Comments (2)
Well, I would say that is up to the user of the library... 😅 It should be their responsibility to choose models that run on their desired platforms.
There are some ways to produce smaller models (e.g., quantization and pruning), but they can only get you so far...
That said, it probably would be good to build up a larger collection of small models (like facebook/detr-resnet-50, which is only 43MB, but can be used for object detection).
If you have some candidates, you can post links to them here
from transformers.js.
Closing issue as there are models which are able to work on mobile devices. For an example, I have tested Whisper Web on my phone and it can run the tiny, quantized versions of whisper (~61MB in size).
Overall, if a user of the library wishes to deploy a version of their app to run in mobile browsers, it is their responsibility to choose models which can run in a mobile browser.
Alternatively, some users have found ways to get Transformers.js running natively via react-native. For more information, see this PR.
from transformers.js.
Related Issues (20)
- The inference results on the same model are inconsistent when using WebGPU and WASM
- TTS for multiple languages HOT 1
- Decode and WebGPU error when using online application Xenova/experimental-phi3-webgpu HOT 8
- Store downloaded model files in IndexedDB HOT 1
- Uncaught Error: Can't create a session for remove-background-client demo HOT 2
- Electron exe build of app crashes to desktop when trying to import Transformers.js
- Upgrade to Jinja v0.3.0 HOT 3
- Unsupported model type: jais HOT 10
- (V3) `examples/demo-stite` had problems generating results HOT 1
- Stable Diffusion HOT 3
- (V3) WebGPU - await extractor() broken - Error: Session already started HOT 3
- [v3] `NoBadWordsLogitsProcessor` breaks when the size of `bad_word_ids` exceeds the size of the input ids HOT 1
- Steps to test out V3 HOT 4
- openvino backend HOT 3
- (v3) speed up NoBadWordsLogitsProcessor HOT 2
- Support wavlm-base-plus-sv with WebGPU HOT 3
- [v3] feature-extraction run with webgpu throw error
- The "path" argument must be of type string or an instance of URL. Received undefined HOT 2
- Where should I get `decoder_model_merged` file from? HOT 3
- When running the feature extraction pipeline with an array of strings as input, the embedding result is different from running the pipeline for one string at a time. HOT 3
Recommend Projects
-
React
A declarative, efficient, and flexible JavaScript library for building user interfaces.
-
Vue.js
🖖 Vue.js is a progressive, incrementally-adoptable JavaScript framework for building UI on the web.
-
Typescript
TypeScript is a superset of JavaScript that compiles to clean JavaScript output.
-
TensorFlow
An Open Source Machine Learning Framework for Everyone
-
Django
The Web framework for perfectionists with deadlines.
-
Laravel
A PHP framework for web artisans
-
D3
Bring data to life with SVG, Canvas and HTML. 📊📈🎉
-
Recommend Topics
-
javascript
JavaScript (JS) is a lightweight interpreted programming language with first-class functions.
-
web
Some thing interesting about web. New door for the world.
-
server
A server is a program made to process requests and deliver data to clients.
-
Machine learning
Machine learning is a way of modeling and interpreting data that allows a piece of software to respond intelligently.
-
Visualization
Some thing interesting about visualization, use data art
-
Game
Some thing interesting about game, make everyone happy.
Recommend Org
-
Facebook
We are working to build community through open source technology. NB: members must have two-factor auth.
-
Microsoft
Open source projects and samples from Microsoft.
-
Google
Google ❤️ Open Source for everyone.
-
Alibaba
Alibaba Open Source for everyone
-
D3
Data-Driven Documents codes.
-
Tencent
China tencent open source team.
from transformers.js.