Downloads · 30 days
23
5% of all-time downloads
brucethemoose/Capybara-Tess-Yi-34B-200K-DARE-Ties
Capybara-Tess-Yi-34B-200K-DARE-Ties is a text generation model from brucethemoose. Use it when you need the model to write or continue text. It is set up for transformers. The card lists the license as other.
NousResearch/Nous-Capybara-34B, migtissera/Tess-M-v1.2 and migtissera/Tess-M-v1.3 merged with a new, experimental implementation of "dare ties" via mergekit. See:
Downloads · 30 days
23
5% of all-time downloads
All-time downloads
451
Public
Parameters
34.4B
68.8 GB on disk
Likes
11
Public
Click a slice to open those files.
.safetensors68.8 GB · 100%
From the Hugging Face model README
NousResearch/Nous-Capybara-34B, migtissera/Tess-M-v1.2 and migtissera/Tess-M-v1.3 merged with a new, experimental implementation of "dare ties" via mergekit. See:
Language Models are Super Mario: Absorbing Abilities from Homologous Models as a Free Lunch
https://github.com/yule-BUAA/MergeLM
https://github.com/cg123/mergekit/tree/dare-tokenizer
Highly experimental and still being tested! But this should yield a better merge than a typical linear/slerp merge or even a ties merge.
Merged with the following config, and the tokenizer from Yi Llamafied:
models:
- model: /home/alpha/Storage/Models/Raw/larryvrh_Yi-34B-200K-Llamafied
# no parameters necessary for base model
- model: /home/alpha/Storage/Models/Raw/migtissera_Tess-M-v1.3
parameters:
weight: 0.50
density: 0.56
- model: /home/alpha/Storage/Models/Raw/migtissera_Tess-M-v1.2
parameters:
weight: 0.20
density: 0.50
- model: /home/alpha/Storage/Models/Raw/Nous-Capybara-34B
parameters:
weight: 0.50
density: 0.56
merge_method: dare_ties
base_model: /home/alpha/Storage/Models/Raw/larryvrh_Yi-34B-200K-Llamafied
parameters:
int8_mask: true
dtype: bfloat16
Tess 1.2 (at a low weight) and 1.3 were used because, according to the trainer, they were trained on different datasets: https://migel.substack.com/p/learnings-from-training-tess
As the Tess creator warned about, if the model repeats at high context like Tess 1.2, let me know!
I chose not to include other finetunes, such as Dolphin, because they aren't trained on the 200K base. If any other 200K finetunes pop up, let me know.
SYSTEM: {system_message}
USER: {prompt}
ASSISTANT:
Being a Yi model, try disabling the BOS token and/or running a lower temperature with MinP if output doesn't seem right.
Sometimes the model "spells out" the stop token as </s> like Capybara, so you may need to add </s> as an additional stopping condition.
Credits:
https://github.com/cg123/mergekit/tree/dare-tokenizer
https://huggingface.co/NousResearch/Nous-Capybara-34B/
https://huggingface.co/migtissera/Tess-M-v1.2
https://huggingface.co/migtissera/Tess-M-v1.3