|
--- |
|
base_model: meta-llama/Meta-Llama-3.1-8B |
|
language: |
|
- en |
|
- hi |
|
datasets: |
|
- student-abdullah/BigPharma_Generic_Q-A_Format_Augemented_Hinglish_Dataset |
|
--- |
|
# LoRA Adapter Layers! |
|
# Uploaded model |
|
|
|
- **Developed by:** student-abdullah |
|
- **Finetuned from model:** meta-llama/Meta-Llama-3.1-8B |
|
- **Created on:** 27th September, 2024 |
|
- **Full model:** student-abdullah/llama3.1_medicine_hinglish_fine-tuned_27-09_8bits_gguf |
|
|
|
--- |
|
# Acknowledgement |
|
[<img src="https://raw.githubusercontent.com/unslothai/unsloth/main/images/unsloth%20made%20with%20love.png" width="200"/>](https://github.com/unslothai/unsloth) |
|
|
|
--- |
|
# Model Description |
|
This LoRA adapter layer model is fine-tuned from the meta-llama/Meta-Llama-3.1-8B base model to specialisation related to generic medications under the PMBJP scheme. The fine-tuning process included the following hyperparameters: |
|
|
|
- Fine Tuning Template: Llama 3.1 Q&A |
|
- Max Tokens: 512 |
|
- LoRA Alpha: 16 |
|
- LoRA Rank (r): 128 |
|
- Learning rate: 2e-4 |
|
- Gradient Accumulation Steps: 2 |
|
- Batch Size: 12 |
|
|
|
--- |
|
# Model Quantitative Performace |
|
- Training Quantitative Loss: 0.141 (at final 300th epoch) |
|
|
|
--- |
|
# Limitations |
|
- This is not a fully compiled model, rather just LoRA layers |
