Comments (6)
I think this is again an issue of the compat specification of Transformers.jl. On 1.10, only GPUCompiler 0.22.0 upwards are available (to my understanding because of this line). However, these never versions of GPUCompiler are only compatible with newer versions of CUDA.jl, but Transformers.jl restricts the CUDA version (explicitly, and implicitly via the Flux compat entry).
from transformers.jl.
Any plans for this?
from transformers.jl.
EDIT: I am wrong -- it's an issue on 1.10, not Transfomers. I don't understand it, but it has to do with CUDA / GPUCompiler.
I believe this makes Transformers.jl uninstallable now.
I'm not sure how it's possible, but I've tried from fresh environment and it fails in resolving. Clearly, some dependency has moved on :/
Example
]activate --temp
]add [email protected]
Resolving package versions...
ERROR: Unsatisfiable requirements detected for package GPUCompiler [61eb1bfa]:
GPUCompiler [61eb1bfa] log:
├─possible versions are: 0.1.0-0.26.4 or uninstalled
├─restricted by julia compatibility requirements to versions: 0.22.0-0.26.4 or uninstalled
└─restricted by compatibility requirements with CUDA [052768ef] to versions: 0.15.1-0.21.4 — no versions left
└─CUDA [052768ef] log:
├─possible versions are: 0.1.0-5.3.1 or uninstalled
├─restricted by julia compatibility requirements to versions: [2.3.0, 2.5.0-5.3.1] or uninstalled
├─restricted by compatibility requirements with Transformers [21ca0261] to versions: 3.10.0-5.3.1
│ └─Transformers [21ca0261] log:
│ ├─possible versions are: 0.0.1-0.2.8 or uninstalled
│ └─restricted to versions 0.2.8 by an explicit requirement, leaving only versions: 0.2.8
└─restricted by compatibility requirements with Flux [587475ba] to versions: 3.0.0-4.4.2, leaving only versions: 3.10.0-4.4.2
└─Flux [587475ba] log:
├─possible versions are: 0.4.1-0.14.15 or uninstalled
└─restricted by compatibility requirements with Transformers [21ca0261] to versions: 0.13.4-0.13.17
└─Transformers [21ca0261] log: see above
from transformers.jl.
I spent a bit trying to chase down the compat issues today and wow what an absolute nightmare.
- GPUCompiler.jl is less of an issue thanks to this.
- A bigger issue is NNlibCUDA.jl, which seems to be vaguely deprecated. Supposedly NNlib.jl contains all the functionality of NNlibCUDA.jl, but this may require some internal code changes and not just compat updates.
from transformers.jl.
Oh this is great, it doesn't seem like the package imports NNlibCUDA.jl anywhere, perhaps we can simply drop the compat. Experimenting.
from transformers.jl.
Opened #185, which should address some of this. Testing still ongoing.
from transformers.jl.
Related Issues (20)
- update NNlib and Flux compat HOT 9
- State of quantization HOT 3
- Dolly example no longer works ... HOT 19
- OWL-ViT HOT 1
- AMDGPU support HOT 1
- DistilBertModel support HOT 1
- Attempting to download CLIP yields UnderVarError `unk_token` not defined
- Performance issue HOT 1
- [Question] Possible to retrieve layer-wise activations? HOT 4
- Adding phi model HOT 5
- Please support Lux.jl HOT 7
- Example Code always produces Max Length Sequences
- how to download model weights on external drive
- Update to newer versions of dependencies
- Improve documentation and take inspiration from python package HOT 6
- Looking to update Transformers.jl and the associated modules HOT 1
- Storage of Downloaded Models from HuggingFace HOT 1
- Converting from integer-tokens to one-hot tokens gives different results. HOT 2
- Dependency resolution fails on Julia 1.10 HOT 4
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.jl.