Instructions to use Boffl/BullingerLM-llama3.1-70B-pretrain-EA with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use Boffl/BullingerLM-llama3.1-70B-pretrain-EA with PEFT:
from peft import PeftModel from transformers import AutoModelForCausalLM base_model = AutoModelForCausalLM.from_pretrained("unsloth/Meta-Llama-3.1-70B-bnb-4bit") model = PeftModel.from_pretrained(base_model, "Boffl/BullingerLM-llama3.1-70B-pretrain-EA") - Notebooks
- Google Colab
- Kaggle
| { | |
| "epoch": 0.9993906154783668, | |
| "total_flos": 5.610936315118879e+18, | |
| "train_loss": 2.5786649890062288, | |
| "train_runtime": 34559.7112, | |
| "train_samples_per_second": 0.38, | |
| "train_steps_per_second": 0.012 | |
| } |