File size: 2,354 Bytes
c2ad5bc 80f36df c2ad5bc 80f36df c2ad5bc 80f36df c2ad5bc 8ff565b c2ad5bc 80f36df c2ad5bc 80f36df c2ad5bc 80f36df c2ad5bc 0e574df ae9f8ee c2ad5bc 80f36df c2ad5bc ae9f8ee 80f36df c2ad5bc 80f36df c2ad5bc 80f36df c2ad5bc 80f36df |
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 |
---
license: cc-by-nc-4.0
base_model: mlabonne/NeuralMonarch-7B
tags:
- generated_from_trainer
- axolotl
- mistral
- instruct
- finetune
- chatml
- gpt4
- synthetic data
- distillation
model-index:
- name: AlphaMonarch-laser
results: []
datasets:
- argilla/OpenHermes2.5-dpo-binarized-alpha
language:
- en
library_name: transformers
pipeline_tag: text-generation
---
<!-- This model card has been generated automatically according to the information the Trainer had access to. You
should probably proofread and complete it, then remove this comment. -->
# AlphaMonarch-daser
AlphaMonarch-daser is a mixture of two techniques that are LaserQlora and Dora. This model is a DPO fine-tuned of [mlabonne/NeuralMonarch-7B](https://huggingface.co/mlabonne/NeuralMonarch-7B/) using the [argilla/OpenHermes2.5-dpo-binarized-alpha](https://huggingface.co/datasets/argilla/OpenHermes2.5-dpo-binarized-alpha) preference dataset. I have fine-tuned this model only on half of the projections, but have achieved better results as compared to the version released [AlphaMonarch-dora](https://huggingface.co/abideen/AlphaMonarch-dora). I have trained this model for 1080 steps. Comparison of AlphaMonarch, AlphaMonarch-laser, AlphaMonarch-daser, and AlphaMonarch-dora on the OpenLLM leaderboard are:
## 🏆 Evaluation results
On YALL leaderboard: AlphaMonarch-daser > AlphaMonarch-dora > AlphaMonarch > AlphaMonarch-laser
On OpenLLM bench: AlphaMonarch-laser > AlphaMonarch > AlphaMonarch-daser > AlphaMonarch-dora
### Training hyperparameters
The following hyperparameters were used during training:
- learning_rate: 5e-07
- train_batch_size: 1
- eval_batch_size: 8
- seed: 42
- gradient_accumulation_steps: 8
- total_train_batch_size: 8
- optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
- lr_scheduler_type: cosine
- lr_scheduler_warmup_steps: 100
- training_steps: 1080
### Framework versions
- Transformers 4.38.0.dev0
- Pytorch 2.1.2+cu118
- Datasets 2.17.0
- Tokenizers 0.15.0 |