> -----Original Message-----
> From: ffmpeg-devel [mailto:[email protected]] On Behalf Of
> Xuewei Meng
> Sent: Tuesday, May 21, 2019 2:53 PM
> To: [email protected]
> Subject: [FFmpeg-devel] [PATCH] libavfilter/dnn_native: Add multiple
> activation
> functions in dnn native
>
> Add "Leaky_relu" and "None" option in activation function.
>
> Signed-off-by: Xuewei Meng <[email protected]>
> ---
> libavfilter/dnn_backend_native.c | 5 +++++
> libavfilter/dnn_backend_native.h | 2 +-
> 2 files changed, 6 insertions(+), 1 deletion(-)
LGTM
>
> diff --git a/libavfilter/dnn_backend_native.c
> b/libavfilter/dnn_backend_native.c
> index fa3f20e4a2..3c8465a283 100644
> --- a/libavfilter/dnn_backend_native.c
> +++ b/libavfilter/dnn_backend_native.c
> @@ -270,6 +270,11 @@ static void convolve(const float *input, float *output,
> const ConvolutionalParam
> break;
> case SIGMOID:
> output[n_filter] = 1.0f / (1.0f +
> exp(-output[n_filter]));
> + break;
> + case NONE:
> + break;
> + case LEAKY_RELU:
> + output[n_filter] = FFMAX(output[n_filter], 0.0) + 0.2 *
> FFMIN(output[n_filter], 0.0);
> }
> }
> output += conv_params->output_num;
> diff --git a/libavfilter/dnn_backend_native.h
> b/libavfilter/dnn_backend_native.h
> index d70cd16387..7e4e943137 100644
> --- a/libavfilter/dnn_backend_native.h
> +++ b/libavfilter/dnn_backend_native.h
> @@ -32,7 +32,7 @@
>
> typedef enum {INPUT, CONV, DEPTH_TO_SPACE} DNNLayerType;
>
> -typedef enum {RELU, TANH, SIGMOID} DNNActivationFunc;
> +typedef enum {RELU, TANH, SIGMOID, NONE, LEAKY_RELU}
> DNNActivationFunc;
>
> typedef enum {VALID, SAME, SAME_CLAMP_TO_EDGE}
> DNNConvPaddingParam;
>
> --
> 2.17.1
>
> _______________________________________________
> ffmpeg-devel mailing list
> [email protected]
> https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
>
> To unsubscribe, visit link above, or email
> [email protected] with subject "unsubscribe".
_______________________________________________
ffmpeg-devel mailing list
[email protected]
https://ffmpeg.org/mailman/listinfo/ffmpeg-devel
To unsubscribe, visit link above, or email
[email protected] with subject "unsubscribe".
