Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference time
Paper • 2203.05482 • Published • 9
How to use from
llama.cppInstall from WinGet (Windows)
winget install llama.cpp
# Start a local OpenAI-compatible server with a web UI:
llama serve -hf QuantFactory/Aurora_faustus-8B-LINEAR-GGUF:# Run inference directly in the terminal:
llama cli -hf QuantFactory/Aurora_faustus-8B-LINEAR-GGUF:Use pre-built binary
# Download pre-built binary from:
# https://github.com/ggerganov/llama.cpp/releases# Start a local OpenAI-compatible server with a web UI:
./llama-server -hf QuantFactory/Aurora_faustus-8B-LINEAR-GGUF:# Run inference directly in the terminal:
./llama-cli -hf QuantFactory/Aurora_faustus-8B-LINEAR-GGUF:Build from source code
git clone https://github.com/ggerganov/llama.cpp.git
cd llama.cpp
cmake -B build
cmake --build build -j --target llama-server llama-cli# Start a local OpenAI-compatible server with a web UI:
./build/bin/llama-server -hf QuantFactory/Aurora_faustus-8B-LINEAR-GGUF:# Run inference directly in the terminal:
./build/bin/llama-cli -hf QuantFactory/Aurora_faustus-8B-LINEAR-GGUF:Use Docker
docker model run hf.co/QuantFactory/Aurora_faustus-8B-LINEAR-GGUF:Quick Links
QuantFactory/Aurora_faustus-8B-LINEAR-GGUF
This is quantized version of DreadPoor/Aurora_faustus-8B-LINEAR created using llama.cpp
Original Model Card
merge
This is a merge of pre-trained language models created using mergekit.
Merge Details
Merge Method
This model was merged using the linear merge method.
Models Merged
The following models were included in the merge:
Configuration
The following YAML configuration was used to produce this model:
models:
- model: DreadPoor/Aspire-8B-model_stock
parameters:
weight: 1.0
- model: DreadPoor/WIP-TEST_PENDING_4
parameters:
weight: 1.0
- model: DreadPoor/Heart_Stolen-8B-Model_Stock
parameters:
weight: 1.0
merge_method: linear
normalize: false
int8_mask: true
dtype: bfloat16
Open LLM Leaderboard Evaluation Results
Detailed results can be found here
| Metric | Value |
|---|---|
| Avg. | 29.31 |
| IFEval (0-Shot) | 72.81 |
| BBH (3-Shot) | 36.26 |
| MATH Lvl 5 (4-Shot) | 15.18 |
| GPQA (0-shot) | 7.61 |
| MuSR (0-shot) | 12.39 |
| MMLU-PRO (5-shot) | 31.58 |
- Downloads last month
- 185
Hardware compatibility
Log In to add your hardware
2-bit
3-bit
4-bit
5-bit
6-bit
8-bit
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support
Model tree for QuantFactory/Aurora_faustus-8B-LINEAR-GGUF
Paper for QuantFactory/Aurora_faustus-8B-LINEAR-GGUF
Evaluation results
- strict accuracy on IFEval (0-Shot)Open LLM Leaderboard72.810
- normalized accuracy on BBH (3-Shot)Open LLM Leaderboard36.260
- exact match on MATH Lvl 5 (4-Shot)Open LLM Leaderboard15.180
- acc_norm on GPQA (0-shot)Open LLM Leaderboard7.610
- acc_norm on MuSR (0-shot)Open LLM Leaderboard12.390
- accuracy on MMLU-PRO (5-shot)test set Open LLM Leaderboard31.580
Install (macOS, Linux)
# Start a local OpenAI-compatible server with a web UI: llama serve -hf QuantFactory/Aurora_faustus-8B-LINEAR-GGUF:# Run inference directly in the terminal: llama cli -hf QuantFactory/Aurora_faustus-8B-LINEAR-GGUF: