Link adaround doc and api pages (quic#652)

* Link Adaround pages to parent page * Address review comments, Fix google collab install instruction links * Fixed apostrophe in model_quantization.rst that resulted in UnicodeDecodeError Co-authored-by: Bharath Ramaswamy <quic_bharathr@quicinc.com>
ethanhe42 · Jun 17, 2021 · 45b0820 · 45b0820
1 parent e5284d6
commit 45b0820
Show file tree

Hide file tree

Showing 8 changed files with 20 additions and 13 deletions.
diff --git a/Docs/api_docs/tensorflow_adaround.rst b/Docs/api_docs/tensorflow_adaround.rst
@@ -1,6 +1,6 @@
 :orphan:
 
-.. _api-tensorflow-adaround:
+.. _api-tf-adaround:
 
 ==================================
 AIMET TensorFlow AdaRound API

diff --git a/Docs/api_docs/tensorflow_bias_correction.rst b/Docs/api_docs/tensorflow_bias_correction.rst
@@ -1,6 +1,6 @@
 :orphan:
 
-.. _tf-bias-correction:
+.. _api-tf-bias-correction:
 
 ====================================
 AIMET TensorFlow Bias Correction API

diff --git a/Docs/api_docs/tensorflow_cross_layer_equalization.rst b/Docs/api_docs/tensorflow_cross_layer_equalization.rst
@@ -1,6 +1,6 @@
 :orphan:
 
-.. _tf-cle:
+.. _api-tf-cle:
 
 ===============================================
 AIMET TensorFlow Cross Layer Equalization APIs

diff --git a/Docs/api_docs/tensorflow_quantization.rst b/Docs/api_docs/tensorflow_quantization.rst
@@ -3,6 +3,7 @@ AIMET TensorFlow Quantization APIs
 ==================================
 
 AIMET Quantization for TensorFlow provides the following functionality
-   - :ref:`Quantization Simulation<tf-quantsim>`: Allows ability to simulate inference and training on quantized hardware
-   - :ref:`Cross-Layer Equalization<tf-cle>`: Post-training quantization technique to equalize layer parameters
-   - :ref:`Bias Correction<tf-bias-correction>`: Post-training quantization technique to correct shift in layer outputs due to quantization noise
+   - :ref:`Quantization Simulation<api-tf-quantsim>`: Allows ability to simulate inference and training on quantized hardware
+   - :ref:`Adaptive Rounding<api-tf-adaround>`: Post-training quantization technique to optimize rounding of weight tensors
+   - :ref:`Cross-Layer Equalization<api-tf-cle>`: Post-training quantization technique to equalize layer parameters
+   - :ref:`Bias Correction<api-tf-bias-correction>`: Post-training quantization technique to correct shift in layer outputs due to quantization noise
diff --git a/Docs/api_docs/tensorflow_quantsim.rst b/Docs/api_docs/tensorflow_quantsim.rst
@@ -1,10 +1,10 @@
 :orphan:
 
-.. _tf-quantsim:
+.. _api-tf-quantsim:
 
-===========================
-QuantSim API for TensorFlow
-===========================
+=====================================
+AIMET TensorFlow Quantization SIM API
+=====================================
 
 Top-level API
 =============

diff --git a/Docs/api_docs/torch_quantization.rst b/Docs/api_docs/torch_quantization.rst
@@ -4,6 +4,7 @@ AIMET PyTorch Quantization APIs
 
 AIMET Quantization for PyTorch provides the following functionality
    - :ref:`Quantization Simulation<api-torch-quantsim>`: Allows ability to simulate inference and training on quantized hardware
+   - :ref:`Adaptive Rounding<api-torch-adaround>`: Post-training quantization technique to optimize rounding of weight tensors
    - :ref:`Cross-Layer Equalization<api-torch-cle>`: Post-training quantization technique to equalize layer parameters
    - :ref:`Bias Correction<api-torch-bias-correction>`: Post-training quantization technique to correct shift in layer outputs due to quantization noise
 
diff --git a/Docs/user_guide/model_quantization.rst b/Docs/user_guide/model_quantization.rst
@@ -15,7 +15,7 @@ AIMET provides multiple techniques and tools that help to create quantized model
 
 Use Cases
 =========
-1. **Predict on-target accuracy**: AIMET enables a user to simulate the effects of quantization to get a first order estimate of the model’s accuracy when run on quantized targets. This is useful to get an estimate of on-target accuracy without needing an actual target platform. Note that to create a simulation model, AIMET uses representative data samples to compute per-layer quantization encodings.
+1. **Predict on-target accuracy**: AIMET enables a user to simulate the effects of quantization to get a first order estimate of the model's accuracy when run on quantized targets. This is useful to get an estimate of on-target accuracy without needing an actual target platform. Note that to create a simulation model, AIMET uses representative data samples to compute per-layer quantization encodings.
 
     .. image:: ../images/quant_use_case_1.PNG
 
@@ -42,10 +42,15 @@ Relevant steps in the diagram are annotated with a green circle with a number. F
 
 .. image:: ../images/flow_diagram.png
 
+AIMET Quantization features
+===========================
+
 #. :ref:`Quantization Simulation<ug-quantsim>`: AIMET enables a user to modify a model to add quantization simulation ops. When an evaluation is run on a model with these quantization simulation ops, the user can observe a first-order simulation of expected accuracy on quantized hardware.
 
 #. :ref:`Quantization Visualization<ug-quantization-visualization>`: AIMET provides visualization tools that help guide the user to determine if AIMET post-training quantization techniques are useful for a given model
 
+#. :ref:`Applying Adaptive Rounding<ug-adaround>`: Determine optimal rounding for weight tensors to improve quantized performance
+
 #. :ref:`Applying Cross-Layer Equalization<ug-post-training-quantization>`: Post-training quantization techniques help a model improve quantized accuracy without needing to re-train. Cross-Layer Equalization equalizes weight ranges in consecutive layers.
 
 #. :ref:`Applying Bias Correction<ug-post-training-quantization>`: Bias Correction corrects for a shift in layer outputs due to quantization noise

diff --git a/README.md b/README.md
@@ -129,11 +129,11 @@ AIMET can also significantly compress models. For popular models, such as Resnet
 ## Installation Instructions
 To install and use the pre-built version of the AIMET package, please follow one of the below links:
 - [Install and run AIMET in *Ubuntu* environment](./packaging/install.md)
-- [Install and run AIMET in *Google Colab* environment](./packaging/google_colab_install.md)
+- [Install and run AIMET in *Google Colab* environment](./packaging/google_colab/google_colab_install.md)
 
 To build, modify (*optionally*) and use the latest AIMET source code, please follow one of the below links:
 - [Instructions to build, install and run AIMET in *Docker* environment](./packaging/docker_install.md)
-- [Instructions to build, install and run AIMET in *Google Colab* environment](./packaging/google_colab_development.md)
+- [Instructions to build, install and run AIMET in *Google Colab* environment](./packaging/google_colab/google_colab_development.md)
 
 ## Resources
 * [User Guide](https://quic.github.io/aimet-pages/releases/1.16.1/user_guide/index.html)