File size: 4,175 Bytes
35471b2 19694dc 35471b2 f4f2d9c 35471b2 60bccbe 35471b2 639cca3 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 |
---
license: cc-by-nc-4.0
---
Replaced Zephyr by Airoboros 2.2 and OpenOrca by SynthIA in the mix, the reason why is to see if using merged Mistral models using all the same prompt format would be a better step or not.
## Description
This repo contains fp16 files of Mistral-11B-SynthIAirOmniMix.
## Model used
- [SynthIA-7B-v1.5](https://huggingface.co/migtissera/SynthIA-7B-v1.5)
- [Mistral-7B-v0.1-Open-Platypus](https://huggingface.co/akjindal53244/Mistral-7B-v0.1-Open-Platypus)
- [CollectiveCognition-v1.1-Mistral-7B](https://huggingface.co/teknium/CollectiveCognition-v1.1-Mistral-7B)
- [airoboros-mistral2.2-7b](https://huggingface.co/teknium/airoboros-mistral2.2-7b)
## Prompt template
3 out of 4 models use the same prompting format in this merge.
The best one should be this one, since Zephyr and OpenOrca is out of the merge:
```
(SYSTEM: {context}) - Not mandatory
USER: {prompt}
ASSISTANT:
```
But this one (maybe) work too:
```
Below is an instruction that describes a task. Write a response that appropriately completes the request.
### Instruction:
{prompt}
### Response:
```
## The secret sauce
Mistral-11B-SynthIAOpenPlatypus :
```
slices:
- sources:
- model: "/content/drive/MyDrive/SynthIA-7B-v1.5-bf16"
layer_range: [0, 24]
- sources:
- model: akjindal53244/Mistral-7B-v0.1-Open-Platypus
layer_range: [8, 32]
merge_method: passthrough
dtype: bfloat16
```
Mistral-11B-CC-Airo :
```
slices:
- sources:
- model: "/content/drive/MyDrive/CC-v1.1-7B-bf16"
layer_range: [0, 24]
- sources:
- model: "/content/drive/MyDrive/Mistral-7B-Airoboros-2.2-bf16"
layer_range: [8, 32]
merge_method: passthrough
dtype: bfloat16
```
Mistral-11B-SynthIAirOmniMix :
```
slices:
- sources:
- model: Mistral-11B-SynthIAOpenPlatypus
layer_range: [0, 48]
- model: Mistral-11B-CC-Airo
layer_range: [0, 48]
merge_method: slerp
base_model: Mistral-11B-OpenOrcaPlatypus
parameters:
t:
- filter: lm_head
value: [0.75]
- filter: embed_tokens
value: [0.75]
- filter: self_attn
value: [0.75, 0.25]
- filter: mlp
value: [0.25, 0.75]
- filter: layernorm
value: [0.5, 0.5]
- filter: modelnorm
value: [0.75]
- value: 0.5 # fallback for rest of tensors
dtype: bfloat16
```
I use [mergekit](https://github.com/cg123/mergekit) for all the manipulation told here.
## Some scoring I done myself
| Task |Version| Metric |Value | |Stderr|
|-------------|------:|--------|-----:|---|-----:|
|arc_challenge| 0|acc |0.5410|± |0.0146|
| | |acc_norm|0.5640|± |0.0145|
|arc_easy | 0|acc |0.8228|± |0.0078|
| | |acc_norm|0.8068|± |0.0081|
|hellaswag | 0|acc |0.6274|± |0.0048|
| | |acc_norm|0.8167|± |0.0039|
|piqa | 0|acc |0.8052|± |0.0092|
| | |acc_norm|0.8232|± |0.0089|
|truthfulqa_mc| 1|mc1 |0.3905|± |0.0171|
| | |mc2 |0.5592|± |0.0155|
|winogrande | 0|acc |0.7364|± |0.0124|
## Others
Special thanks to Sushi, [Henky](https://github.com/KoboldAI/KoboldAI-Client) for the machine he give me for big task, and [Charles Goddard](https://github.com/cg123) for his amazing tool.
If you want to support me, you can [here](https://ko-fi.com/undiai).
# [Open LLM Leaderboard Evaluation Results](https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard)
Detailed results can be found [here](https://huggingface.co/datasets/open-llm-leaderboard/details_NeverSleep__Mistral-11B-SynthIAirOmniMix)
| Metric | Value |
|-----------------------|---------------------------|
| Avg. | 54.56 |
| ARC (25-shot) | 62.46 |
| HellaSwag (10-shot) | 83.13 |
| MMLU (5-shot) | 63.47 |
| TruthfulQA (0-shot) | 55.69 |
| Winogrande (5-shot) | 76.4 |
| GSM8K (5-shot) | 11.9 |
| DROP (3-shot) | 28.88 |
|