BigMistral-11b-GLUE_LORA

<img src="https://raw.githubusercontent.com/OpenAccess-AI-Collective/axolotl/main/image/axolotl-badge-web.png" alt="Built with Axolotl" width="200" height="32"/>

<p align="center"><font size="5"> <b>Regret: Should have not targeted Q, V, K, O; as those are less impactful for "healing" but more impactful on performance otherwise. Still works great!</b> </font></p>

qlora

This model is a fine-tuned version of athirdpath/BigMistral-11b on the athirdpath/Merge_Glue dataset. It achieves the following results on the evaluation set:

• Loss: 0.9174

<p align="center"><font size="7"> <b>Before and After Example</b></font></p> <p align="center"><font size="4"> <b>Example model is athirdpath/CleverMage-11b</b></font></p> <p align="center"><font size="5"> <b>Example with LoRA (min_p, alpaca)</b></font></p> <p align="center"><img src="https://iili.io/JzsmqzJ.png"/> <p align="center"><font size="5"> <b>Example without LoRA (min_p, chatML)</b></font></p> <p align="center"><img src="https://iili.io/JzsmCsR.png"/>

Training hyperparameters

The following hyperparameters were used during training:

• learning_rate: 0.0005
• train_batch_size: 10
• eval_batch_size: 10
• seed: 42
• gradient_accumulation_steps: 4
• total_train_batch_size: 40
• optimizer: Adam with betas=(0.9,0.999) and epsilon=1e-08
• lr_scheduler_type: cosine
• lr_scheduler_warmup_steps: 10
• num_epochs: 3

Training results

Framework versions

• Transformers 4.35.2
• Pytorch 2.0.1+cu118
• Datasets 2.15.0
• Tokenizers 0.15.0