Exllama v2 Quantizations of Einstein-v3-7B
Using turboderp's ExLlamaV2 v0.0.13 for quantization.
The "main" branch only contains the measurement.json, download one of the other branches for the model (see below)
Each branch contains an individual bits per weight, with the main one containing only the meaurement.json for further conversions.
Original model: https://huggingface.co/Weyaxi/Einstein-v3-7B
Branch | Bits | lm_head bits | VRAM (4k) | VRAM (16k) | VRAM (32k) | Description |
---|---|---|---|---|---|---|
8_0 | 8.0 | 8.0 | 8.4 GB | 9.8 GB | 11.8 GB | Maximum quality that ExLlamaV2 can produce, near unquantized performance. |
6_5 | 6.5 | 8.0 | 7.2 GB | 8.6 GB | 10.6 GB | Very similar to 8.0, good tradeoff of size vs performance, recommended. |
5_0 | 5.0 | 6.0 | 6.0 GB | 7.4 GB | 9.4 GB | Slightly lower quality vs 6.5, but usable on 8GB cards. |
4_25 | 4.25 | 6.0 | 5.3 GB | 6.7 GB | 8.7 GB | GPTQ equivalent bits per weight, slightly higher quality. |
3_5 | 3.5 | 6.0 | 4.7 GB | 6.1 GB | 8.1 GB | Lower quality, only use if you have to. |
Download instructions
With git:
git clone --single-branch --branch 6_5 https://huggingface.co/bartowski/Einstein-v3-7B-exl2 Einstein-v3-7B-exl2-6_5
With huggingface hub (credit to TheBloke for instructions):
pip3 install huggingface-hub
To download the main
(only useful if you only care about measurement.json) branch to a folder called Einstein-v3-7B-exl2
:
mkdir Einstein-v3-7B-exl2
huggingface-cli download bartowski/Einstein-v3-7B-exl2 --local-dir Einstein-v3-7B-exl2 --local-dir-use-symlinks False
To download from a different branch, add the --revision
parameter:
Linux:
mkdir Einstein-v3-7B-exl2-6_5
huggingface-cli download bartowski/Einstein-v3-7B-exl2 --revision 6_5 --local-dir Einstein-v3-7B-exl2-6_5 --local-dir-use-symlinks False
Windows (which apparently doesn't like _ in folders sometimes?):
mkdir Einstein-v3-7B-exl2-6.5
huggingface-cli download bartowski/Einstein-v3-7B-exl2 --revision 6_5 --local-dir Einstein-v3-7B-exl2-6.5 --local-dir-use-symlinks False
Want to support my work? Visit my ko-fi page here: https://ko-fi.com/bartowski
Finetuned from
Evaluation results
- normalized accuracy on AI2 Reasoning Challenge (25-Shot)test set Open LLM Leaderboard62.290
- normalized accuracy on HellaSwag (10-Shot)validation set Open LLM Leaderboard83.010
- accuracy on MMLU (5-Shot)test set Open LLM Leaderboard63.320
- mc2 on TruthfulQA (0-shot)validation set Open LLM Leaderboard51.180
- accuracy on Winogrande (5-shot)validation set Open LLM Leaderboard79.950
- accuracy on GSM8k (5-shot)test set Open LLM Leaderboard44.810