@@ -28,6 +28,7 @@
#include "dnn_backend_native_layer_conv2d.h"
#include "dnn_backend_native_layer_depth2space.h"
#include "libavformat/avio.h"
+#include "libavformat/internal.h"
#include "libavutil/avassert.h"
#include "../internal.h"
#include "dnn_backend_native_layer_pad.h"
@@ -202,35 +203,24 @@ static DNNReturnType load_tf_model(TFModel *tf_model, const char *model_filename
TF_SessionOptions *sess_opts;
const TF_Operation *init_op;
uint8_t *sess_config = NULL;
- int sess_config_length = 0;
+ int sess_config_length;
// prepare the sess config data
if (tf_model->ctx.options.sess_config != NULL) {
+ const char *config;
/*
tf_model->ctx.options.sess_config is hex to present the serialized proto
required by TF_SetConfig below, so we need to first generate the serialized
- proto in a python script, the following is a script example to generate
- serialized proto which specifies one GPU, we can change the script to add
- more options.
-
- import tensorflow as tf
- gpu_options = tf.GPUOptions(visible_device_list='0')
- config = tf.ConfigProto(gpu_options=gpu_options)
- s = config.SerializeToString()
- b = ''.join("%02x" % int(ord(b)) for b in s[::-1])
- print('0x%s' % b)
-
- the script output looks like: 0xab...cd, and then pass 0xab...cd to sess_config.
+ proto in a python script, tools/python/tf_sess_config.py is a script example
+ to generate the configs of sess_config.
*/
- char tmp[3];
- tmp[2] = '\0';
-
if (strncmp(tf_model->ctx.options.sess_config, "0x", 2) != 0) {
av_log(ctx, AV_LOG_ERROR, "sess_config should start with '0x'\n");
return DNN_ERROR;
}
+ config = tf_model->ctx.options.sess_config + 2;
+ sess_config_length = ff_hex_to_data(NULL, config);
- sess_config_length = strlen(tf_model->ctx.options.sess_config);
if (sess_config_length % 2 != 0) {
av_log(ctx, AV_LOG_ERROR, "the length of sess_config is not even (%s), "
"please re-generate the config.\n",
@@ -238,21 +228,12 @@ static DNNReturnType load_tf_model(TFModel *tf_model, const char *model_filename
return DNN_ERROR;
}
- sess_config_length -= 2; //ignore the first '0x'
- sess_config_length /= 2; //get the data length in byte
-
- sess_config = av_malloc(sess_config_length);
+ sess_config = av_mallocz(sess_config_length + AV_INPUT_BUFFER_PADDING_SIZE);
if (!sess_config) {
av_log(ctx, AV_LOG_ERROR, "failed to allocate memory\n");
return DNN_ERROR;
}
-
- for (int i = 0; i < sess_config_length; i++) {
- int index = 2 + (sess_config_length - 1 - i) * 2;
- tmp[0] = tf_model->ctx.options.sess_config[index];
- tmp[1] = tf_model->ctx.options.sess_config[index + 1];
- sess_config[i] = strtol(tmp, NULL, 16);
- }
+ ff_hex_to_data(sess_config, config);
}
graph_def = read_graph(model_filename);
new file mode 100644
@@ -0,0 +1,44 @@
+# Copyright (c) 2021
+#
+# This file is part of FFmpeg.
+#
+# FFmpeg is free software; you can redistribute it and/or
+# modify it under the terms of the GNU Lesser General Public
+# License as published by the Free Software Foundation; either
+# version 2.1 of the License, or (at your option) any later version.
+#
+# FFmpeg is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
+# Lesser General Public License for more details.
+#
+# You should have received a copy of the GNU Lesser General Public
+# License along with FFmpeg; if not, write to the Free Software
+# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
+# ==============================================================================
+
+# verified with Python 3.6.8 on CentOS 7.2
+import tensorflow as tf
+
+visible_device_list = '0' # use , separator for more GPUs like '0, 1'
+per_process_gpu_memory_fraction = 0.9 # avoid out of memory
+intra_op_parallelism_threads = 2 # default in tensorflow
+inter_op_parallelism_threads = 5 # default in tensorflow
+
+gpu_options = tf.compat.v1.GPUOptions(
+ per_process_gpu_memory_fraction = per_process_gpu_memory_fraction,
+ visible_device_list = visible_device_list,
+ allow_growth = True)
+
+config = tf.compat.v1.ConfigProto(
+ allow_soft_placement = True,
+ log_device_placement = False,
+ intra_op_parallelism_threads = intra_op_parallelism_threads,
+ inter_op_parallelism_threads = inter_op_parallelism_threads,
+ gpu_options = gpu_options)
+
+s = config.SerializeToString()
+# print(list(map(hex, s))) # print by json if need
+
+b = ''.join(format(b,'02x') for b in s)
+print('0x%s' % b) # print by hex format