base_model: flammenai/Mahou-1.5-mistral-nemo-12B
datasets:
- flammenai/MahouMix-v1
library_name: transformers
license: apache-2.0
tags:
- llama-cpp
- gguf-my-repo
model-index:
- name: Mahou-1.5-mistral-nemo-12B
results:
- task:
type: text-generation
name: Text Generation
dataset:
name: IFEval (0-Shot)
type: HuggingFaceH4/ifeval
args:
num_few_shot: 0
metrics:
- type: inst_level_strict_acc and prompt_level_strict_acc
value: 67.51
name: strict accuracy
source:
url: >-
https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=flammenai/Mahou-1.5-mistral-nemo-12B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: BBH (3-Shot)
type: BBH
args:
num_few_shot: 3
metrics:
- type: acc_norm
value: 36.26
name: normalized accuracy
source:
url: >-
https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=flammenai/Mahou-1.5-mistral-nemo-12B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: MATH Lvl 5 (4-Shot)
type: hendrycks/competition_math
args:
num_few_shot: 4
metrics:
- type: exact_match
value: 5.06
name: exact match
source:
url: >-
https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=flammenai/Mahou-1.5-mistral-nemo-12B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: GPQA (0-shot)
type: Idavidrein/gpqa
args:
num_few_shot: 0
metrics:
- type: acc_norm
value: 3.47
name: acc_norm
source:
url: >-
https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=flammenai/Mahou-1.5-mistral-nemo-12B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: MuSR (0-shot)
type: TAUR-Lab/MuSR
args:
num_few_shot: 0
metrics:
- type: acc_norm
value: 16.47
name: acc_norm
source:
url: >-
https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=flammenai/Mahou-1.5-mistral-nemo-12B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: MMLU-PRO (5-shot)
type: TIGER-Lab/MMLU-Pro
config: main
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 28.91
name: accuracy
source:
url: >-
https://huggingface.co/spaces/open-llm-leaderboard/open_llm_leaderboard?query=flammenai/Mahou-1.5-mistral-nemo-12B
name: Open LLM Leaderboard
Triangle104/Mahou-1.5-mistral-nemo-12B-Q5_K_M-GGUF
This model was converted to GGUF format from flammenai/Mahou-1.5-mistral-nemo-12B using llama.cpp via the ggml.ai's GGUF-my-repo space.
Refer to the original model card for more details on the model.
Model details:
Mahou-1.5-mistral-nemo-12B
Mahou is designed to provide short messages in a conversational context. It is capable of casual conversation and character roleplay. Chat Format
This model has been trained to use ChatML format.
<|im_start|>system {{system}}<|im_end|> <|im_start|>{{char}} {{message}}<|im_end|> <|im_start|>{{user}} {{message}}<|im_end|>
Roleplay Format
Speech without quotes.
Actions in *asterisks*
leans against wall cooly so like, i just casted a super strong spell at magician academy today, not gonna lie, felt badass.
SillyTavern Settings
Use ChatML for the Context Template.
Enable Instruct Mode.
Use the Mahou ChatML Instruct preset.
Use the Mahou Sampler preset.
Method
ORPO finetuned with 4x H100 for 3 epochs. Open LLM Leaderboard Evaluation Results
Detailed results can be found here Metric Value Avg. 26.28 IFEval (0-Shot) 67.51 BBH (3-Shot) 36.26 MATH Lvl 5 (4-Shot) 5.06 GPQA (0-shot) 3.47 MuSR (0-shot) 16.47 MMLU-PRO (5-shot) 28.91
Use with llama.cpp
Install llama.cpp through brew (works on Mac and Linux)
brew install llama.cpp
Invoke the llama.cpp server or the CLI.
CLI:
llama-cli --hf-repo Triangle104/Mahou-1.5-mistral-nemo-12B-Q5_K_M-GGUF --hf-file mahou-1.5-mistral-nemo-12b-q5_k_m.gguf -p "The meaning to life and the universe is"
Server:
llama-server --hf-repo Triangle104/Mahou-1.5-mistral-nemo-12B-Q5_K_M-GGUF --hf-file mahou-1.5-mistral-nemo-12b-q5_k_m.gguf -c 2048
Note: You can also use this checkpoint directly through the usage steps listed in the Llama.cpp repo as well.
Step 1: Clone llama.cpp from GitHub.
git clone https://github.com/ggerganov/llama.cpp
Step 2: Move into the llama.cpp folder and build it with LLAMA_CURL=1 flag along with other hardware-specific flags (for ex: LLAMA_CUDA=1 for Nvidia GPUs on Linux).
cd llama.cpp && LLAMA_CURL=1 make
Step 3: Run inference through the main binary.
./llama-cli --hf-repo Triangle104/Mahou-1.5-mistral-nemo-12B-Q5_K_M-GGUF --hf-file mahou-1.5-mistral-nemo-12b-q5_k_m.gguf -p "The meaning to life and the universe is"
or
./llama-server --hf-repo Triangle104/Mahou-1.5-mistral-nemo-12B-Q5_K_M-GGUF --hf-file mahou-1.5-mistral-nemo-12b-q5_k_m.gguf -c 2048