
- break;
- default:
- v_info->format = GST_VIDEO_FORMAT_UNKNOWN;
- break;
+ if (info->num_tensors < 1) {
+ return FALSE;
}
- v_info->w = config->dimension[1];
- v_info->h = config->dimension[2];
- v_info->frames = config->dimension[3];
- v_info->fn = config->rate_n;
- v_info->fd = config->rate_d;
- return TRUE;
-}
-
-/**
- * @brief Set the audio info structure from tensor config
- * @param a_info audio info structure to be filled
- * @param config tensor config structure to be interpreted
- * @note We cannot get the exact media info from tensor config, you have to check media info after calling this function.
- * @return TRUE if no error
- */
-gboolean
-gst_tensor_audio_info_from_config (GstTensorAudioInfo * a_info,
- const GstTensorConfig * config)
-{
- g_return_val_if_fail (config != NULL, FALSE);
- g_return_val_if_fail (a_info != NULL, FALSE);
-
- gst_tensor_audio_info_init (a_info);
-
- /**
- * Set audio format with the type.
- */
- switch (config->type) {
- case _NNS_INT8:
- a_info->format = GST_AUDIO_FORMAT_S8;
- break;
- case _NNS_UINT8:
- a_info->format = GST_AUDIO_FORMAT_U8;
- break;
- case _NNS_INT16:
- a_info->format = GST_AUDIO_FORMAT_S16;
- break;
- case _NNS_UINT16:
- a_info->format = GST_AUDIO_FORMAT_U16;
- break;
- default:
- a_info->format = GST_AUDIO_FORMAT_UNKNOWN;
- break;
+ for (i = 0; i < info->num_tensors; i++) {
+ if (!gst_tensor_info_validate (&info->info[i])) {
+ return FALSE;
+ }
}
- a_info->ch = config->dimension[0];
- a_info->frames = config->dimension[1];
- a_info->rate = config->rate_n;
return TRUE;
}
/**
- * @brief Set the text info structure from tensor config
- * @param t_info text info structure to be filled
- * @param config tensor config structure to be interpreted
- * @note We cannot get the exact media info from tensor config, you have to check media info after calling this function.
- * @return TRUE if no error
+ * @brief Compare tensors info
+ * @param TRUE if equal
*/
gboolean
-gst_tensor_text_info_from_config (GstTensorTextInfo * t_info,
- const GstTensorConfig * config)
+gst_tensors_info_is_equal (const GstTensorsInfo * i1, const GstTensorsInfo * i2)
{
- g_return_val_if_fail (config != NULL, FALSE);
- g_return_val_if_fail (t_info != NULL, FALSE);
+ guint i;
- gst_tensor_text_info_init (t_info);
+ g_return_val_if_fail (i1 != NULL, FALSE);
+ g_return_val_if_fail (i2 != NULL, FALSE);
- /**
- * Set text format. Supposed utf8 if type is int8.
- */
- if (config->type == _NNS_INT8) {
- /** Supposed utf8 */
- t_info->format = 1;
+ if (i1->num_tensors != i2->num_tensors) {
+ return FALSE;
+ }
+
+ for (i = 0; i < i1->num_tensors; i++) {
+ if (!gst_tensor_info_is_equal (&i1->info[i], &i2->info[i])) {
+ return FALSE;
+ }
}
- t_info->size = config->dimension[0];
- t_info->frames = config->dimension[1];
+ /** matched all */
return TRUE;
}
void
gst_tensor_config_init (GstTensorConfig * config)
{
- guint i;
-
g_return_if_fail (config != NULL);
- config->type = _NNS_END;
-
- for (i = 0; i < NNS_TENSOR_RANK_LIMIT; i++) {
- config->dimension[i] = 0;
- }
+ gst_tensor_info_init (&config->info);
config->rate_n = -1;
config->rate_d = -1;
gboolean
gst_tensor_config_validate (const GstTensorConfig * config)
{
- guint i;
-
g_return_val_if_fail (config != NULL, FALSE);
- if (config->type == _NNS_END) {
- return FALSE;
- }
-
- for (i = 0; i < NNS_TENSOR_RANK_LIMIT; i++) {
- if (config->dimension[i] == 0) {
- return FALSE;
- }
- }
-
if (config->rate_n < 0 || config->rate_d < 0) {
return FALSE;
}
- return TRUE;
+ return gst_tensor_info_validate (&config->info);
}
/**
* @brief Compare tensor config info
- * @param TRUE if same
+ * @param TRUE if equal
*/
gboolean
-gst_tensor_config_is_same (const GstTensorConfig * c1,
+gst_tensor_config_is_equal (const GstTensorConfig * c1,
const GstTensorConfig * c2)
{
- guint i;
-
g_return_val_if_fail (c1 != NULL, FALSE);
g_return_val_if_fail (c2 != NULL, FALSE);
- if (c1->type != c2->type) {
- return FALSE;
- }
-
if (c1->rate_n != c2->rate_n || c1->rate_d != c2->rate_d) {
return FALSE;
}
- for (i = 0; i < NNS_TENSOR_RANK_LIMIT; i++) {
- if (c1->dimension[i] != c2->dimension[i]) {
- return FALSE;
- }
- }
-
- /** matched all */
- return TRUE;
+ return gst_tensor_info_is_equal (&c1->info, &c2->info);
}
/**
- * @brief Parse structure and set tensor config info
+ * @brief Parse structure and set tensor config info (internal static function)
* @param config tensor config structure to be filled
* @param structure structure to be interpreted
* @return TRUE if ok
gst_tensor_config_from_tensor_structure (GstTensorConfig * config,
const GstStructure * structure)
{
+ GstTensorInfo *info;
const gchar *type_string;
g_return_val_if_fail (config != NULL, FALSE);
- g_return_val_if_fail (structure != NULL, FALSE);
-
gst_tensor_config_init (config);
+ info = &config->info;
+
+ g_return_val_if_fail (structure != NULL, FALSE);
if (!gst_structure_has_name (structure, "other/tensor")) {
err_print ("caps is not tensor %s\n", gst_structure_get_name (structure));
return FALSE;
}
- gst_structure_get_int (structure, "dim1", (gint *) (&config->dimension[0]));
- gst_structure_get_int (structure, "dim2", (gint *) (&config->dimension[1]));
- gst_structure_get_int (structure, "dim3", (gint *) (&config->dimension[2]));
- gst_structure_get_int (structure, "dim4", (gint *) (&config->dimension[3]));
+ gst_structure_get_int (structure, "dim1", (gint *) (&info->dimension[0]));
+ gst_structure_get_int (structure, "dim2", (gint *) (&info->dimension[1]));
+ gst_structure_get_int (structure, "dim3", (gint *) (&info->dimension[2]));
+ gst_structure_get_int (structure, "dim4", (gint *) (&info->dimension[3]));
type_string = gst_structure_get_string (structure, "type");
- config->type = get_tensor_type (type_string);
+ info->type = get_tensor_type (type_string);
gst_structure_get_fraction (structure, "framerate", &config->rate_n,
&config->rate_d);
}
/**
- * @brief Parse caps and set tensor config info
- * @param config tensor config structure to be filled
- * @param structure structure to be interpreted
- * @return TRUE if ok
- */
-static gboolean
-gst_tensor_config_from_media_structure (GstTensorConfig * config,
- const GstStructure * structure)
-{
- media_type m_type;
-
- g_return_val_if_fail (config != NULL, FALSE);
- g_return_val_if_fail (structure != NULL, FALSE);
-
- gst_tensor_config_init (config);
-
- m_type = gst_tensor_media_type_from_structure (structure);
-
- switch (m_type) {
- case _NNS_VIDEO:
- {
- GstTensorVideoInfo v_info;
-
- gst_tensor_video_info_from_structure (&v_info, structure);
-
- /** supposed 1 frame in tensor */
- v_info.frames = 1;
-
- gst_tensor_config_from_video_info (config, &v_info);
- break;
- }
- case _NNS_AUDIO:
- {
- GstTensorAudioInfo a_info;
-
- gst_tensor_audio_info_from_structure (&a_info, structure);
-
- /** supposed 1 frame in tensor */
- a_info.frames = 1;
-
- gst_tensor_config_from_audio_info (config, &a_info);
- break;
- }
- case _NNS_STRING:
- {
- GstTensorTextInfo t_info;
-
- gst_tensor_text_info_from_structure (&t_info, structure);
-
- /** fixed size of string */
- t_info.size = GST_TENSOR_STRING_SIZE;
- /** supposed 1 frame in tensor */
- t_info.frames = 1;
-
- gst_tensor_config_from_text_info (config, &t_info);
- break;
- }
- default:
- err_print ("Unsupported type %d\n", m_type);
- return FALSE;
- }
-
- return TRUE;
-}
-
-/**
- * @brief Parse structure and set tensor config info
- * @param config tensor config structure to be filled
- * @param structure structure to be interpreted
- * @note Change dimention if tensor contains N frames.
- * @return TRUE if no error
- */
-gboolean
-gst_tensor_config_from_structure (GstTensorConfig * config,
- const GstStructure * structure)
-{
- g_return_val_if_fail (config != NULL, FALSE);
- g_return_val_if_fail (structure != NULL, FALSE);
-
- if (gst_structure_has_name (structure, "other/tensor")) {
- return gst_tensor_config_from_tensor_structure (config, structure);
- } else {
- return gst_tensor_config_from_media_structure (config, structure);
- }
-}
-
-/**
- * @brief Set the tensor config structure from video info
+ * @brief Set the tensor config structure from video info (internal static function)
* @param config tensor config structure to be filled
- * @param v_info video info structure to be interpreted
+ * @param structure caps structure
* @note Change dimention if tensor contains N frames.
* @return TRUE if supported type
*/
-gboolean
+static gboolean
gst_tensor_config_from_video_info (GstTensorConfig * config,
- const GstTensorVideoInfo * v_info)
+ const GstStructure * structure)
{
/**
* Refer: https://www.tensorflow.org/api_docs/python/tf/summary/image
* A 4-D uint8 or float32 Tensor of shape [batch_size, height, width, channels]
* where channels is 1, 3, or 4.
*/
- g_return_val_if_fail (config != NULL, FALSE);
- g_return_val_if_fail (v_info != NULL, FALSE);
+ const gchar *format_string;
+ GstVideoFormat format = GST_VIDEO_FORMAT_UNKNOWN;
+ gint width = 0;
+ gint height = 0;
+ gint fn = -1; /** numerator */
+ gint fd = -1; /** denominator */
+ g_return_val_if_fail (config != NULL, FALSE);
gst_tensor_config_init (config);
+ g_return_val_if_fail (structure != NULL, FALSE);
+
+ format_string = gst_structure_get_string (structure, "format");
+ if (format_string) {
+ format = gst_video_format_from_string (format_string);
+ }
+
+ gst_structure_get_int (structure, "width", &width);
+ gst_structure_get_int (structure, "height", &height);
+ gst_structure_get_fraction (structure, "framerate", &fn, &fd);
+
/** [color-space][width][height][frames] */
- switch (v_info->format) {
+ switch (format) {
case GST_VIDEO_FORMAT_GRAY8:
- config->type = _NNS_UINT8;
- config->dimension[0] = 1;
+ config->info.type = _NNS_UINT8;
+ config->info.dimension[0] = 1;
break;
case GST_VIDEO_FORMAT_RGB:
- config->type = _NNS_UINT8;
- config->dimension[0] = 3;
+ config->info.type = _NNS_UINT8;
+ config->info.dimension[0] = 3;
break;
case GST_VIDEO_FORMAT_BGRx:
- config->type = _NNS_UINT8;
- config->dimension[0] = 4;
+ config->info.type = _NNS_UINT8;
+ config->info.dimension[0] = 4;
break;
default:
/** unsupported format */
- err_print ("Unsupported format = %d\n", v_info->format);
+ err_print ("Unsupported format = %d\n", format);
break;
}
- config->dimension[1] = v_info->w;
- config->dimension[2] = v_info->h;
- config->dimension[3] = v_info->frames; /** change this if tensor contains N frames */
+ config->info.dimension[1] = width;
+ config->info.dimension[2] = height;
+ config->info.dimension[3] = 1; /** Supposed 1 frame in tensor, change this if tensor contains N frames. */
- config->rate_n = v_info->fn;
- config->rate_d = v_info->fd;
+ config->rate_n = fn;
+ config->rate_d = fd;
- return (config->type != _NNS_END);
+ return (config->info.type != _NNS_END);
}
/**
- * @brief Set the tensor config structure from audio info
+ * @brief Set the tensor config structure from audio info (internal static function)
* @param config tensor config structure to be filled
- * @param a_info audio info structure to be interpreted
+ * @param structure caps structure
* @note Change dimention if tensor contains N frames.
* @return TRUE if supported type
*/
-gboolean
+static gboolean
gst_tensor_config_from_audio_info (GstTensorConfig * config,
- const GstTensorAudioInfo * a_info)
+ const GstStructure * structure)
{
/**
* Refer: https://www.tensorflow.org/api_docs/python/tf/summary/audio
* A 3-D float32 Tensor of shape [batch_size, frames, channels]
* or a 2-D float32 Tensor of shape [batch_size, frames].
*/
- g_return_val_if_fail (config != NULL, FALSE);
- g_return_val_if_fail (a_info != NULL, FALSE);
+ const gchar *format_string;
+ GstAudioFormat format = GST_AUDIO_FORMAT_UNKNOWN;
+ gint channels = 0;
+ gint rate = 0;
+ g_return_val_if_fail (config != NULL, FALSE);
gst_tensor_config_init (config);
+ g_return_val_if_fail (structure != NULL, FALSE);
+
+ format_string = gst_structure_get_string (structure, "format");
+ if (format_string) {
+ format = gst_audio_format_from_string (format_string);
+ }
+
+ gst_structure_get_int (structure, "channels", &channels);
+ gst_structure_get_int (structure, "rate", &rate);
+
/** [channels][frames] */
- switch (a_info->format) {
+ switch (format) {
case GST_AUDIO_FORMAT_S8:
- config->type = _NNS_INT8;
+ config->info.type = _NNS_INT8;
break;
case GST_AUDIO_FORMAT_U8:
- config->type = _NNS_UINT8;
+ config->info.type = _NNS_UINT8;
break;
case GST_AUDIO_FORMAT_S16:
- config->type = _NNS_INT16;
+ config->info.type = _NNS_INT16;
break;
case GST_AUDIO_FORMAT_U16:
- config->type = _NNS_UINT16;
+ config->info.type = _NNS_UINT16;
break;
default:
/** unsupported format */
- err_print ("Unsupported format = %d\n", a_info->format);
+ err_print ("Unsupported format = %d\n", format);
break;
}
- config->dimension[0] = a_info->ch;
- config->dimension[1] = a_info->frames; /** change this if tensor contains N frames */
- config->dimension[2] = 1;
- config->dimension[3] = 1;
+ config->info.dimension[0] = channels;
+ config->info.dimension[1] = 1; /** Supposed 1 frame in tensor, change this if tensor contains N frames */
+ config->info.dimension[2] = 1;
+ config->info.dimension[3] = 1;
- if (a_info->rate > 0) {
- config->rate_n = a_info->rate;
+ if (rate > 0) {
+ config->rate_n = rate;
config->rate_d = 1;
}
- return (config->type != _NNS_END);
+ return (config->info.type != _NNS_END);
}
/**
- * @brief Set the tensor config structure from text info
+ * @brief Set the tensor config structure from text info (internal static function)
* @param config tensor config structure to be filled
- * @param t_info text info structure to be interpreted
+ * @param structure caps structure
* @note Change dimention if tensor contains N frames.
* @return TRUE if supported type
*/
-gboolean
+static gboolean
gst_tensor_config_from_text_info (GstTensorConfig * config,
- const GstTensorTextInfo * t_info)
+ const GstStructure * structure)
{
/**
* Refer: https://www.tensorflow.org/api_docs/python/tf/summary/text
* A string-type Tensor
*/
- g_return_val_if_fail (config != NULL, FALSE);
- g_return_val_if_fail (t_info != NULL, FALSE);
+ const gchar *format_string;
+ g_return_val_if_fail (config != NULL, FALSE);
gst_tensor_config_init (config);
- /** [size][frames] */
- switch (t_info->format) {
- case 1:
- /** utf8 */
- config->type = _NNS_INT8;
- break;
- default:
+ g_return_val_if_fail (structure != NULL, FALSE);
+
+ format_string = gst_structure_get_string (structure, "format");
+ if (format_string) {
+ if (g_str_equal (format_string, "utf8")) {
+ config->info.type = _NNS_INT8;
+ } else {
/** unsupported format */
- err_print ("Unsupported format = %d\n", t_info->format);
- break;
+ err_print ("Unsupported format\n");
+ }
}
- config->dimension[0] = t_info->size;
- config->dimension[1] = t_info->frames; /** change this if tensor contains N frames */
- config->dimension[2] = 1;
- config->dimension[3] = 1;
+ /** [size][frames] */
+ config->info.dimension[0] = GST_TENSOR_STRING_SIZE; /** fixed size of string */
+ config->info.dimension[1] = 1; /** Supposed 1 frame in tensor,, change this if tensor contains N frames */
+ config->info.dimension[2] = 1;
+ config->info.dimension[3] = 1;
/** cannot get the framerate for text type */
config->rate_n = 0;
config->rate_d = 1;
- return (config->type != _NNS_END);
+ return (config->info.type != _NNS_END);
+}
+
+/**
+ * @brief Parse caps and set tensor config info
+ * @param config tensor config structure to be filled
+ * @param structure structure to be interpreted
+ * @return TRUE if ok
+ */
+static gboolean
+gst_tensor_config_from_media_structure (GstTensorConfig * config,
+ const GstStructure * structure)
+{
+ media_type m_type;
+
+ g_return_val_if_fail (config != NULL, FALSE);
+ gst_tensor_config_init (config);
+
+ g_return_val_if_fail (structure != NULL, FALSE);
+
+ m_type = gst_tensor_media_type_from_structure (structure);
+
+ switch (m_type) {
+ case _NNS_VIDEO:
+ gst_tensor_config_from_video_info (config, structure);
+ break;
+ case _NNS_AUDIO:
+ gst_tensor_config_from_audio_info (config, structure);
+ break;
+ case _NNS_STRING:
+ gst_tensor_config_from_text_info (config, structure);
+ break;
+ default:
+ err_print ("Unsupported type %d\n", m_type);
+ return FALSE;
+ }
+
+ return TRUE;
+}
+
+/**
+ * @brief Parse structure and set tensor config info
+ * @param config tensor config structure to be filled
+ * @param structure structure to be interpreted
+ * @note Change dimention if tensor contains N frames.
+ * @return TRUE if no error
+ */
+gboolean
+gst_tensor_config_from_structure (GstTensorConfig * config,
+ const GstStructure * structure)
+{
+ g_return_val_if_fail (config != NULL, FALSE);
+ g_return_val_if_fail (structure != NULL, FALSE);
+
+ if (gst_structure_has_name (structure, "other/tensor")) {
+ return gst_tensor_config_from_tensor_structure (config, structure);
+ } else {
+ /** tensor config from media stream */
+ return gst_tensor_config_from_media_structure (config, structure);
+ }
}
/**
caps = gst_caps_from_string (GST_TENSOR_CAP_DEFAULT);
- if (config->dimension[0] > 0) {
- gst_caps_set_simple (caps, "dim1", G_TYPE_INT, config->dimension[0], NULL);
+ if (config->info.dimension[0] > 0) {
+ gst_caps_set_simple (caps, "dim1", G_TYPE_INT, config->info.dimension[0],
+ NULL);
}
- if (config->dimension[1] > 0) {
- gst_caps_set_simple (caps, "dim2", G_TYPE_INT, config->dimension[1], NULL);
+ if (config->info.dimension[1] > 0) {
+ gst_caps_set_simple (caps, "dim2", G_TYPE_INT, config->info.dimension[1],
+ NULL);
}
- if (config->dimension[2] > 0) {
- gst_caps_set_simple (caps, "dim3", G_TYPE_INT, config->dimension[2], NULL);
+ if (config->info.dimension[2] > 0) {
+ gst_caps_set_simple (caps, "dim3", G_TYPE_INT, config->info.dimension[2],
+ NULL);
}
- if (config->dimension[3] > 0) {
- gst_caps_set_simple (caps, "dim4", G_TYPE_INT, config->dimension[3], NULL);
+ if (config->info.dimension[3] > 0) {
+ gst_caps_set_simple (caps, "dim4", G_TYPE_INT, config->info.dimension[3],
+ NULL);
}
- if (config->type != _NNS_END) {
+ if (config->info.type != _NNS_END) {
gst_caps_set_simple (caps, "type", G_TYPE_STRING,
- tensor_element_typename[config->type], NULL);
+ tensor_element_typename[config->info.type], NULL);
+ }
+
+ if (config->rate_n >= 0 && config->rate_d > 0) {
+ gst_caps_set_simple (caps, "framerate", GST_TYPE_FRACTION,
+ config->rate_n, config->rate_d, NULL);
+ }
+
+ return gst_caps_simplify (caps);
+}
+
+/**
+ * @brief Initialize the tensors config info structure (for other/tensors)
+ * @param config tensors config structure to be initialized
+ */
+void
+gst_tensors_config_init (GstTensorsConfig * config)
+{
+ g_return_if_fail (config != NULL);
+
+ gst_tensors_info_init (&config->info);
+
+ config->rate_n = -1;
+ config->rate_d = -1;
+}
+
+/**
+ * @brief Check the tensors are all configured
+ * @param config tensor config structure
+ * @return TRUE if configured
+ */
+gboolean
+gst_tensors_config_validate (const GstTensorsConfig * config)
+{
+ g_return_val_if_fail (config != NULL, FALSE);
+
+ if (config->rate_n < 0 || config->rate_d < 0) {
+ return FALSE;
+ }
+
+ return gst_tensors_info_validate (&config->info);
+}
+
+/**
+ * @brief Compare tensor config info
+ * @param TRUE if equal
+ */
+gboolean
+gst_tensors_config_is_equal (const GstTensorsConfig * c1,
+ const GstTensorsConfig * c2)
+{
+ g_return_val_if_fail (c1 != NULL, FALSE);
+ g_return_val_if_fail (c2 != NULL, FALSE);
+
+ if (c1->rate_n != c2->rate_n || c1->rate_d != c2->rate_d) {
+ return FALSE;
+ }
+
+ return gst_tensors_info_is_equal (&c1->info, &c2->info);
+}
+
+/**
+ * @brief Parse structure and set tensors config (for other/tensors)
+ * @param config tensors config structure to be filled
+ * @param structure structure to be interpreted
+ * @return TRUE if no error
+ */
+gboolean
+gst_tensors_config_from_structure (GstTensorsConfig * config,
+ const GstStructure * structure)
+{
+ const gchar *name;
+ guint i, j;
+
+ g_return_val_if_fail (config != NULL, FALSE);
+ gst_tensors_config_init (config);
+
+ g_return_val_if_fail (structure != NULL, FALSE);
+
+ name = gst_structure_get_name (structure);
+
+ if (g_str_equal (name, "other/tensor")) {
+ GstTensorConfig c;
+
+ gst_tensor_config_from_tensor_structure (&c, structure);
+
+ config->info.num_tensors = 1;
+ config->info.info[0] = c.info;
+ config->rate_d = c.rate_d;
+ config->rate_n = c.rate_n;
+ } else if (g_str_equal (name, "other/tensors")) {
+ const gchar *dims_string;
+ const gchar *types_string;
+
+ gst_structure_get_uint (structure, "num_tensors",
+ &config->info.num_tensors);
+ gst_structure_get_fraction (structure, "framerate", &config->rate_n,
+ &config->rate_d);
+
+ /* parse dimensions */
+ dims_string = gst_structure_get_string (structure, "dimensions");
+ if (dims_string) {
+ gchar **str_dim, **str_dims;
+ gint num_dim, num_dims;
+ guint64 val;
+
+ str_dims = g_strsplit (dims_string, ",", -1);
+ num_dims = g_strv_length (str_dims);
+
+ if (config->info.num_tensors != num_dims) {
+ err_print ("Invalid param, dimensions (%d) tensors (%d)\n",
+ num_dims, config->info.num_tensors);
+ }
+
+ for (i = 0; i < num_dims; i++) {
+ str_dim = g_strsplit (str_dims[i], ":", NNS_TENSOR_RANK_LIMIT);
+ num_dim = g_strv_length (str_dim);
+
+ for (j = 0; j < num_dim; j++) {
+ val = g_ascii_strtoull (str_dim[j], NULL, 10);
+ config->info.info[i].dimension[j] = (uint32_t) val;
+ }
+
+ for (; j < NNS_TENSOR_RANK_LIMIT; j++) {
+ config->info.info[i].dimension[j] = 1;
+ }
+
+ g_strfreev (str_dim);
+ }
+
+ g_strfreev (str_dims);
+ }
+
+ /** parse types */
+ types_string = gst_structure_get_string (structure, "types");
+ if (types_string) {
+ gchar **str_types;
+ gint num_types;
+
+ str_types = g_strsplit (dims_string, ",", -1);
+ num_types = g_strv_length (str_types);
+
+ if (config->info.num_tensors != num_types) {
+ err_print ("Invalid param, types (%d) tensors (%d)\n",
+ num_types, config->info.num_tensors);
+ }
+
+ for (i = 0; i < num_types; i++) {
+ config->info.info[i].type = get_tensor_type (str_types[i]);
+ }
+
+ g_strfreev (str_types);
+ }
+ } else {
+ err_print ("Unsupported type = %s\n", name);
+ return FALSE;
+ }
+
+ return TRUE;
+}
+
+/**
+ * @brief Get caps from tensors config (for other/tensors)
+ * @param config tensors config info
+ * @return caps for given config
+ */
+GstCaps *
+gst_tensors_caps_from_config (const GstTensorsConfig * config)
+{
+ GstCaps *caps;
+ guint i;
+
+ g_return_val_if_fail (config != NULL, NULL);
+
+ caps = gst_caps_from_string (GST_TENSORS_CAP_DEFAULT);
+
+ if (config->info.num_tensors > 0) {
+ GString *dimensions = g_string_new (NULL);
+ GString *types = g_string_new (NULL);
+
+ /** dimensions and types */
+ for (i = 0; i < config->info.num_tensors; i++) {
+ /**
+ * @note supposed dimension with rank 4 (NNS_TENSOR_RANK_LIMIT)
+ */
+ gchar *dim_string = g_strdup_printf ("%d:%d:%d:%d",
+ config->info.info[i].dimension[0], config->info.info[i].dimension[1],
+ config->info.info[i].dimension[2], config->info.info[i].dimension[3]);
+
+ dimensions = g_string_append (dimensions, dim_string);
+ types =
+ g_string_append (types,
+ tensor_element_typename[config->info.info[i].type]);
+
+ if (i < config->info.num_tensors - 1) {
+ dimensions = g_string_append (dimensions, ",");
+ types = g_string_append (types, ",");
+ }
+
+ g_free (dim_string);
+ }
+
+ gst_caps_set_simple (caps, "num_tensors", G_TYPE_INT,
+ config->info.num_tensors, NULL);
+ gst_caps_set_simple (caps, "dimensions", G_TYPE_STRING, dimensions->str,
+ NULL);
+ gst_caps_set_simple (caps, "types", G_TYPE_STRING, types->str, NULL);
+
+ g_string_free (dimensions, TRUE);
+ g_string_free (types, TRUE);
}
if (config->rate_n >= 0 && config->rate_d > 0) {
*/
int
get_tensor_dimension (const gchar * param,
- uint32_t dim[NNS_TENSOR_SIZE_LIMIT][NNS_TENSOR_RANK_LIMIT],
- int rank[NNS_TENSOR_SIZE_LIMIT])
+ uint32_t dim[NNS_TENSOR_SIZE_LIMIT][NNS_TENSOR_RANK_LIMIT])
{
int i, j = 0;
guint64 val;
break;
val = g_ascii_strtoull (strv[j], NULL, 10);
dim[i][j] = val;
- rank[i] = j + 1;
}
for (; j < NNS_TENSOR_RANK_LIMIT; j++)
dim[i][j] = 1;
GstTensor_TensorsMeta * meta, int *framerate_num, int *framerate_denom)
{
int num = 0;
- int rank[NNS_TENSOR_SIZE_LIMIT];
const gchar *strval;
gint fn = 0, fd = 0;
gchar **strv;
strval = gst_structure_get_string (str, "dimensions");
- counter = get_tensor_dimension (strval, meta->dims, rank);
+ counter = get_tensor_dimension (strval, meta->dims);
if (counter != num) {
err_print
/** update dimension in output tensor */
if (self->frames_dim >= 0) {
- config.dimension[self->frames_dim] = self->frames_out;
+ config.info.dimension[self->frames_dim] = self->frames_out;
}
self->out_config = config;
* <refsect2>
* <title>Example launch line</title>
* |[
- * gst-launch-1.0 videotestsrc ! video/x-raw,format=RGB,width=640,height=480 ! tensor_convert ! tensor_sink
+ * gst-launch-1.0 videotestsrc ! video/x-raw,format=RGB,width=640,height=480 ! tensor_converter ! tensor_sink
* ]|
* </refsect2>
*/
switch (self->in_media_type) {
case _NNS_VIDEO:
frame_size =
- config->dimension[0] * config->dimension[1] * config->dimension[2] *
- tensor_element_size[config->type];
+ config->info.dimension[0] * config->info.dimension[1] *
+ config->info.dimension[2] * tensor_element_size[config->info.type];
frames_in = 1; /** supposed 1 frame in buffer */
if (self->remove_padding) {
/**
* Refer: https://gstreamer.freedesktop.org/documentation/design/mediatype-video-raw.html
*/
- size = offset = config->dimension[0] * config->dimension[1];
+ size = offset = config->info.dimension[0] * config->info.dimension[1];
g_assert (offset % 4);
if (offset % 4) {
for (d0 = 0; d0 < frames_in; d0++) {
/** Supposed to be 0 only, 1 frame in buffer. */
g_assert (d0 == 0);
- for (d1 = 0; d1 < config->dimension[2]; d1++) { /** Height */
+ for (d1 = 0; d1 < config->info.dimension[2]; d1++) { /** Height */
memcpy (destptr + dest_idx, srcptr + src_idx, size);
dest_idx += size;
src_idx += offset;
}
/** set the number of frames in dimension */
- config.dimension[frames_dim] = self->frames_per_tensor;
+ config.info.dimension[frames_dim] = self->frames_per_tensor;
self->in_media_type = in_type;
self->tensor_configured = TRUE;
gst_tensordec_video_caps_from_config (GstTensorDec * self,
const GstTensorConfig * config)
{
- GstTensorVideoInfo v_info;
+ GstVideoFormat format;
+ gint width, height, fn, fd;
GstCaps *caps;
g_return_val_if_fail (config != NULL, NULL);
caps = gst_caps_from_string (GST_TENSOR_VIDEO_CAPS_STR);
- gst_tensor_video_info_from_config (&v_info, config);
+ switch (config->info.dimension[0]) {
+ case 1:
+ format = GST_VIDEO_FORMAT_GRAY8;
+ break;
+ case 3:
+ format = GST_VIDEO_FORMAT_RGB;
+ break;
+ case 4:
+ format = GST_VIDEO_FORMAT_BGRx;
+ break;
+ default:
+ format = GST_VIDEO_FORMAT_UNKNOWN;
+ break;
+ }
+
+ width = config->info.dimension[1];
+ height = config->info.dimension[2];
+ fn = config->rate_n;
+ fd = config->rate_d;
- if (v_info.format != GST_VIDEO_FORMAT_UNKNOWN) {
- const gchar *format_string = gst_video_format_to_string (v_info.format);
+ if (format != GST_VIDEO_FORMAT_UNKNOWN) {
+ const gchar *format_string = gst_video_format_to_string (format);
gst_caps_set_simple (caps, "format", G_TYPE_STRING, format_string, NULL);
}
- if (v_info.w > 0) {
- gst_caps_set_simple (caps, "width", G_TYPE_INT, v_info.w, NULL);
+ if (width > 0) {
+ gst_caps_set_simple (caps, "width", G_TYPE_INT, width, NULL);
}
- if (v_info.h > 0) {
- gst_caps_set_simple (caps, "height", G_TYPE_INT, v_info.h, NULL);
+ if (height > 0) {
+ gst_caps_set_simple (caps, "height", G_TYPE_INT, height, NULL);
}
- if (v_info.fn > 0 && v_info.fd > 0) {
- gst_caps_set_simple (caps, "framerate", GST_TYPE_FRACTION,
- v_info.fn, v_info.fd, NULL);
+ if (fn > 0 && fd > 0) {
+ gst_caps_set_simple (caps, "framerate", GST_TYPE_FRACTION, fn, fd, NULL);
}
return gst_caps_simplify (caps);
gst_tensordec_audio_caps_from_config (GstTensorDec * self,
const GstTensorConfig * config)
{
- GstTensorAudioInfo a_info;
+ GstAudioFormat format;
+ gint ch, rate;
GstCaps *caps;
g_return_val_if_fail (config != NULL, NULL);
caps = gst_caps_from_string (GST_TENSOR_AUDIO_CAPS_STR);
- gst_tensor_audio_info_from_config (&a_info, config);
+ switch (config->info.type) {
+ case _NNS_INT8:
+ format = GST_AUDIO_FORMAT_S8;
+ break;
+ case _NNS_UINT8:
+ format = GST_AUDIO_FORMAT_U8;
+ break;
+ case _NNS_INT16:
+ format = GST_AUDIO_FORMAT_S16;
+ break;
+ case _NNS_UINT16:
+ format = GST_AUDIO_FORMAT_U16;
+ break;
+ default:
+ format = GST_AUDIO_FORMAT_UNKNOWN;
+ break;
+ }
+
+ ch = config->info.dimension[0];
+ rate = config->rate_n;
- if (a_info.format != GST_AUDIO_FORMAT_UNKNOWN) {
- const gchar *format_string = gst_audio_format_to_string (a_info.format);
+ if (format != GST_AUDIO_FORMAT_UNKNOWN) {
+ const gchar *format_string = gst_audio_format_to_string (format);
gst_caps_set_simple (caps, "format", G_TYPE_STRING, format_string, NULL);
}
- if (a_info.ch > 0) {
- gst_caps_set_simple (caps, "channels", G_TYPE_INT, a_info.ch, NULL);
+ if (ch > 0) {
+ gst_caps_set_simple (caps, "channels", G_TYPE_INT, ch, NULL);
}
- if (a_info.rate > 0) {
- gst_caps_set_simple (caps, "rate", G_TYPE_INT, a_info.rate, NULL);
+ if (rate > 0) {
+ gst_caps_set_simple (caps, "rate", G_TYPE_INT, rate, NULL);
}
return gst_caps_simplify (caps);
gst_tensordec_text_caps_from_config (GstTensorDec * self,
const GstTensorConfig * config)
{
- GstTensorTextInfo t_info;
-
g_return_val_if_fail (config != NULL, NULL);
- gst_tensor_text_info_from_config (&t_info, config);
-
- /** utf8 */
- g_return_val_if_fail (t_info.format != 1, NULL);
+ /**
+ * Set text format. Supposed utf8 if type is int8.
+ */
+ g_return_val_if_fail (config->info.type == _NNS_INT8, NULL);
return gst_caps_from_string (GST_TENSOR_TEXT_CAPS_STR);
}
g_return_val_if_fail (config != NULL, FALSE);
if (self->configured) {
- return gst_tensor_config_is_same (&self->tensor_config, config);
+ return gst_tensor_config_is_equal (&self->tensor_config, config);
}
/** not configured yet */
switch (self->output_type) {
case OUTPUT_VIDEO:
{
- GstTensorVideoInfo v_info;
+ GstVideoFormat format;
+ gint width;
+
+ switch (config.info.dimension[0]) {
+ case 1:
+ format = GST_VIDEO_FORMAT_GRAY8;
+ break;
+ case 3:
+ format = GST_VIDEO_FORMAT_RGB;
+ break;
+ case 4:
+ format = GST_VIDEO_FORMAT_BGRx;
+ break;
+ default:
+ format = GST_VIDEO_FORMAT_UNKNOWN;
+ break;
+ }
+ width = config.info.dimension[1];
- if (gst_tensor_video_info_from_config (&v_info, &config)) {
- if (gst_tensor_video_stride_padding_per_row (v_info.format, v_info.w)) {
- self->add_padding = TRUE;
- }
+ if (gst_tensor_video_stride_padding_per_row (format, width)) {
+ self->add_padding = TRUE;
}
break;
g_assert (self->add_padding);
g_assert (self->output_type == OUTPUT_VIDEO);
- size = offset = config->dimension[0] * config->dimension[1];
+ size = offset = config->info.dimension[0] * config->info.dimension[1];
if (offset % 4)
offset += 4 - (offset % 4);
- size_out = offset * config->dimension[2] * config->dimension[3];
+ size_out = offset * config->info.dimension[2] * config->info.dimension[3];
if (gst_buffer_get_size (outbuf) < size_out) {
gst_buffer_set_size (outbuf, size_out);
inptr = inInfo.data;
outptr = outInfo.data;
- for (d0 = 0; d0 < config->dimension[3]; d0++) {
+ for (d0 = 0; d0 < config->info.dimension[3]; d0++) {
g_assert (d0 == 0);
- for (row = 0; row < config->dimension[2]; row++) {
+ for (row = 0; row < config->info.dimension[2]; row++) {
memcpy (outptr + dest_idx, inptr + src_idx, size);
dest_idx += offset;
src_idx += size;
/** flag add_padding only for video */
g_assert (self->output_type == OUTPUT_VIDEO);
- offset = config->dimension[0] * config->dimension[1];
+ offset = config->info.dimension[0] * config->info.dimension[1];
if (offset % 4)
offset += 4 - (offset % 4);
- *othersize = offset * config->dimension[2] * config->dimension[3];
+ *othersize = offset * config->info.dimension[2] * config->info.dimension[3];
} else {
*othersize = size;
}
int i;
prop->inputMeta.num_tensors =
get_tensor_dimension (g_value_get_string (value),
- prop->inputMeta.dims, prop->inputMeta.ranks);
+ prop->inputMeta.dims);
for (i = 0; i < prop->inputMeta.num_tensors; i++) {
- g_assert (prop->inputMeta.ranks[i] > 0
- && prop->inputMeta.ranks[i] <= NNS_TENSOR_RANK_LIMIT);
silent_debug ("Input Prop: %d:%d:%d:%d Rank %d\n",
prop->inputMeta.dims[i][0], prop->inputMeta.dims[i][1],
prop->inputMeta.dims[i][2], prop->inputMeta.dims[i][3],
int i;
prop->outputMeta.num_tensors =
get_tensor_dimension (g_value_get_string (value),
- prop->outputMeta.dims, prop->outputMeta.ranks);
+ prop->outputMeta.dims);
for (i = 0; i < prop->outputMeta.num_tensors; i++) {
- g_assert (prop->outputMeta.ranks[i] > 0
- && prop->outputMeta.ranks[i] <= NNS_TENSOR_RANK_LIMIT);
silent_debug ("Output Prop: %d:%d:%d:%d Rank %d\n",
prop->outputMeta.dims[i][0], prop->outputMeta.dims[i][1],
prop->outputMeta.dims[i][2], prop->outputMeta.dims[i][3],
dim =
g_array_index (tensor_split->tensorseg, tensor_dim *,
tensor_split->num_srcpads);
- type = tensor_split->sink_tensor_conf.type;
+ type = tensor_split->sink_tensor_conf.info.type;
tensor_split->num_srcpads++;
dim = g_array_index (split->tensorseg, tensor_dim *, nth);
for (i = 0; i < NNS_TENSOR_RANK_LIMIT; i++)
temp *= (*dim)[i];
- size += temp * tensor_element_size[split->sink_tensor_conf.type];
+ size += temp * tensor_element_size[split->sink_tensor_conf.info.type];
mem = gst_allocator_alloc (NULL, size, NULL);
gst_memory_map (mem, &dest_info, GST_MAP_WRITE);
gst_buffer_map (buffer, &src_info, GST_MAP_READ);
"num_tensors = " GST_TENSOR_NUM_TENSORS_RANGE ", "\
"framerate = " GST_TENSOR_RATE_RANGE
/**
- * type should be one of
- * { float32, float64, int32, uint32, int16, uint16, int8, uint8 }
+ * type should be one of types in GST_TENSOR_TYPE_ALL
* "types = (string) uint8, uint8, uint8"
* Dimensions of Tensors for negotiation. It's comment out here,
but when we call gst_structure_get_string, it actually is working well
} media_type;
/**
- * @brief Internal data structure for video info to configure tensor.
+ * @brief Internal data structure for configured tensor info (for other/tensor).
*/
typedef struct
{
- GstVideoFormat format; /**< video format */
- gint w; /**< width */
- gint h; /**< height */
- gint fn; /**< framerate numerator */
- gint fd; /**< framerate denominator */
- gint frames; /**< number of frames per tensor */
-} GstTensorVideoInfo;
-
-/**
- * @brief Internal data structure for audio info to configure tensor.
- */
-typedef struct
-{
- GstAudioFormat format; /**< audio format */
- gint ch; /**< channels */
- gint rate; /**< rate */
- gint frames; /**< samples per tensor */
-} GstTensorAudioInfo;
-
-/**
- * @brief Internal data structure for text info to configure tensor.
- */
-typedef struct
-{
- gint format; /**< text format (0:unknown, 1:utf8) */
- gint size; /**< string length (now it is fixed size GST_TENSOR_STRING_SIZE) */
- gint frames; /**< number of frames per tensor */
-} GstTensorTextInfo;
+ GstTensorInfo info; /**< tensor info*/
+ gint rate_n; /**< framerate is in fraction, which is numerator/denominator */
+ gint rate_d; /**< framerate is in fraction, which is numerator/denominator */
+} GstTensorConfig;
/**
- * @brief Internal data structure for configured tensor info.
+ * @brief Internal data structure for configured tensors info (for other/tensors).
*/
typedef struct
{
- tensor_type type; /**< Type of each element in the tensor. User must designate this. Otherwise, this is UINT8 for video/x-raw byte stream */
- tensor_dim dimension; /**< Dimensions. We support up to 4th ranks. */
+ GstTensorsInfo info; /**< tensor info*/
gint rate_n; /**< framerate is in fraction, which is numerator/denominator */
gint rate_d; /**< framerate is in fraction, which is numerator/denominator */
-} GstTensorConfig;
+} GstTensorsConfig;
/**
* @brief String representations for each tensor element type.
gst_tensor_media_type_from_structure (const GstStructure * structure);
/**
- * @brief Initialize the video info structure
- * @param v_info video info structure to be initialized
- */
-extern void
-gst_tensor_video_info_init (GstTensorVideoInfo * v_info);
-
-/**
- * @brief Initialize the audio info structure
- * @param a_info audio info structure to be initialized
- */
-extern void
-gst_tensor_audio_info_init (GstTensorAudioInfo * a_info);
-
-/**
- * @brief Initialize the text info structure
- * @param t_info text info structure to be initialized
+ * @brief Initialize the tensor info structure
+ * @param info tensor info structure to be initialized
*/
extern void
-gst_tensor_text_info_init (GstTensorTextInfo * t_info);
+gst_tensor_info_init (GstTensorInfo * info);
/**
- * @brief Set video info to configure tensor
- * @param v_info video info structure to be filled
- * @param structure caps structure
- * @note Fill frames in GstTensorVideoInfo after calling this function.
+ * @brief Check the tensor info is valid
+ * @param info tensor info structure
+ * @return TRUE if info is valid
*/
-extern void
-gst_tensor_video_info_from_structure (GstTensorVideoInfo * v_info,
- const GstStructure * structure);
+extern gboolean
+gst_tensor_info_validate (const GstTensorInfo * info);
/**
- * @brief Set audio info to configure tensor
- * @param a_info audio info structure to be filled
- * @param structure caps structure
- * @note Fill frames in GstTensorAudioInfo after calling this function.
+ * @brief Compare tensor info
+ * @param TRUE if equal
*/
-extern void
-gst_tensor_audio_info_from_structure (GstTensorAudioInfo * a_info,
- const GstStructure * structure);
+extern gboolean
+gst_tensor_info_is_equal (const GstTensorInfo * i1, const GstTensorInfo * i2);
/**
- * @brief Set text info to configure tensor
- * @param t_info text info structure to be filled
- * @param structure caps structure
- * @note Fill size and frames in GstTensorTextInfo after calling this function.
+ * @brief Initialize the tensors info structure
+ * @param info tensors info structure to be initialized
*/
extern void
-gst_tensor_text_info_from_structure (GstTensorTextInfo * t_info,
- const GstStructure * structure);
-
-/**
- * @brief Set the video info structure from tensor config
- * @param v_info video info structure to be filled
- * @param config tensor config structure to be interpreted
- * @note We cannot get the exact media info from tensor config, you have to check media info after calling this function.
- * @return TRUE if no error
- */
-extern gboolean
-gst_tensor_video_info_from_config (GstTensorVideoInfo * v_info,
- const GstTensorConfig * config);
+gst_tensors_info_init (GstTensorsInfo * info);
/**
- * @brief Set the audio info structure from tensor config
- * @param a_info audio info structure to be filled
- * @param config tensor config structure to be interpreted
- * @note We cannot get the exact media info from tensor config, you have to check media info after calling this function.
- * @return TRUE if no error
+ * @brief Check the tensors info is valid
+ * @param info tensors info structure
+ * @return TRUE if info is valid
*/
extern gboolean
-gst_tensor_audio_info_from_config (GstTensorAudioInfo * a_info,
- const GstTensorConfig * config);
+gst_tensors_info_validate (const GstTensorsInfo * info);
/**
- * @brief Set the text info structure from tensor config
- * @param t_info text info structure to be filled
- * @param config tensor config structure to be interpreted
- * @note We cannot get the exact media info from tensor config, you have to check media info after calling this function.
- * @return TRUE if no error
+ * @brief Compare tensors info
+ * @param TRUE if equal
*/
extern gboolean
-gst_tensor_text_info_from_config (GstTensorTextInfo * t_info,
- const GstTensorConfig * config);
+gst_tensors_info_is_equal (const GstTensorsInfo * i1, const GstTensorsInfo * i2);
/**
* @brief Initialize the tensor config info structure
/**
* @brief Compare tensor config info
- * @param TRUE if same
+ * @param TRUE if equal
*/
extern gboolean
-gst_tensor_config_is_same (const GstTensorConfig * c1,
+gst_tensor_config_is_equal (const GstTensorConfig * c1,
const GstTensorConfig * c2);
/**
- * @brief Parse structure and set tensor config info
+ * @brief Parse structure and set tensor config info (for other/tensor)
* @param config tensor config structure to be filled
* @param structure structure to be interpreted
* @note Change dimention if tensor contains N frames.
const GstStructure * structure);
/**
- * @brief Set the tensor config structure from video info
- * @param config tensor config structure to be filled
- * @param v_info video info structure to be interpreted
- * @note Change dimention if tensor contains N frames.
- * @return TRUE if supported type
+ * @brief Get tensor caps from tensor config (for other/tensor)
+ * @param config tensor config info
+ * @return caps for given config
+ */
+extern GstCaps *
+gst_tensor_caps_from_config (const GstTensorConfig * config);
+
+/**
+ * @brief Initialize the tensors config info structure (for other/tensors)
+ * @param config tensors config structure to be initialized
+ */
+extern void
+gst_tensors_config_init (GstTensorsConfig * config);
+
+/**
+ * @brief Check the tensors are all configured (for other/tensors)
+ * @param config tensor config structure
+ * @return TRUE if configured
*/
extern gboolean
-gst_tensor_config_from_video_info (GstTensorConfig * config,
- const GstTensorVideoInfo * v_info);
+gst_tensors_config_validate (const GstTensorsConfig * config);
/**
- * @brief Set the tensor config structure from audio info
- * @param config tensor config structure to be filled
- * @param a_info audio info structure to be interpreted
- * @note Change dimention if tensor contains N frames.
- * @return TRUE if supported type
+ * @brief Compare tensor config info (for other/tensors)
+ * @param TRUE if equal
*/
extern gboolean
-gst_tensor_config_from_audio_info (GstTensorConfig * config,
- const GstTensorAudioInfo * a_info);
+gst_tensors_config_is_equal (const GstTensorsConfig * c1,
+ const GstTensorsConfig * c2);
/**
- * @brief Set the tensor config structure from text info
- * @param config tensor config structure to be filled
- * @param t_info text info structure to be interpreted
- * @note Change dimention if tensor contains N frames.
- * @return TRUE if supported type
+ * @brief Parse structure and set tensors config (for other/tensors)
+ * @param config tensors config structure to be filled
+ * @param structure structure to be interpreted
+ * @return TRUE if no error
*/
extern gboolean
-gst_tensor_config_from_text_info (GstTensorConfig * config,
- const GstTensorTextInfo * t_info);
+gst_tensors_config_from_structure (GstTensorsConfig * config,
+ const GstStructure * structure);
/**
- * @brief Get tensor caps from tensor config
- * @param config tensor config info
+ * @brief Get caps from tensors config (for other/tensors)
+ * @param config tensors config info
* @return caps for given config
*/
extern GstCaps *
-gst_tensor_caps_from_config (const GstTensorConfig * config);
+gst_tensors_caps_from_config (const GstTensorsConfig * config);
/**
* @brief Determine if we need zero-padding
* @param param The parameter string in the format of d1:d2:d3:d4, d1:d2:d3, d1:d2, or d1, where dN is a positive integer and d1 is the innermost dimension; i.e., dim[d4][d3][d2][d1];
*/
extern int get_tensor_dimension (const gchar * param,
- uint32_t dim[NNS_TENSOR_SIZE_LIMIT][NNS_TENSOR_RANK_LIMIT],
- int rank[NNS_TENSOR_SIZE_LIMIT]);
+ uint32_t dim[NNS_TENSOR_SIZE_LIMIT][NNS_TENSOR_RANK_LIMIT]);
/**
* @brief Count the number of elemnts of a tensor
/**
* @brief Internal meta data exchange format for a other/tensors instance
+ * @todo replace this to GstTensorsInfo
*/
typedef struct
{
} GstTensor_TensorsMeta;
/**
+ * @brief Internal data structure for tensor info.
+ */
+typedef struct
+{
+ tensor_type type; /**< Type of each element in the tensor. User must designate this. */
+ tensor_dim dimension; /**< Dimension. We support up to 4th ranks. */
+} GstTensorInfo;
+
+/**
+ * @brief Internal meta data exchange format for a other/tensors instance
+ */
+typedef struct
+{
+ unsigned int num_tensors; /**< The number of tensors */
+ GstTensorInfo info[NNS_TENSOR_SIZE_LIMIT]; /**< The list of tensor info */
+} GstTensorsInfo;
+
+/**
* @brief Tensor_Filter's properties (internal data structure)
*
* Because custom filters of tensor_filter may need to access internal data
TEST (common_get_tensor_dimension, case1)
{
uint32_t dim[NNS_TENSOR_SIZE_LIMIT][NNS_TENSOR_RANK_LIMIT];
- int rank[NNS_TENSOR_RANK_LIMIT];
- int num_tensors = get_tensor_dimension ("345:123:433:177", dim, rank);
+ int num_tensors = get_tensor_dimension ("345:123:433:177", dim);
EXPECT_EQ (num_tensors, 1);
- EXPECT_EQ (rank[0], 4);
EXPECT_EQ (dim[0][0], 345);
EXPECT_EQ (dim[0][1], 123);
EXPECT_EQ (dim[0][2], 433);
TEST (common_get_tensor_dimension, case2)
{
uint32_t dim[NNS_TENSOR_SIZE_LIMIT][NNS_TENSOR_RANK_LIMIT];
- int rank[NNS_TENSOR_RANK_LIMIT];
- int num_tensors = get_tensor_dimension ("345:123:433", dim, rank);
+ int num_tensors = get_tensor_dimension ("345:123:433", dim);
EXPECT_EQ (num_tensors, 1);
- EXPECT_EQ (rank[0], 3);
EXPECT_EQ (dim[0][0], 345);
EXPECT_EQ (dim[0][1], 123);
EXPECT_EQ (dim[0][2], 433);
TEST (common_get_tensor_dimension, case3)
{
uint32_t dim[NNS_TENSOR_SIZE_LIMIT][NNS_TENSOR_RANK_LIMIT];
- int rank[NNS_TENSOR_RANK_LIMIT];
- int num_tensors = get_tensor_dimension ("345:123", dim, rank);
+ int num_tensors = get_tensor_dimension ("345:123", dim);
EXPECT_EQ (num_tensors, 1);
- EXPECT_EQ (rank[0], 2);
EXPECT_EQ (dim[0][0], 345);
EXPECT_EQ (dim[0][1], 123);
EXPECT_EQ (dim[0][2], 1);
TEST (common_get_tensor_dimension, case4)
{
uint32_t dim[NNS_TENSOR_SIZE_LIMIT][NNS_TENSOR_RANK_LIMIT];
- int rank[NNS_TENSOR_RANK_LIMIT];
- int num_tensors = get_tensor_dimension ("345", dim, rank);
+ int num_tensors = get_tensor_dimension ("345", dim);
EXPECT_EQ (num_tensors, 1);
- EXPECT_EQ (rank[0], 1);
EXPECT_EQ (dim[0][0], 345);
EXPECT_EQ (dim[0][1], 1);
EXPECT_EQ (dim[0][2], 1);
/** check tensor config for video */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_UINT8);
- EXPECT_EQ (g_test_data.config.dimension[0], 3);
- EXPECT_EQ (g_test_data.config.dimension[1], 160);
- EXPECT_EQ (g_test_data.config.dimension[2], 120);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_UINT8);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 3);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 160);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 120);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 30);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for video */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_UINT8);
- EXPECT_EQ (g_test_data.config.dimension[0], 3);
- EXPECT_EQ (g_test_data.config.dimension[1], 162);
- EXPECT_EQ (g_test_data.config.dimension[2], 120);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_UINT8);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 3);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 162);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 120);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 30);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for video */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_UINT8);
- EXPECT_EQ (g_test_data.config.dimension[0], 3);
- EXPECT_EQ (g_test_data.config.dimension[1], 160);
- EXPECT_EQ (g_test_data.config.dimension[2], 120);
- EXPECT_EQ (g_test_data.config.dimension[3], 3);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_UINT8);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 3);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 160);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 120);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 3);
EXPECT_EQ (g_test_data.config.rate_n, 30);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for video */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_UINT8);
- EXPECT_EQ (g_test_data.config.dimension[0], 4);
- EXPECT_EQ (g_test_data.config.dimension[1], 160);
- EXPECT_EQ (g_test_data.config.dimension[2], 120);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_UINT8);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 4);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 160);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 120);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 30);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for video */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_UINT8);
- EXPECT_EQ (g_test_data.config.dimension[0], 4);
- EXPECT_EQ (g_test_data.config.dimension[1], 160);
- EXPECT_EQ (g_test_data.config.dimension[2], 120);
- EXPECT_EQ (g_test_data.config.dimension[3], 2);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_UINT8);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 4);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 160);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 120);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 2);
EXPECT_EQ (g_test_data.config.rate_n, 30);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for video */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_UINT8);
- EXPECT_EQ (g_test_data.config.dimension[0], 1);
- EXPECT_EQ (g_test_data.config.dimension[1], 160);
- EXPECT_EQ (g_test_data.config.dimension[2], 120);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_UINT8);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 160);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 120);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 30);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for video */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_UINT8);
- EXPECT_EQ (g_test_data.config.dimension[0], 1);
- EXPECT_EQ (g_test_data.config.dimension[1], 162);
- EXPECT_EQ (g_test_data.config.dimension[2], 120);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_UINT8);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 162);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 120);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 30);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for video */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_UINT8);
- EXPECT_EQ (g_test_data.config.dimension[0], 1);
- EXPECT_EQ (g_test_data.config.dimension[1], 162);
- EXPECT_EQ (g_test_data.config.dimension[2], 120);
- EXPECT_EQ (g_test_data.config.dimension[3], 3);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_UINT8);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 162);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 120);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 3);
EXPECT_EQ (g_test_data.config.rate_n, 30);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for audio */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_INT8);
- EXPECT_EQ (g_test_data.config.dimension[0], 1);
- EXPECT_EQ (g_test_data.config.dimension[1], 500);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_INT8);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 500);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 16000);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for audio */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_UINT8);
- EXPECT_EQ (g_test_data.config.dimension[0], 1);
- EXPECT_EQ (g_test_data.config.dimension[1], 100);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_UINT8);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 100);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 16000);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for audio */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_INT16);
- EXPECT_EQ (g_test_data.config.dimension[0], 1);
- EXPECT_EQ (g_test_data.config.dimension[1], 500);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_INT16);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 500);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 16000);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for audio */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_UINT16);
- EXPECT_EQ (g_test_data.config.dimension[0], 1);
- EXPECT_EQ (g_test_data.config.dimension[1], 1000);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_UINT16);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 1000);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 16000);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for text */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_INT8);
- EXPECT_EQ (g_test_data.config.dimension[0], GST_TENSOR_STRING_SIZE);
- EXPECT_EQ (g_test_data.config.dimension[1], 1);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_INT8);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], GST_TENSOR_STRING_SIZE);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 0);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for text */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_INT8);
- EXPECT_EQ (g_test_data.config.dimension[0], GST_TENSOR_STRING_SIZE);
- EXPECT_EQ (g_test_data.config.dimension[1], 3);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_INT8);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], GST_TENSOR_STRING_SIZE);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 3);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 0);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for text */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, t_type);
- EXPECT_EQ (g_test_data.config.dimension[0], GST_TENSOR_STRING_SIZE);
- EXPECT_EQ (g_test_data.config.dimension[1], 1);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, t_type);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], GST_TENSOR_STRING_SIZE);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 0);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for text */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, t_type);
- EXPECT_EQ (g_test_data.config.dimension[0], GST_TENSOR_STRING_SIZE);
- EXPECT_EQ (g_test_data.config.dimension[1], 1);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, t_type);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], GST_TENSOR_STRING_SIZE);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 0);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for text */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, t_type);
- EXPECT_EQ (g_test_data.config.dimension[0], GST_TENSOR_STRING_SIZE);
- EXPECT_EQ (g_test_data.config.dimension[1], 1);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, t_type);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], GST_TENSOR_STRING_SIZE);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 0);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for text */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, t_type);
- EXPECT_EQ (g_test_data.config.dimension[0], GST_TENSOR_STRING_SIZE);
- EXPECT_EQ (g_test_data.config.dimension[1], 1);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, t_type);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], GST_TENSOR_STRING_SIZE);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 0);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for text */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, t_type);
- EXPECT_EQ (g_test_data.config.dimension[0], GST_TENSOR_STRING_SIZE);
- EXPECT_EQ (g_test_data.config.dimension[1], 1);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, t_type);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], GST_TENSOR_STRING_SIZE);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 0);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for text */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, t_type);
- EXPECT_EQ (g_test_data.config.dimension[0], GST_TENSOR_STRING_SIZE);
- EXPECT_EQ (g_test_data.config.dimension[1], 1);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, t_type);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], GST_TENSOR_STRING_SIZE);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 0);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for text */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, t_type);
- EXPECT_EQ (g_test_data.config.dimension[0], GST_TENSOR_STRING_SIZE);
- EXPECT_EQ (g_test_data.config.dimension[1], 1);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, t_type);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], GST_TENSOR_STRING_SIZE);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 0);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for text */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, t_type);
- EXPECT_EQ (g_test_data.config.dimension[0], GST_TENSOR_STRING_SIZE);
- EXPECT_EQ (g_test_data.config.dimension[1], 1);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, t_type);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], GST_TENSOR_STRING_SIZE);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 0);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for video */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_UINT8);
- EXPECT_EQ (g_test_data.config.dimension[0], 3);
- EXPECT_EQ (g_test_data.config.dimension[1], 160);
- EXPECT_EQ (g_test_data.config.dimension[2], 120);
- EXPECT_EQ (g_test_data.config.dimension[3], 10);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_UINT8);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 3);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 160);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 120);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 10);
EXPECT_EQ (g_test_data.config.rate_n, 30);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for audio */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_INT16);
- EXPECT_EQ (g_test_data.config.dimension[0], 1);
- EXPECT_EQ (g_test_data.config.dimension[1], 2000);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_INT16);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 2000);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 16000);
EXPECT_EQ (g_test_data.config.rate_d, 1);
/** check tensor config for audio */
EXPECT_TRUE (gst_tensor_config_validate (&g_test_data.config));
- EXPECT_EQ (g_test_data.config.type, _NNS_UINT16);
- EXPECT_EQ (g_test_data.config.dimension[0], 1);
- EXPECT_EQ (g_test_data.config.dimension[1], 100);
- EXPECT_EQ (g_test_data.config.dimension[2], 1);
- EXPECT_EQ (g_test_data.config.dimension[3], 1);
+ EXPECT_EQ (g_test_data.config.info.type, _NNS_UINT16);
+ EXPECT_EQ (g_test_data.config.info.dimension[0], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[1], 100);
+ EXPECT_EQ (g_test_data.config.info.dimension[2], 1);
+ EXPECT_EQ (g_test_data.config.info.dimension[3], 1);
EXPECT_EQ (g_test_data.config.rate_n, 16000);
EXPECT_EQ (g_test_data.config.rate_d, 1);
projects / platform / upstream / nnstreamer.git / commitdiff