post-training quantization tensorflow model to float16

tf_model_post_training_quantization.py

import tensorflow as tf
from tensorflow.core.framework import types_pb2, graph_pb2, attr_value_pb2
from tensorflow.tools.graph_transforms import TransformGraph
from google.protobuf import text_format
import numpy as np

# object detection api input and output nodes
input_name = "image_tensor"
output_names = ["detection_boxes", "detection_classes", "detection_scores", "num_detections"]
# Const should be float32 in object detection api during nms (see here: https://www.tensorflow.org/api_docs/cc/class/tensorflow/ops/non-max-suppression-v4.html)
keep_fp32_node_name = ["Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/non_max_suppression/iou_threshold",
                       "Postprocessor/BatchMultiClassNonMaxSuppression/MultiClassNonMaxSuppression/non_max_suppression/score_threshold"]

def load_graph(model_path):
    graph = tf.Graph()
    with graph.as_default():
        graph_def = tf.GraphDef()
        if model_path.endswith("pb"):
            with open(model_path, "rb") as f:
                graph_def.ParseFromString(f.read())
        else:
            with open(model_path, "r") as pf:
                text_format.Parse(pf.read(), graph_def)
        tf.import_graph_def(graph_def, name="")
        sess = tf.Session(graph=graph)
        return sess

def rewrite_batch_norm_node_v2(node, graph_def, target_type='fp16'):
    """
    Rewrite FusedBatchNorm with FusedBatchNormV2 for reserve_space_1 and reserve_space_2 in FusedBatchNorm require float32 for 
    gradient calculation (See here: https://www.tensorflow.org/api_docs/cc/class/tensorflow/ops/fused-batch-norm)
    """
    if target_type == 'fp16':
        dtype = types_pb2.DT_HALF
    elif target_type == 'fp64':
        dtype = types_pb2.DT_DOUBLE
    else:
        dtype = types_pb2.DT_FLOAT
    new_node = graph_def.node.add()
    new_node.op = "FusedBatchNormV2"
    new_node.name = node.name
    new_node.input.extend(node.input)
    new_node.attr["U"].CopyFrom(attr_value_pb2.AttrValue(type=types_pb2.DT_FLOAT))
    for attr in list(node.attr.keys()):
        if attr == "T":
            node.attr[attr].type = dtype
        new_node.attr[attr].CopyFrom(node.attr[attr])
    print("rewrite fused_batch_norm done!")

def convert_graph_to_fp16(model_path, save_path, name, as_text=False, target_type='fp16', input_name=None, output_names=None):
    if target_type == 'fp16':
        dtype = types_pb2.DT_HALF
    elif target_type == 'fp64':
        dtype = types_pb2.DT_DOUBLE
    else:
        dtype = types_pb2.DT_FLOAT
    source_sess = load_graph(model_path)
    source_graph_def = source_sess.graph.as_graph_def()
    target_graph_def = graph_pb2.GraphDef()
    target_graph_def.versions.CopyFrom(source_graph_def.versions)
    for node in source_graph_def.node:
        # fused batch norm node
        if node.op == "FusedBatchNorm":
            rewrite_batch_norm_node_v2(node, target_graph_def, target_type=target_type)
            continue
        # replicate node
        new_node = target_graph_def.node.add()
        new_node.op = node.op
        new_node.name = node.name
        new_node.input.extend(node.input)
        attrs = list(node.attr.keys())
        # keep batch norm params node
        if ("BatchNorm" in node.name) or ('batch_normalization' in node.name):
            for attr in attrs:
                new_node.attr[attr].CopyFrom(node.attr[attr])
            continue
        # replace dtype in node attr with target dtype
        for attr in attrs:
            # keep special node in fp32
            if node.name in keep_fp32_node_name:
                new_node.attr[attr].CopyFrom(node.attr[attr])
                continue
            if node.attr[attr].type == types_pb2.DT_FLOAT:
                # modify node dtype
                new_node.attr[attr].type = dtype
            if attr == "value":
                tensor = node.attr[attr].tensor
                if tensor.dtype == types_pb2.DT_FLOAT:
                    # if float_val exists
                    if tensor.float_val:
                        float_val = tf.make_ndarray(node.attr[attr].tensor)
                        new_node.attr[attr].tensor.CopyFrom(tf.make_tensor_proto(float_val, dtype=dtype))
                        continue
                    # if tensor content exists
                    if tensor.tensor_content:
                        tensor_shape = [x.size for x in tensor.tensor_shape.dim]
                        tensor_weights = tf.make_ndarray(tensor)
                        # reshape tensor
                        tensor_weights = np.reshape(tensor_weights, tensor_shape)
                        tensor_proto = tf.make_tensor_proto(tensor_weights, dtype=dtype)
                        new_node.attr[attr].tensor.CopyFrom(tensor_proto)
                        continue
            new_node.attr[attr].CopyFrom(node.attr[attr])
    # transform graph
    if output_names:
        if not input_name:
            input_name = []
        transforms = ["strip_unused_nodes"]
        target_graph_def = TransformGraph(target_graph_def, input_name, output_names, transforms)
    # write graph_def to model
    tf.io.write_graph(target_graph_def, logdir=save_path, name=name, as_text=as_text)
    print("Converting done ...")

save_path = "test"
name = "test.pb"
as_text = False
target_type = 'fp16'
convert_graph_to_fp16(model_path, save_path, name, as_text=as_text, target_type=target_type, input_name=input_name, output_names=output_names)
# test loading
# ISSUE: loading detection model is extremely slow while loading classification model is normal
sess = load_graph(save_path+"/"+name)

My fault. Thanks for fixing this issue! @sayradley

Hi,fan,do you have any idea about add a new node into pb?

@longside I think you may need to the network structure file. This repo may help you

thanks, fan. Actually, I didnot supposed to add node while frozen graph.

I find it is work to use your method to add a new node in pb, LOL.
like this:

def add_cast_node(node, src_graph, graph_def, target_type='fp16', node_name='import/cast'):
    new_node = graph_def.node.add()
    new_node.op = "Cast"
    new_node.name = node_name
    new_node.input.append(node.input[0])
    new_node.attr["SrcT"].type = types_pb2.DT_INT32
    new_node.attr["DstT"].type = types_pb2.DT_INT64
    new_node.attr["Truncate"].b = False

    return new_node

ValueError: Input 0 of node InceptionResnetV1/Conv2d_1a_3x3/BatchNorm/cond/FusedBatchNorm was passed float from InceptionResnetV1/Conv2d_1a_3x3/BatchNorm/cond/FusedBatchNorm/Switch:1 incompatible with expected half.

im trying to convert pretrained model in facenet with input_names='input and output_names='embeddings' ,using tensorflow 1.14.0

@glennford49 As it said, try to keep the attributes of the Switch input node in the FusedBatchNorm, rather than being converted to ft16.

i have copied entirely your code but stil gives me same error, i will try to downgrade tf to 1.13.1

@glennford49 I'm not sure about this error. Could you provide your model?

I just downloaded the model 20180402-114759.pb trained in vggface2 in david sanberg github, i have used this model without converting to fp16 it works fine.. when i try to convert this using your code replacing the input name with input and ouput name with embeddings, but i get errors

ValueError: Input 0 of node InceptionResnetV1/Conv2d_1a_3x3/BatchNorm/cond/FusedBatchNorm was passed float from InceptionResnetV1/Conv2d_1a_3x3/BatchNorm/cond/FusedBatchNorm/Switch:1 incompatible with expected half.

im trying to convert pretrained model in facenet with input_names='input and output_names='embeddings' ,using tensorflow 1.14.0

same issue here. any idea to solve it ?

some bugs:
the fp16 is not more than tf.float16.max and makes some value be inf.
for the attr "SrcT","T","Tparams","DstT", this code doesn't make any change.