Ericu950/Epigr_3_Llama-3.1-8B-Instruct_text
TEXT GENERATIONConcurrency Cost:1Model Size:8BQuant:FP8Ctx Length:32kPublished:Nov 16, 2024Architecture:Transformer Cold

Eicu950/Epigr_3_Llama-3.1-8B-Instruct_text is an 8 billion parameter instruction-tuned language model, fine-tuned from Ericu950/Papy_2_Llama-3.1-8B-Instruct_text. This model was trained on the augmented Ericu950/Inscriptions_2 dataset, using an 80/10/10 split for training, testing, and validation. It serves as an alternative to Ericu950/Epigr_2_Llama-3.1-8B-Instruct_text, focusing on specialized text generation based on its unique training data.

Loading preview...

Model Overview

Eicu950/Epigr_3_Llama-3.1-8B-Instruct_text is an 8 billion parameter instruction-tuned language model developed by Ericu950. It is an alternative iteration to the previously released Ericu950/Epigr_2_Llama-3.1-8B-Instruct_text.

Training Details

This model was fine-tuned starting from the base model Ericu950/Papy_2_Llama-3.1-8B-Instruct_text. The training utilized the augmented dataset Ericu950/Inscriptions_2, which was split into an 80% training set, 10% test set, and 10% validation set. Specifically:

  • Test Set: Inscriptions with PHI IDs ending in 3.
  • Validation Set: Inscriptions with PHI IDs ending in 4.
  • Training Set: All other inscriptions.

Key Characteristics

This model is designed for specialized text generation, leveraging its unique training on the Inscriptions_2 dataset. For detailed inference guidelines and further technical specifications, users should refer to the model card for Ericu950/Epigr_2_Llama-3.1-8B-Instruct_text, as this model builds upon that foundation.