Comments (7)
I tested 13B and 30B without problems. Just download the models and start the program with the correct file, e.g.:
-m ggml-alpaca-30b-q4.bin
from alpaca.cpp.
Download of the 30B Model:
https://huggingface.co/Pi3141/alpaca-lora-30B-ggml/tree/main
4_1 does not work for me, i use the ggml-model-q4_0.bin file and renamed it to ggml-alpaca-30b-q4.bin.
It's the same file, that was downloaded inside the Dalai docker container:
sudo docker compose run dalai md5sum /root/dalai/alpaca/models/30B/ggml-model-q4_0.bin && md5sum /daten/alpaca/ggml-alpaca-30b-q4.bin
7bc55515cb128cd2ded9403d35c25244 /root/dalai/alpaca/models/30B/ggml-model-q4_0.bin
7bc55515cb128cd2ded9403d35c25244 /daten/alpaca/ggml-alpaca-30b-q4.bin
My system is a notebook and i reduced the cpu frequency a little, one token tooks over 600 ms.
from alpaca.cpp.
There was in the README the option to choose between 3 models and it disappeared a couple of weeks ago.
I found the 13b model, I wasn't able to find the 30b. I am still looking.
from alpaca.cpp.
@supportend could you please link the bin files for 13B and 30B?
how many tokens/s are you getting with the larger models?
from alpaca.cpp.
could you please link the bin files for 13B
Download of the 30B Model: https://huggingface.co/Pi3141/alpaca-lora-30B-ggml/tree/main 4_1 does not work for me, i use the ggml-model-q4_0.bin file and renamed it to ggml-alpaca-30b-q4.bin.
It's the same file, that was downloaded inside the Dalai docker container:
sudo docker compose run dalai md5sum /root/dalai/alpaca/models/30B/ggml-model-q4_0.bin && md5sum /daten/alpaca/ggml-alpaca-30b-q4.bin 7bc55515cb128cd2ded9403d35c25244 /root/dalai/alpaca/models/30B/ggml-model-q4_0.bin 7bc55515cb128cd2ded9403d35c25244 /daten/alpaca/ggml-alpaca-30b-q4.bin
My system is a notebook and i reduced the cpu frequency a little, one token tooks over 600 ms.
could you please link the bin files for 13B???
from alpaca.cpp.
I think, this is it:
https://huggingface.co/Pi3141/alpaca-lora-13B-ggml/tree/main
again 4_0 file.
There is a native version too on the site. But i dit not download a 4_1 file or the.
from alpaca.cpp.
I think, this is it:
https://huggingface.co/Pi3141/alpaca-lora-13B-ggml/tree/main again 4_0 file.
There is a native version too on the site. But i dit not download a 4_1 file or the.
thanks a lot i download it
from alpaca.cpp.
Related Issues (20)
- windows 10, fails to get to chat prompt HOT 8
- Issues with Monte Carlo Tree Search on alpaca.cpp
- CUDA support? HOT 3
- Input Handling HOT 1
- Segmentation Fault with 7b on Raspberry Pi 3 HOT 5
- ### Instruction: ### Response: HOT 3
- Random seed not working correctly, "too many inputs", chinese symbols?
- How to use with python HOT 1
- Can i use it on linux with nvidia gpu HOT 4
- minimum specs
- Does not perceive other languages HOT 5
- Running very slow HOT 7
- Redpajama dataset?
- using ggml-alpaca-13b-q4.bin HOT 1
- Archive Repository HOT 1
- main: failed to load model from 'ggml-alpaca-7b-q4.bin' HOT 4
- llama_model_load: failed to open 'ggml-alpaca-7b-native-q4.bin' HOT 1
- HuggingFaceH4/starchat-alpha Cpp LLM
- Adding risc-v architecture support 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 alpaca.cpp.