File size: 11,658 Bytes
cb7a0bc 06dbf61 cb7a0bc 06dbf61 cb7a0bc 92329e6 cb7a0bc adb769e 70922ab 92329e6 a4d0d79 92329e6 cb7a0bc 8425fdc 70922ab 476767c a93afef 476767c 4f21e9e 476767c 467ec5c 476767c 467ec5c 476767c 467ec5c 476767c adb769e 467ec5c 476767c 467ec5c adb769e 467ec5c 71e12be adb769e 71e12be e4eedc9 467ec5c 06dbf61 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 |
---
language:
- en
license: apache-2.0
tags:
- distilabel
- dpo
- rlaif
- rlhf
datasets:
- argilla/distilabel-intel-orca-dpo-pairs
model-index:
- name: distilabeled-Hermes-2.5-Mistral-7B
results:
- task:
type: text-generation
name: Text Generation
dataset:
name: AI2 Reasoning Challenge (25-Shot)
type: ai2_arc
config: ARC-Challenge
split: test
args:
num_few_shot: 25
metrics:
- type: acc_norm
value: 66.3
name: normalized accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=argilla/distilabeled-Hermes-2.5-Mistral-7B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: HellaSwag (10-Shot)
type: hellaswag
split: validation
args:
num_few_shot: 10
metrics:
- type: acc_norm
value: 85.15
name: normalized accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=argilla/distilabeled-Hermes-2.5-Mistral-7B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: MMLU (5-Shot)
type: cais/mmlu
config: all
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 63.5
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=argilla/distilabeled-Hermes-2.5-Mistral-7B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: TruthfulQA (0-shot)
type: truthful_qa
config: multiple_choice
split: validation
args:
num_few_shot: 0
metrics:
- type: mc2
value: 55.75
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=argilla/distilabeled-Hermes-2.5-Mistral-7B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: Winogrande (5-shot)
type: winogrande
config: winogrande_xl
split: validation
args:
num_few_shot: 5
metrics:
- type: acc
value: 78.93
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=argilla/distilabeled-Hermes-2.5-Mistral-7B
name: Open LLM Leaderboard
- task:
type: text-generation
name: Text Generation
dataset:
name: GSM8k (5-shot)
type: gsm8k
config: main
split: test
args:
num_few_shot: 5
metrics:
- type: acc
value: 60.88
name: accuracy
source:
url: https://huggingface.co/spaces/HuggingFaceH4/open_llm_leaderboard?query=argilla/distilabeled-Hermes-2.5-Mistral-7B
name: Open LLM Leaderboard
---
# ⚗️ distilabeled OpenHermes 2.5 Mistral 7B
> A Neural DPO of OpenHermes 2.5, high quality matters for DPO!
<div>
<img src="https://cdn-uploads.huggingface.co/production/uploads/60420dccc15e823a685f2b03/yWdvBtKKfJdpdnPiSlNb9.png">
</div>
<p align="center">
<a href="https://github.com/argilla-io/distilabel">
<img src="https://raw.githubusercontent.com/argilla-io/distilabel/main/docs/assets/distilabel-badge-light.png" alt="Built with Distilabel" width="200" height="32"/>
</a>
</p>
## Introduction
This model is the virtual launching partner of our new open dataset [argilla/distilabel-intel-orca-dpo-pairs](https://huggingface.co/datasets/argilla/distilabel-intel-orca-dpo-pairs). It's a DPO fine tune of [OpenHermes-2.5-Mistral-7B](https://huggingface.co/teknium/OpenHermes-2.5-Mistral-7B). It outperforms the awesome `mlabonne/NeuralHermes-2.5-Mistral-7B` with the **exact same DPO recipe but using our new orca-pairs dataset**.
The dataset is a "distilabeled" version of the widely used dataset: [Intel/orca_dpo_pairs](https://huggingface.co/datasets/Intel/orca_dpo_pairs). The original dataset has been used by 100s of open source practitioners and models. We knew from fixing UltraFeedback (and before that, Alpacas and Dollys) that this dataset could be highly improved.
Continuing with our mission to build the best alignment datasets for open source LLMs and the community, we spent a few hours to improve it with [distilabel](https://github.com/argilla-io/distilabel).
The main intuition was: the original dataset just assumes gpt4/3.5-turbo are always the best response. We know from UltraFeedback that's not always the case. Moreover, DPO fine-tuning benefits from diversity of preference pairs.
This is what it took to build a real preference dataset with distilabel:
```python
from distilabel.llm import OpenAILLM
from distilabel.tasks import JudgeLMTask
from distilabel.pipeline import Pipeline
from datasets import load_dataset
dataset = load_dataset("Intel/orca_dpo_pairs", split="train")
# this shuffles the pairs to mitigate positional bias
dataset = dataset.map(lambda x: shuffle_and_track(x["chosen"], x["rejected"]))
# we use our JudgeLM implementation to rate the original pairs
labeler = OpenAILLM(
task=JudgeLMTask(),
model="gpt-4-1106-preview",
num_threads=16,
max_new_tokens=512,
)
dataset = dataset.rename_columns({"question": "input"})
distipipe = Pipeline(
labeller=labeler
)
# this computes ratings and natural language critiques for each pair
ds = distipipe.generate(dataset=dataset, num_generations=2)
```
The resulting dataset is now much more useful: we know which response is preferred (by gpt-4-turbo), which ones have low scores, and we even have natural language explanations. But what did we find? Was our intuition confirmed?
The above chart shows the following:
* ~4,000 pairs were given the same rating (a tie).
* ~7,000 pairs were correct according to our AI judge (`unchanged`).
* and ~2,000 times the rejected response was preferred (`swapped`).
Now the next question is: can we build better models with this new knowledge? The answer is "distilabeled Hermes" so let's get back to the model!
> If you love datasets as much as we do, check the [dataset](https://huggingface.co/datasets/argilla/distilabel-intel-orca-dpo-pairs) and share it with your friends and colleagues.
## Training details
As we did with [Notus](https://argilla.io/blog/notus7b/), we wanted a reproducible recipe to test the impact of data quality.
And we're lucky to have so many amazing folks in the open community contributing reproducible, easy-to-use training scripts and recipes. This time, [Maxime Labonne](https://twitter.com/maximelabonne) had shared a [Colab](https://colab.research.google.com/drive/15iFBr1xWgztXvhrj5I9fBv20c7CFOPBE?usp=sharing) to fine-tune OpenHermes with DPO and the original Intel's dataset, perfect! (funnily enough this exact recipe has been used recently to fine-tune the [top ranked 7B model](https://huggingface.co/CultriX/MistralTrix-v1)).
And that's all for the model part: we reused a good, reproducible recipe.
Once we had created the dataset, the training data part is also kind of boring: we just filtered the samples based on our intuition and with the goal of reducing the dataset size:
* Ties probably won't help the DPO tuning to learn something meaningful: both responses are similarly good or bad (filter out `ties`)
* Very good chosen responses will steer the model to generate good responses (score of chosen response >=8)
Additionally, we did some "decontamination" of gsm8k prompts (very few that were present in the train split of gsm8k).
In code, using our new dataset this translates into:
```python
from datasets import load_dataset
# Instead of this:
# dataset = load_dataset("Intel/orca_dpo_pairs", split="train")
# we did this
dataset = load_dataset("argilla/distilabel-intel-orca-dpo-pairs", split="train")
dataset = dataset.filter(
lambda r:
r["status"] != "tie" and
r["chosen_score"] >= 8 and
not r["in_gsm8k_train"]
)
```
This resulted in `5,922` instead of `12,859` samples (54% reduction) and we run it for 200 steps (using around ~3.2K samples).
## Benchmark results
For benchmarking we used the famous "Nous" or "Teknium" benchmark. You can find below an overview, including our first experiment with a less ambitious dataset filtering (removing ties and `score>5`).
For running the benchmark we used another awesome contribution from Maxime: [LLM AutoEval](https://github.com/mlabonne/llm-autoeval), check it out!
| Model | AGIEval | GPT4All | TruthfulQA | Bigbench | Average |
|-------------------------------------------------------------------------------------------------------------------|--------:|--------:|-----------:|---------:|--------:|
| [argilla/distilabeled-Hermes-2.5-Mistral-7B](https://huggingface.co/argilla/distilabeled-Hermes-2.5-Mistral-7B) | **44.64** | **73.35** | 55.96 | 42.21 | **54.04** |
| [dvilasuero/NeuralHermes-2.5-Mistral-7B-distilabel](https://huggingface.co/dvilasuero/NeuralHermes-2.5-Mistral-7B-distilabel) (first experiment) | 44.27 | 73.3 | **56.26** | **42.25** | 54.02 |
| mlabonne/NeuralHermes-2.5-Mistral-7B (original recipe) | 43.67 | 73.24 | 55.37 | 41.76 | 53.51 |
| teknium/OpenHermes-2.5-Mistral-7B | 42.75 | 72.99 | 52.99 | 40.94 | 52.42|
> Update: we now include llm-harness results too!
| Model | ARC | HellaSwag | MMLU | TruthfulQA | Winogrande | GSM8K |
|------------------------------------------------------|-------|-----------|------|-----------:|------------|-------|
| [argilla/distilabeled-Hermes-2.5-Mistral-7B](https://huggingface.co/argilla/distilabeled-Hermes-2.5-Mistral-7B) | 66.04 | **85.07** | Pending | 55.96 | **79.56** | **66.34** |
| [dvilasuero/NeuralHermes-2.5-Mistral-7B-distilabel](https://huggingface.co/dvilasuero/NeuralHermes-2.5-Mistral-7B-distilabel) | 65.36 | 84.74 | Pending | **56.26** | 79.24 | 65.13 |
| [mlabonne/NeuralHermes-2.5-Mistral-7B](https://huggingface.co/mlabonne/NeuralHermes-2.5-Mistral-7B) | **66.55** | 84.90 | **63.32** | 54.93 | 78.30 | 61.30 |
| [teknium/OpenHermes-2.5-Mistral-7B](https://huggingface.co/teknium/OpenHermes-2.5-Mistral-7B) | 64.93 | 84.18 | 63.64 | 52.24 | 78.06 | 26.08 |
### Training Hardware
We used 1 x A100 40GB in runpod for less than 1 hour.
## Acknowledgements
We'd like to thank the amazing open community and in particular:
* The Intel team for publishing a great open dataset and show how well it worked in the first place
* Teknium and NousResearch for their awesome work and models.
* Maxime for sharing such great resources.
# [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_argilla__distilabeled-Hermes-2.5-Mistral-7B)
| Metric |Value|
|---------------------------------|----:|
|Avg. |68.42|
|AI2 Reasoning Challenge (25-Shot)|66.30|
|HellaSwag (10-Shot) |85.15|
|MMLU (5-Shot) |63.50|
|TruthfulQA (0-shot) |55.75|
|Winogrande (5-shot) |78.93|
|GSM8k (5-shot) |60.88|
|