Xinging/llama2-13b_sft_0.1_ratio_alpaca_gpt4_proj_by_human_eval_ntrain_378

TEXT GENERATIONConcurrency Cost:1Model Size:13BQuant:FP8Ctx Length:4kPublished:Feb 11, 2025License:otherArchitecture:Transformer Cold

The Xinging/llama2-13b_sft_0.1_ratio_alpaca_gpt4_proj_by_human_eval_ntrain_378 model is a 13 billion parameter language model fine-tuned from Meta's Llama-2-13b-hf. It was fine-tuned on the 0.1_ratio_alpaca_gpt4_proj_by_human_eval_ntrain_378 dataset, suggesting a focus on instruction-following and potentially code-related tasks. This model is designed for applications requiring a Llama 2-based instruction-tuned model with a 4096-token context length.

Loading preview...

Model Overview

Xinging/llama2-13b_sft_0.1_ratio_alpaca_gpt4_proj_by_human_eval_ntrain_378 is a 13 billion parameter language model derived from the meta-llama/Llama-2-13b-hf base model. It has been fine-tuned using the 0.1_ratio_alpaca_gpt4_proj_by_human_eval_ntrain_378 dataset, indicating an instruction-following specialization, potentially with a focus on code generation or evaluation based on the dataset name.

Training Details

The model was trained with the following key hyperparameters:

  • Learning Rate: 2e-05
  • Batch Size: 16 (train), 8 (eval)
  • Optimizer: AdamW with betas=(0.9, 0.999) and epsilon=1e-08
  • Scheduler: Cosine learning rate scheduler with a 0.03 warmup ratio
  • Epochs: 1.0

Intended Use Cases

This model is suitable for applications that benefit from a Llama 2-based instruction-tuned model, particularly those requiring a 13 billion parameter model with a 4096-token context window. Its fine-tuning dataset suggests potential strengths in tasks related to instruction following and possibly code-centric prompts, though specific performance metrics are not provided in the README.