libavfiter/dnn_backend_openvino: Add multiple output support

Add multiple output support to openvino backend. You can use '&' to
split different output when you set output name using command line.

Signed-off-by: Wenbin Chen <wenbin.chen@intel.com>
Reviewed-by: Guo Yejun <yejun.guo@intel.com>

2 files changed, 147 insertions, 76 deletions

diff --git a/libavfilter/dnn/dnn_backend_common.c b/libavfilter/dnn/dnn_backend_common.c
index 91a4a3c4bf..632832ec36 100644
--- a/libavfilter/dnn/dnn_backend_common.c
+++ b/libavfilter/dnn/dnn_backend_common.c
@@ -43,13 +43,6 @@ int ff_check_exec_params(void *ctx, DNNBackendType backend, DNNFunctionType func
         return AVERROR(EINVAL);
     }
 
-    if (exec_params->nb_output != 1 && backend != DNN_TF) {
-        // currently, the filter does not need multiple outputs,
-        // so we just pending the support until we really need it.
-        avpriv_report_missing_feature(ctx, "multiple outputs");
-        return AVERROR(ENOSYS);
-    }
-
     return 0;
 }
 
diff --git a/libavfilter/dnn/dnn_backend_openvino.c b/libavfilter/dnn/dnn_backend_openvino.c
index 6fe8b9c243..089e028818 100644
--- a/libavfilter/dnn/dnn_backend_openvino.c
+++ b/libavfilter/dnn/dnn_backend_openvino.c
@@ -64,7 +64,7 @@ typedef struct OVModel{
     ov_compiled_model_t *compiled_model;
     ov_output_const_port_t* input_port;
     ov_preprocess_input_info_t* input_info;
-    ov_output_const_port_t* output_port;
+    ov_output_const_port_t** output_ports;
     ov_preprocess_output_info_t* output_info;
     ov_preprocess_prepostprocessor_t* preprocess;
 #else
@@ -77,6 +77,7 @@ typedef struct OVModel{
     SafeQueue *request_queue;   // holds OVRequestItem
     Queue *task_queue;          // holds TaskItem
     Queue *lltask_queue;     // holds LastLevelTaskItem
+    int nb_outputs;
 } OVModel;
 
 // one request for one call to openvino
@@ -349,7 +350,7 @@ static void infer_completion_callback(void *args)
     TaskItem *task = lltask->task;
     OVModel *ov_model = task->model;
     SafeQueue *requestq = ov_model->request_queue;
-    DNNData output;
+    DNNData *outputs;
     OVContext *ctx = &ov_model->ctx;
 #if HAVE_OPENVINO2
     size_t* dims;
@@ -358,45 +359,61 @@ static void infer_completion_callback(void *args)
     ov_shape_t output_shape = {0};
     ov_element_type_e precision;
 
-    memset(&output, 0, sizeof(output));
-    status = ov_infer_request_get_output_tensor_by_index(request->infer_request, 0, &output_tensor);
-    if (status != OK) {
-        av_log(ctx, AV_LOG_ERROR,
-               "Failed to get output tensor.");
+    outputs = av_calloc(ov_model->nb_outputs, sizeof(*outputs));
+    if (!outputs) {
+        av_log(ctx, AV_LOG_ERROR, "Failed to alloc outputs.");
         return;
     }
 
-    status = ov_tensor_data(output_tensor, &output.data);
-    if (status != OK) {
-        av_log(ctx, AV_LOG_ERROR,
-               "Failed to get output data.");
-        return;
-    }
+    for (int i = 0; i < ov_model->nb_outputs; i++) {
+        status = ov_infer_request_get_tensor_by_const_port(request->infer_request,
+                                                           ov_model->output_ports[i],
+                                                           &output_tensor);
+        if (status != OK) {
+            av_log(ctx, AV_LOG_ERROR,
+                "Failed to get output tensor.");
+            goto end;
+        }
 
-    status = ov_tensor_get_shape(output_tensor, &output_shape);
-    if (status != OK) {
-        av_log(ctx, AV_LOG_ERROR, "Failed to get output port shape.\n");
-        return;
-    }
-    dims = output_shape.dims;
+        status = ov_tensor_data(output_tensor, &outputs[i].data);
+        if (status != OK) {
+            av_log(ctx, AV_LOG_ERROR,
+                "Failed to get output data.");
+            goto end;
+        }
 
-    status = ov_port_get_element_type(ov_model->output_port, &precision);
-    if (status != OK) {
-        av_log(ctx, AV_LOG_ERROR, "Failed to get output port data type.\n");
+        status = ov_tensor_get_shape(output_tensor, &output_shape);
+        if (status != OK) {
+            av_log(ctx, AV_LOG_ERROR, "Failed to get output port shape.\n");
+            goto end;
+        }
+        dims = output_shape.dims;
+
+        status = ov_port_get_element_type(ov_model->output_ports[i], &precision);
+        if (status != OK) {
+            av_log(ctx, AV_LOG_ERROR, "Failed to get output port data type.\n");
+            goto end;
+        }
+        outputs[i].dt       = precision_to_datatype(precision);
+
+        outputs[i].channels = output_shape.rank > 2 ? dims[output_shape.rank - 3] : 1;
+        outputs[i].height   = output_shape.rank > 1 ? dims[output_shape.rank - 2] : 1;
+        outputs[i].width    = output_shape.rank > 0 ? dims[output_shape.rank - 1] : 1;
+        av_assert0(request->lltask_count <= dims[0]);
+        outputs[i].layout   = ctx->options.layout;
+        outputs[i].scale    = ctx->options.scale;
+        outputs[i].mean     = ctx->options.mean;
         ov_shape_free(&output_shape);
-        return;
+        ov_tensor_free(output_tensor);
+        output_tensor = NULL;
     }
-    output.channels = output_shape.rank > 2 ? dims[output_shape.rank - 3] : 1;
-    output.height   = output_shape.rank > 1 ? dims[output_shape.rank - 2] : 1;
-    output.width    = output_shape.rank > 0 ? dims[output_shape.rank - 1] : 1;
-    av_assert0(request->lltask_count <= dims[0]);
-    ov_shape_free(&output_shape);
 #else
     IEStatusCode status;
     dimensions_t dims;
     ie_blob_t *output_blob = NULL;
     ie_blob_buffer_t blob_buffer;
     precision_e precision;
+    DNNData output;
     status = ie_infer_request_get_blob(request->infer_request, task->output_names[0], &output_blob);
     if (status != OK) {
         av_log(ctx, AV_LOG_ERROR,
@@ -424,11 +441,12 @@ static void infer_completion_callback(void *args)
     output.height   = dims.dims[2];
     output.width    = dims.dims[3];
     av_assert0(request->lltask_count <= dims.dims[0]);
-#endif
     output.dt       = precision_to_datatype(precision);
     output.layout   = ctx->options.layout;
     output.scale    = ctx->options.scale;
     output.mean     = ctx->options.mean;
+    outputs = &output;
+#endif
 
     av_assert0(request->lltask_count >= 1);
     for (int i = 0; i < request->lltask_count; ++i) {
@@ -438,28 +456,33 @@ static void infer_completion_callback(void *args)
         case DFT_PROCESS_FRAME:
             if (task->do_ioproc) {
                 if (ov_model->model->frame_post_proc != NULL) {
-                    ov_model->model->frame_post_proc(task->out_frame, &output, ov_model->model->filter_ctx);
+                    ov_model->model->frame_post_proc(task->out_frame, outputs, ov_model->model->filter_ctx);
                 } else {
-                    ff_proc_from_dnn_to_frame(task->out_frame, &output, ctx);
+                    ff_proc_from_dnn_to_frame(task->out_frame, outputs, ctx);
                 }
             } else {
-                task->out_frame->width = output.width;
-                task->out_frame->height = output.height;
+                task->out_frame->width = outputs[0].width;
+                task->out_frame->height = outputs[0].height;
             }
             break;
         case DFT_ANALYTICS_DETECT:
             if (!ov_model->model->detect_post_proc) {
                 av_log(ctx, AV_LOG_ERROR, "detect filter needs to provide post proc\n");
-                return;
+                goto end;
             }
-            ov_model->model->detect_post_proc(task->in_frame, &output, 1, ov_model->model->filter_ctx);
+            ov_model->model->detect_post_proc(task->in_frame, outputs,
+                                              ov_model->nb_outputs,
+                                              ov_model->model->filter_ctx);
             break;
         case DFT_ANALYTICS_CLASSIFY:
             if (!ov_model->model->classify_post_proc) {
                 av_log(ctx, AV_LOG_ERROR, "classify filter needs to provide post proc\n");
-                return;
+                goto end;
             }
-            ov_model->model->classify_post_proc(task->in_frame, &output, request->lltasks[i]->bbox_index, ov_model->model->filter_ctx);
+            for (int output_i = 0; output_i < ov_model->nb_outputs; output_i++)
+                ov_model->model->classify_post_proc(task->in_frame, outputs,
+                                                    request->lltasks[i]->bbox_index,
+                                                    ov_model->model->filter_ctx);
             break;
         default:
             av_assert0(!"should not reach here");
@@ -468,10 +491,17 @@ static void infer_completion_callback(void *args)
 
         task->inference_done++;
         av_freep(&request->lltasks[i]);
-        output.data = (uint8_t *)output.data
-                      + output.width * output.height * output.channels * get_datatype_size(output.dt);
+        for (int i = 0; i < ov_model->nb_outputs; i++)
+            outputs[i].data = (uint8_t *)outputs[i].data +
+                outputs[i].width * outputs[i].height * outputs[i].channels * get_datatype_size(outputs[i].dt);
     }
-#if !HAVE_OPENVINO2
+end:
+#if HAVE_OPENVINO2
+    av_freep(&outputs);
+    ov_shape_free(&output_shape);
+    if (output_tensor)
+        ov_tensor_free(output_tensor);
+#else
     ie_blob_free(&output_blob);
 #endif
     request->lltask_count = 0;
@@ -525,8 +555,10 @@ static void dnn_free_model_ov(DNNModel **model)
 #if HAVE_OPENVINO2
     if (ov_model->input_port)
         ov_output_const_port_free(ov_model->input_port);
-    if (ov_model->output_port)
-        ov_output_const_port_free(ov_model->output_port);
+    for (int i = 0; i < ov_model->nb_outputs; i++)
+        if (ov_model->output_ports[i])
+            ov_output_const_port_free(ov_model->output_ports[i]);
+    av_freep(&ov_model->output_ports);
     if (ov_model->preprocess)
         ov_preprocess_prepostprocessor_free(ov_model->preprocess);
     if (ov_model->compiled_model)
@@ -551,7 +583,7 @@ static void dnn_free_model_ov(DNNModel **model)
 }
 
 
-static int init_model_ov(OVModel *ov_model, const char *input_name, const char *output_name)
+static int init_model_ov(OVModel *ov_model, const char *input_name, const char **output_names, int nb_outputs)
 {
     int ret = 0;
     OVContext *ctx = &ov_model->ctx;
@@ -594,17 +626,15 @@ static int init_model_ov(OVModel *ov_model, const char *input_name, const char *
     }
 
     status = ov_preprocess_prepostprocessor_get_input_info_by_name(ov_model->preprocess, input_name, &ov_model->input_info);
-    status |= ov_preprocess_prepostprocessor_get_output_info_by_name(ov_model->preprocess, output_name, &ov_model->output_info);
     if (status != OK) {
-        av_log(ctx, AV_LOG_ERROR, "Failed to get input/output info from preprocess.\n");
+        av_log(ctx, AV_LOG_ERROR, "Failed to get input info from preprocess.\n");
         ret = ov2_map_error(status, NULL);
         goto err;
     }
 
     status = ov_preprocess_input_info_get_tensor_info(ov_model->input_info, &input_tensor_info);
-    status |= ov_preprocess_output_info_get_tensor_info(ov_model->output_info, &output_tensor_info);
     if (status != OK) {
-        av_log(ctx, AV_LOG_ERROR, "Failed to get tensor info from input/output.\n");
+        av_log(ctx, AV_LOG_ERROR, "Failed to get tensor info from input.\n");
         ret = ov2_map_error(status, NULL);
         goto err;
     }
@@ -642,17 +672,43 @@ static int init_model_ov(OVModel *ov_model, const char *input_name, const char *
     }
 
     status = ov_preprocess_input_tensor_info_set_element_type(input_tensor_info, U8);
-    if (ov_model->model->func_type != DFT_PROCESS_FRAME)
-        status |= ov_preprocess_output_set_element_type(output_tensor_info, F32);
-    else if (fabsf(ctx->options.scale - 1) > 1e-6f || fabsf(ctx->options.mean) > 1e-6f)
-        status |= ov_preprocess_output_set_element_type(output_tensor_info, F32);
-    else
-        status |= ov_preprocess_output_set_element_type(output_tensor_info, U8);
     if (status != OK) {
-        av_log(ctx, AV_LOG_ERROR, "Failed to set input/output element type\n");
+        av_log(ctx, AV_LOG_ERROR, "Failed to set input element type\n");
         ret = ov2_map_error(status, NULL);
         goto err;
     }
+
+    ov_model->nb_outputs = nb_outputs;
+    for (int i = 0; i < nb_outputs; i++) {
+        status = ov_preprocess_prepostprocessor_get_output_info_by_name(
+                ov_model->preprocess, output_names[i], &ov_model->output_info);
+        if (status != OK) {
+            av_log(ctx, AV_LOG_ERROR, "Failed to get output info from preprocess.\n");
+            ret = ov2_map_error(status, NULL);
+            goto err;
+        }
+        status |= ov_preprocess_output_info_get_tensor_info(ov_model->output_info, &output_tensor_info);
+        if (status != OK) {
+            av_log(ctx, AV_LOG_ERROR, "Failed to get tensor info from input/output.\n");
+            ret = ov2_map_error(status, NULL);
+            goto err;
+        }
+        if (ov_model->model->func_type != DFT_PROCESS_FRAME)
+            status |= ov_preprocess_output_set_element_type(output_tensor_info, F32);
+        else if (fabsf(ctx->options.scale - 1) > 1e-6f || fabsf(ctx->options.mean) > 1e-6f)
+            status |= ov_preprocess_output_set_element_type(output_tensor_info, F32);
+        else
+            status |= ov_preprocess_output_set_element_type(output_tensor_info, U8);
+        if (status != OK) {
+            av_log(ctx, AV_LOG_ERROR, "Failed to set output element type\n");
+            ret = ov2_map_error(status, NULL);
+            goto err;
+        }
+        ov_preprocess_output_tensor_info_free(output_tensor_info);
+        output_tensor_info = NULL;
+        ov_preprocess_output_info_free(ov_model->output_info);
+        ov_model->output_info = NULL;
+    }
     // set preprocess steps.
     if (fabsf(ctx->options.scale - 1) > 1e-6f || fabsf(ctx->options.mean) > 1e-6f) {
         ov_preprocess_preprocess_steps_t* input_process_steps = NULL;
@@ -667,11 +723,18 @@ static int init_model_ov(OVModel *ov_model, const char *input_name, const char *
         status |= ov_preprocess_preprocess_steps_scale(input_process_steps, ctx->options.scale);
         if (status != OK) {
             av_log(ctx, AV_LOG_ERROR, "Failed to set preprocess steps\n");
+            ov_preprocess_preprocess_steps_free(input_process_steps);
+            input_process_steps = NULL;
             ret = ov2_map_error(status, NULL);
             goto err;
         }
         ov_preprocess_preprocess_steps_free(input_process_steps);
+        input_process_steps = NULL;
     }
+    ov_preprocess_input_tensor_info_free(input_tensor_info);
+    input_tensor_info = NULL;
+    ov_preprocess_input_info_free(ov_model->input_info);
+    ov_model->input_info = NULL;
 
     //update model
     if(ov_model->ov_model)
@@ -679,20 +742,33 @@ static int init_model_ov(OVModel *ov_model, const char *input_name, const char *
     status = ov_preprocess_prepostprocessor_build(ov_model->preprocess, &ov_model->ov_model);
     if (status != OK) {
         av_log(ctx, AV_LOG_ERROR, "Failed to update OV model\n");
+        ov_model_free(tmp_ov_model);
+        tmp_ov_model = NULL;
         ret = ov2_map_error(status, NULL);
         goto err;
     }
     ov_model_free(tmp_ov_model);
 
     //update output_port
-    if (ov_model->output_port) {
-        ov_output_const_port_free(ov_model->output_port);
-        ov_model->output_port = NULL;
-    }
-    status = ov_model_const_output_by_name(ov_model->ov_model, output_name, &ov_model->output_port);
-    if (status != OK) {
-        av_log(ctx, AV_LOG_ERROR, "Failed to get output port.\n");
-        goto err;
+    if (!ov_model->output_ports) {
+        ov_model->output_ports = av_calloc(nb_outputs, sizeof(*ov_model->output_ports));
+        if (!ov_model->output_ports) {
+            ret = AVERROR(ENOMEM);
+            goto err;
+        }
+    } else
+        for (int i = 0; i < nb_outputs; i++) {
+            ov_output_const_port_free(ov_model->output_ports[i]);
+            ov_model->output_ports[i] = NULL;
+        }
+
+    for (int i = 0; i < nb_outputs; i++) {
+        status = ov_model_const_output_by_name(ov_model->ov_model, output_names[i],
+                                               &ov_model->output_ports[i]);
+        if (status != OK) {
+            av_log(ctx, AV_LOG_ERROR, "Failed to get output port %s.\n", output_names[i]);
+            goto err;
+        }
     }
     //compile network
     status = ov_core_compile_model(ov_model->core, ov_model->ov_model, device, 0, &ov_model->compiled_model);
@@ -701,6 +777,7 @@ static int init_model_ov(OVModel *ov_model, const char *input_name, const char *
         goto err;
     }
     ov_preprocess_input_model_info_free(input_model_info);
+    input_model_info = NULL;
     ov_layout_free(NCHW_layout);
     ov_layout_free(NHWC_layout);
 #else
@@ -745,6 +822,7 @@ static int init_model_ov(OVModel *ov_model, const char *input_name, const char *
         ret = DNN_GENERIC_ERROR;
         goto err;
     }
+    ov_model->nb_outputs = 1;
 
     // all models in openvino open model zoo use BGR with range [0.0f, 255.0f] as input,
     // we don't have a AVPixelFormat to describe it, so we'll use AV_PIX_FMT_BGR24 and
@@ -848,6 +926,10 @@ static int init_model_ov(OVModel *ov_model, const char *input_name, const char *
 
 err:
 #if HAVE_OPENVINO2
+    if (output_tensor_info)
+        ov_preprocess_output_tensor_info_free(output_tensor_info);
+    if (ov_model->output_info)
+        ov_preprocess_output_info_free(ov_model->output_info);
     if (NCHW_layout)
         ov_layout_free(NCHW_layout);
     if (NHWC_layout)
@@ -1204,11 +1286,6 @@ static int get_output_ov(void *model, const char *input_name, int input_width, i
         }
     }
 
-    status = ov_model_const_output_by_name(ov_model->ov_model, output_name, &ov_model->output_port);
-    if (status != OK) {
-        av_log(ctx, AV_LOG_ERROR, "Failed to get output port.\n");
-        return ov2_map_error(status, NULL);
-    }
     if (!ov_model->compiled_model) {
 #else
     if (ctx->options.input_resizable) {
@@ -1224,7 +1301,7 @@ static int get_output_ov(void *model, const char *input_name, int input_width, i
     }
     if (!ov_model->exe_network) {
 #endif
-        ret = init_model_ov(ov_model, input_name, output_name);
+        ret = init_model_ov(ov_model, input_name, &output_name, 1);
         if (ret != 0) {
             av_log(ctx, AV_LOG_ERROR, "Failed init OpenVINO exectuable network or inference request\n");
             return ret;
@@ -1397,7 +1474,8 @@ static int dnn_execute_model_ov(const DNNModel *model, DNNExecBaseParams *exec_p
 #else
     if (!ov_model->exe_network) {
 #endif
-        ret = init_model_ov(ov_model, exec_params->input_name, exec_params->output_names[0]);
+        ret = init_model_ov(ov_model, exec_params->input_name,
+                            exec_params->output_names, exec_params->nb_output);
         if (ret != 0) {
             av_log(ctx, AV_LOG_ERROR, "Failed init OpenVINO exectuable network or inference request\n");
             return ret;