models(gallery): add rombos-coder-v2.5-qwen-32b (#4134)

Signed-off-by: Ettore Di Giacinto <mudler@localai.io>
This commit is contained in:
Ettore Di Giacinto 2024-11-13 09:37:08 +01:00 committed by GitHub
parent 1db504353c
commit aebaf71be6
No known key found for this signature in database
GPG Key ID: B5690EEEBB952194

View File

@ -136,6 +136,23 @@
- filename: Rombos-Coder-V2.5-Qwen-7b-Q4_K_M.gguf
sha256: ca16a550f1be00b7e92f94c0c18ea6af1e5c158d5d1cb3994f9f0a0d13922272
uri: huggingface://bartowski/Rombos-Coder-V2.5-Qwen-7b-GGUF/Rombos-Coder-V2.5-Qwen-7b-Q4_K_M.gguf
- !!merge <<: *qwen25coder
name: "rombos-coder-v2.5-qwen-32b"
icon: https://cdn-uploads.huggingface.co/production/uploads/642cc1c253e76b4c2286c58e/QErypCEKD5OZLxUcSmYaR.jpeg
urls:
- https://huggingface.co/rombodawg/Rombos-Coder-V2.5-Qwen-32b
- https://huggingface.co/bartowski/Rombos-Coder-V2.5-Qwen-32b-GGUF
- https://docs.google.com/document/d/1OjbjU5AOz4Ftn9xHQrX3oFQGhQ6RDUuXQipnQ9gn6tU/edit?usp=sharing
description: |
Rombos-Coder-V2.5-Qwen-32b is a continues finetuned version of Qwen2.5-Coder-32B-Instruct. I took it upon myself to merge the instruct model with the base model myself using the Ties merge method as demonstrated in my own "Continuous Finetuning" method (link available).
This version of the model shows higher performance than the original instruct and base models.
overrides:
parameters:
model: Rombos-Coder-V2.5-Qwen-32b-Q4_K_M.gguf
files:
- filename: Rombos-Coder-V2.5-Qwen-32b-Q4_K_M.gguf
sha256: 821ea2a13d96354db1368986700b1189938fbbc56ca6bb9d0c39f752580de71a
uri: huggingface://bartowski/Rombos-Coder-V2.5-Qwen-32b-GGUF/Rombos-Coder-V2.5-Qwen-32b-Q4_K_M.gguf
- &opencoder
name: "opencoder-8b-base"
icon: https://github.com/OpenCoder-llm/opencoder-llm.github.io/blob/main/static/images/opencoder_icon.jpg?raw=true