Comments (5)
Can you try changing the GGML_ASSERT
in ggml_cuda:2392
to CUDA_CHECK(err);
? That should give more information about the error.
from llama.cpp.
having the exact same issue only with hip instead of cuda
changed that in the file and got this
ggml_cuda_compute_forward: RMS_NORM failed
CUDA error: invalid device function
current device: 0, in function ggml_cuda_compute_forward at C:/Users/gaming/src/llama.cpp/ggml-cuda.cu:2392
err
from llama.cpp.
"invalid device function" suggests that the code was not compiled for the device being used, or with a version of the CUDA toolkit not compatible with the installed driver version.
from llama.cpp.
Can you try changing the
GGML_ASSERT
inggml_cuda:2392
toCUDA_CHECK(err);
? That should give more information about the error.
Thanks. Found the problem:
ml_cuda_compute_forward: RMS_NORM failed
CUDA error: the provided PTX was compiled with an unsupported toolchain.
current device: 0, in function ggml_cuda_compute_forward at F:\LLM\llama.cpp\ggml-cuda.cu:2392
err
GGML_ASSERT: F:\LLM\llama.cpp\ggml-cuda.cu:60: !"CUDA error"
from llama.cpp.
I downgrade the cuda version. It works now. But while answering, cuda seems does not works?
from llama.cpp.
Related Issues (20)
- Proposal: An alternative to chat templates HOT 5
- gcc: fatal error: cannot execute ‘cc1plus’: execvp: No such file or directory HOT 1
- mixtral 8x22b instruct Q2 core dumped HOT 1
- Offload kv to smaller GPU HOT 2
- safetensors_rust.SafetensorError: Error while deserializing header: MetadataIncompleteBuffer
- Output of main and server is totally different HOT 3
- GPU is not working well. HOT 2
- make -j LLAMA_CUDA=1 CUDA_DOCKER_ARCH=sm_53 HOT 1
- Low prompt processing speed with mixtral? HOT 7
- Llama.cpp GPU Offloading Issue - Unexpected Switch to CPU HOT 7
- Llama only uses dedicated memory when both shared and dedicated are available. HOT 2
- Update convert.py to accept these argument choices for quantization HOT 1
- llama3 family support HOT 78
- AMD GPU is slower than expected HOT 6
- write mean pooled embedding to callers vector to simplify using SoTA embedding models and language bindings HOT 1
- quantize serious memory leak! HOT 2
- ggml : add GPU support for Mamba models
- Does GGUF consist of a feedforward architecture model with weights? HOT 2
- llama3 model inference will keep repeating the output
- New optimization from NVIDIA to use CUDA Graphs in llama.cpp HOT 11
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 llama.cpp.