Comments (3)
Fixed by #15?
If not, what is left to do?
from onnxscript.
I think the following needs to be done (mostly to ensure consistency between eager-mode and onnx-conversion):
- Will need type-checking to ensure types of arguments passed in (for eager-mode) are of correct expected types
- This will be better off if we define our own types/classes (e.g. our own tensor-types, that may be a wrapper around numpy arrays) and if these are consistent with the type-annotations we use (such as FLOAT[100]). At any rate, we should have a clear definition of what representation-type we use for what onnx-types.
Further, we should be able to do the following (not sure it is there yet): simply import the function-definitions, and call them as normal python functions. Just like in the example shown above for gemmgelu
, but I think some clean up may be required (such as special handling of the identifier oxs
etc.)
from onnxscript.
Looks like this is done.
from onnxscript.
Related Issues (20)
- crash in rewriter/broadcast_to_matmul.py (proposed fix) HOT 3
- tiiuae/falcon-rw-1b: Dynamo export error during optimization step HOT 9
- There should be a method to set size threshold for constant folding HOT 1
- [IR] Create a DEBUG flag to enable addtional checks
- Create method to turn external tensor to ir.Tensor
- Remove `get_numpy_from_ir_value`
- [IR] Pass manager follow up
- [IR] Figure out how we can share tensors
- Create a node iterator
- how to set opset in generated model by to_model_proto? HOT 4
- proper way to construct a tensor on the fly? HOT 4
- SequenceEmpty parameter dtype seems not work HOT 2
- Handle attribute constraints in multi-output pattern-matcher
- Keyword arg not respected and was treated as positional arg HOT 2
- op.Pad behavior is inconsistent with onnx ReferenceEvaluator and documentation HOT 2
- converted model with loop cannot be load by onnxruntime
- [IR] Handle nodes with defined output names
- [IR] Fix Value repr
- How to set value in ConstantOfShape? HOT 2
- op.Resize parameter sizes cannot accept Type tensor(int64) HOT 2
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 onnxscript.