tensorflow/model-optimization

View on GitHub

A toolkit to optimize ML models for deployment for Keras and TensorFlow, including quantization and pruning.

Stars: 1.6k
Forks: 349
Open beginner issues: 1
Indexed issues: 201
Avg merge: 3d 2h
Merged PRs (30d): 1

Dominant language: Python
License: Apache-2.0
Last GitHub push: Aug 24, 2026
Latest indexed: Sep 19, 2026
Contributing guide: Contributing guide
Code of conduct: No code of conduct
Beginner labels: No beginner labels indexed

Beginner friendly Everything

Open All

201 open issues indexed

tf.quantization.quantize fails when converting to TFLite Open

Difficulty 4/5 3-5 days Newbie friendliness 38/100

tensorflow/model-optimization#840 · 2 comments · Opened Sep 25, 2021
QAT for object detection models? Open

feature request

tensorflow/model-optimization#838 · 4 comments · 1 assignee · Opened Sep 22, 2021
Can't run prunning with XNNPack Open

tensorflow/model-optimization#837 · 1 comment · 1 assignee · Opened Sep 17, 2021
Is it possible to have quantised fused_tanh (with add op ) via TFlite converter Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

tensorflow/model-optimization#833 · 2 comments · Opened Sep 14, 2021
perform inference after QAT Open

tensorflow/model-optimization#832 · 3 comments · 1 reaction · 1 assignee · Opened Sep 13, 2021
quantization not happening? Open

tensorflow/model-optimization#829 · 4 comments · 1 assignee · Opened Sep 12, 2021
Model with SeparableConvs not converting to QAT Open

bug

tensorflow/model-optimization#823 · 4 comments · 1 assignee · Opened Sep 6, 2021
Support `float16` input for `tf.quantization.fake_quant_*` ops Open

tensorflow/model-optimization#824 · 2 comments · 1 assignee · Opened Sep 6, 2021
Weight in fully connected layers don't follow tensorflow quantization spec (zero-point!=0) Open

tensorflow/model-optimization#822 · 3 comments · 1 assignee · Opened Sep 6, 2021
Quantization aware training on untrained model Open

Difficulty 5/5 Over a week Newbie friendliness 10/100

tensorflow/model-optimization#820 · Opened Sep 3, 2021
WARNING:tensorflow:AutoGraph could not transform <tensorflow_model_optimization.python.core.quantization.keras.quantizers.AllValuesQuantizer object at 0x7f92927f3f90> and will run it as-is. Open

bug

tensorflow/model-optimization#813 · 1 comment · 1 assignee · Opened Sep 2, 2021
modify_model_interface removing input float tensor is not completed, memory allocator occupied this dummy tensor Open

feature request

Difficulty 4/5 3-5 days Newbie friendliness 35/100

tensorflow/model-optimization#814 · 2 comments · Opened Sep 2, 2021
model written with tf2.6.0 (not using keras) cannot find instructions to do quantization aware training, only keras model based quantization example can be found Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

tensorflow/model-optimization#811 · 1 comment · Opened Sep 1, 2021
Depthwise convolution min_max dimension Open

tensorflow/model-optimization#808 · 2 comments · 1 assignee · Opened Aug 30, 2021
Activity Regularizer not working with quantization aware training (QAT) Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 35/100

tensorflow/model-optimization#802 · 1 comment · 1 reaction · Opened Aug 25, 2021
[RNN] Stateful LSTM can't be converted to TF Lite with Integer Quantization Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

tensorflow/model-optimization#803 · 21 comments · Opened Aug 25, 2021
Is there a way to do symmetric quantization for activations in post-training quantization Open

Difficulty 5/5 Over a week Newbie friendliness 20/100

tensorflow/model-optimization#801 · 1 comment · Opened Aug 25, 2021
[clustering] Possible wrong implementation of get_weight_from_layer Open

bug technique:clustering

tensorflow/model-optimization#799 · 6 comments · 2 assignees · Opened Aug 23, 2021
Quantization aware training with mask_rcnn model error: tensorflow.python.framework.errors_impl.NotFoundError: Key SecondStageFeatureExtractor_1/InceptionV2/Mixed_5a/Branch_0/Conv2d_0a_1x1/act_quant/max not found in checkpoint [[{{node save/RestoreV2}}]] Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

tensorflow/model-optimization#798 · 1 comment · Opened Aug 23, 2021
TFLite quantization, the int tflite model is much slower than the float tflite model Open

tensorflow/model-optimization#797 · 5 comments · 1 assignee · Opened Aug 20, 2021
Integer tflite model slow than float32 model Open

tensorflow/model-optimization#796 · 11 comments · 1 assignee · Opened Aug 19, 2021
Performance issue in the program Open

bug

Difficulty 2/5 1-3 hours Newbie friendliness 55/100

tensorflow/model-optimization#795 · 2 comments · Opened Aug 18, 2021
Tensorflow1.15 and TensorFlow2.4 were used for int8 precision quantization of the same Pb model, and the quantization results were different. Where, the BIAS of float32 bit is the same, but the weight of int8 bit convolution kernel is different Open

tensorflow/model-optimization#794 · 4 comments · 1 assignee · Opened Aug 18, 2021
loc("model/FPN/FPN/FPN//Concatenate_p_1/concat"): error: 'tfl.concatenation' op quantization parameters violate the same scale constraint: !quant.uniform<i8:f32, 9.9999999999999995E-7> vs. !quant.uniform<i8:f32, 4.9130543629871681E-5:-128> Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

tensorflow/model-optimization#793 · 6 comments · Opened Aug 17, 2021
Tensorflow lite INT8 Convert Issue Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

tensorflow/model-optimization#791 · 3 comments · Opened Aug 17, 2021
Tensorflow Lite interpreter `invoke()` hangs upon attempting int8 quantization with convertor Open

Difficulty 4/5 3-5 days Newbie friendliness 35/100

tensorflow/model-optimization#788 · 1 comment · Opened Aug 16, 2021
Quantization Aware Training support for KerasLayer Open

feature request

Difficulty 5/5 Over a week Newbie friendliness 25/100

tensorflow/model-optimization#786 · Opened Aug 13, 2021
Use TensorFlow Lite quantize RetinaFace get wrong result Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

tensorflow/model-optimization#784 · 1 comment · Opened Aug 12, 2021
Disable quantization range during training Open

tensorflow/model-optimization#777 · 1 comment · 1 assignee · Opened Aug 5, 2021
pruning example code for GNMT Open

feature request

tensorflow/model-optimization#776 · 1 comment · 1 assignee · Opened Aug 4, 2021
Tflite full Integer quantization bug. Open

tensorflow/model-optimization#775 · 8 comments · 1 assignee · Opened Aug 4, 2021
wrong inference results of qat int8 model with Conv2DTranspose. Open

bug

tensorflow/model-optimization#771 · 1 assignee · Opened Jul 28, 2021
Convert model to full int after quantization aware training Open

tensorflow/model-optimization#765 · 1 comment · 1 assignee · Opened Jul 19, 2021
RuntimeError: Layer backbone.stage0.0.fuse_layersb2.1.0.0.1:<class 'tensorflow.python.keras.layers.normalization_v2.BatchNormalization'> is not supported. You can quantize this layer by passing a `tfmot.quantization.keras.QuantizeConfig` instance to the `quantize_annotate_layer` API. Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 35/100

tensorflow/model-optimization#763 · 3 comments · Opened Jul 16, 2021
pruning cannnot reduce .h5 model size Open

bug technique:pruning

Difficulty 4/5 3-5 days Newbie friendliness 25/100

tensorflow/model-optimization#760 · 3 comments · Opened Jul 12, 2021
Support for custom layers composed of standard keras layers Open

Difficulty 5/5 Over a week Newbie friendliness 25/100

tensorflow/model-optimization#756 · 2 comments · Opened Jul 6, 2021
Issue loading model with Conv2DTranspose Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 35/100

tensorflow/model-optimization#751 · 3 comments · Opened Jun 29, 2021
Pruning Transfer Learning Models Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

tensorflow/model-optimization#749 · 1 comment · Opened Jun 27, 2021
TypeError: '<' not supported between instances of 'InputLayer' and 'Sequential' Open

Difficulty 4/5 3-5 days Newbie friendliness 25/100

tensorflow/model-optimization#748 · 3 comments · Opened Jun 25, 2021
Running MobileNet on XNNPACK Open

bug technique:pruning

tensorflow/model-optimization#735 · 1 reaction · 2 assignees · Opened Jun 14, 2021
Support for Multiply layer Open

feature request

tensorflow/model-optimization#733 · 1 comment · 5 reactions · 1 assignee · Opened Jun 10, 2021
Structural (filter) pruning for convolutional layers Open

feature request technique:pruning

tensorflow/model-optimization#732 · 5 comments · 3 reactions · 2 assignees · Opened Jun 8, 2021
Assertion error with custom quantize config with per_axis=True Open

bug feature request

tensorflow/model-optimization#723 · 1 comment · 1 assignee · Opened May 31, 2021
Puring cannot reduce tflite model size Open

bug

Difficulty 4/5 3-5 days Newbie friendliness 35/100

tensorflow/model-optimization#722 · 5 comments · Opened May 29, 2021
" Depth of input is not a multiple of input depth of filter" error when quantizing loaded model Open

tensorflow/model-optimization#713 · 2 assignees · Opened May 19, 2021
Converting a pruning model in a TFlite model Open

bug

tensorflow/model-optimization#668 · 1 comment · 1 assignee · Opened Apr 18, 2021
Error reported when prunning one tf-keras layer Open

bug

tensorflow/model-optimization#662 · 1 comment · 1 assignee · Opened Apr 14, 2021
Pruning models from tf.keras.applications (efficientnet, resnet etc) Open

tensorflow/model-optimization#656 · 12 comments · 1 assignee · Opened Apr 5, 2021
QAT training support strided_slice Open

feature request

Difficulty 4/5 3-5 days Newbie friendliness 15/100

tensorflow/model-optimization#650 · 2 comments · Opened Mar 22, 2021
Problem regarding loading a QAT SavedModel Open

bug

tensorflow/model-optimization#644 · 11 comments · 1 reaction · 1 assignee · Opened Mar 10, 2021

tensorflow/model-optimization

Get new issues in your inbox