ExternalVendorCode/LocalAI
1
0
Fork 0
mirror of https://github.com/mudler/LocalAI.git synced 2025-01-18 10:46:46 +00:00
Code Issues Actions 17 Packages Projects Releases Wiki Activity

models(gallery): add arliai-llama-3-8b-dolfin-v0.5 (#2852)

Browse Source 
Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
This commit is contained in:
Ettore Di Giacinto 2024-07-13 22:40:00 +02:00 committed by GitHub
parent 405794d4ca
commit 9eb27c563c
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194
1 changed files with 23 additions and 0 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

23
gallery/index.yaml
View File

@ -2750,6 +2750,29 @@
- filename: LLaMAX3-8B.Q4_K_M.gguf
sha256: 862fb2be5d74b171f4294f862f43e7cb6e6dbecce29a9f9167da4f1db230daac
uri: huggingface://mradermacher/LLaMAX3-8B-GGUF/LLaMAX3-8B.Q4_K_M.gguf
- !!merge <<: *llama3
name: "arliai-llama-3-8b-dolfin-v0.5"
urls:
- https://huggingface.co/OwenArli/ArliAI-Llama-3-8B-Dolfin-v0.5
- https://huggingface.co/QuantFactory/ArliAI-Llama-3-8B-Dolfin-v0.5-GGUF
description: |
Based on Meta-Llama-3-8b-Instruct, and is governed by Meta Llama 3 License agreement: https://huggingface.co/meta-llama/Meta-Llama-3-8B-Instruct
This is a fine tune using an improved Dolphin and WizardLM dataset intended to make the model follow instructions better and refuse less.
OpenLLM Benchmark:
Training:
2048 sequence length since the dataset has an average length of under 1000 tokens, while the base model is 8192 sequence length. From testing it still performs the same 8192 context just fine.
Training duration is around 2 days on 2xRTX 3090, using 4-bit loading and Qlora 64-rank 128-alpha resulting in ~2% trainable weights.
overrides:
parameters:
model: ArliAI-Llama-3-8B-Dolfin-v0.5.Q4_K_M.gguf
files:
- filename: ArliAI-Llama-3-8B-Dolfin-v0.5.Q4_K_M.gguf
sha256: 71fef02915c606b438ccff2cae6b7760bbb54a558d5f2d39c2421d97b6682fea
uri: huggingface://QuantFactory/ArliAI-Llama-3-8B-Dolfin-v0.5-GGUF/ArliAI-Llama-3-8B-Dolfin-v0.5.Q4_K_M.gguf
- &chatml
### ChatML
url: "github:mudler/LocalAI/gallery/chatml.yaml@master"