Comments (7)
I am having the same issue on windows 10, core i5-8600K, 32GB of RAM, GTX 3080
from dalai.
I'm having this issue but 7B won't quantize
from dalai.
In my case : Ryzen 9 - Ubu 22.10
alpaca 7B [running good] and llama 7B [not loading ...yet]
[...]
/root/dalai/venv/bin/python convert-pth-to-ggml.py models/7B/ 1
{'dim': 4096, 'multiple_of': 256, 'n_heads': 32, 'n_layers': 32, 'norm_eps': 1e-06, 'vocab_size': -1}
Namespace(dir_model='models/7B/', ftype=1, vocab_only=0)
n_parts = 1
Processing part 0
[...]
- it seems that using just models 7B from Alpaca/llama - no ggml-model-f16.bin file was produced.
the only file found with a similar name is : ggml-vocab.bin [at same level as folder 7B in /llama/models. [432,6 Ko]
I've copied it to the quantize folder and tried to run :
sudo ./quantize ggml-vocab.bin ggml-model-q4_0.bin 2
but process has failed. (Same result while renaming it "***-f16.bin )
llama_model_quantize: loading model from 'ggml-vocab.bin'
llama_model_quantize: invalid model file 'ggml-vocab.bin' (bad magic)
main: failed to quantize model from 'ggml-vocab.bin'
any other work arround welcomed, thanks ;)
from dalai.
same issue on win 10 (docker)
from dalai.
find the quantize executable for simplicity copy it to your folder wher ggml-model-f16.bin is and run ./quantize ggml-model-f16.bin ggml-model-q4_0.bin 2
or on windows quantize.exe ggml-model-f16.bin ggml-model-q4_0.bin 2 it takes a few minutes
Listen to this guy!
from dalai.
similar issue
from dalai.
find the quantize executable
for simplicity copy it to your folder wher ggml-model-f16.bin is
and run ./quantize ggml-model-f16.bin ggml-model-q4_0.bin 2
or on windows
quantize.exe ggml-model-f16.bin ggml-model-q4_0.bin 2
it takes a few minutes
from dalai.
Related Issues (20)
- "npx dalai llama install 7B" fails with "./quantize : The term './quantize' is not recognized" HOT 1
- Dalai server doesn't start without internet connection HOT 1
- Possible solution for Windows users - LLama not working. HOT 4
- Low CPU, Low Memory, Low GPU usage via Docker HOT 3
- Typo in README.md HOT 1
- llama_model_load: loading model issue in Docker HOT 5
- Docker compose never responds HOT 5
- [EndeavorOS] model hangs
- can't install HOT 7
- any way this can run Falcon?
- silent fail in llama/main HOT 1
- Support for Llama-2? HOT 4
- Error installing llama using docker compose (logs attached) HOT 1
- Alpaca doesn't respond HOT 1
- boucle inf on javascript call to Dalai
- Does not work on Windows 11 or Linux Mint Cinnamon HOT 1
- llama 7B is talking jibberish. Does not respond intelligently to my command.
- On Mac OS, I got this error
- On Mac, Alpaca is stuck. Did I uninstall correctly?
- Models repository HOT 1
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 dalai.