OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Models

Conformer — Rishit-dagli

Conformer is a Python implementation of the convolution-augmented Transformer architecture for speech recognition, using TensorFlow and Keras. It provides components such as a convolutional module and a Conformer block that developers can instantiate with configurable model dimensions and other parameters.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Conformer — Rishit-dagli — OpenViST project profile in AI Models.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use it when you need TensorFlow/Keras building blocks for experimenting with Conformer architecture in speech-recognition modeling, rather than a complete transcription product.

Who it is for

  • TensorFlow/Keras developers prototyping speech-recognition models.
  • Researchers comparing convolution-augmented transformers with other speech-model approaches.
  • Python developers who want to inspect or adapt an Apache-2.0-licensed implementation of Conformer components.

Documented uses

  • Create a Conformer convolutional module and apply it to a TensorFlow tensor.
  • Instantiate a Conformer block with configurable attention and convolution parameters.
  • Set up the repository's development dependencies and run its documented rank and shape tests.
Before you choose

Requirements & limits

The public GitHub repository lists an Apache-2.0 license. The README documents installation through pip and development installation from a clone; it does not state account or pricing requirements.

  • The implementation is built around TensorFlow/Keras rather than presented as a framework-neutral model.
  • Documented usage covers model components and tensor inputs, not an end-to-end speech-recognition or transcription pipeline.
  • The README gives a package-install command and a separate source-based development setup.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Create and apply a Conformer block to a sample tensor

Instantiate the documented Conformer block and apply it to a random TensorFlow tensor.

Before you start

  • Install conformer-tf.
  • Use TensorFlow.

Steps

  1. Import tensorflow as tf and ConformerBlock from conformer_tf.

  2. Instantiate ConformerBlock with dim=512, dim_head=64, heads=8, ff_mult=4, conv_expansion_factor=2, conv_kernel_size=31, and all three dropout parameters set to 0..

  3. Create a random tensor with tf.random.normal([1, 1024, 512]).

  4. Call conformer_block(x). The example indicates an output shape of (1, 1024, 512).

Expected outcome

A Conformer block is applied to the sample tensor, with the example showing output shape `(1, 1024, 512)`.

Guide limitations

  • The example uses random data and demonstrates block application, not speech recognition, model training, or evaluation.
  • The README does not specify a required TensorFlow version.
Documentation 1 ↗
Evidence and transparency

Sources & review

Official repository ↗
Source record details
Source availability
Active
Sources checked
18 October 2026
Original website title
Rishit-dagli/Conformer · Official homepage title · retained in its source language
How we review sourcesSuggest a correction