munish0838's picture
Update README.md
47323ba verified
|
raw
history blame
2.12 kB
metadata
license: apache-2.0
datasets:
  - Locutusque/hercules-v5.0
base_model: M4-ai/Hercules-5.0-Qwen2-1.5B
language:
  - en
inference:
  parameters:
    do_sample: true
    temperature: 0.8
    top_p: 0.95
    top_k: 40
    min_p: 0.1
    max_new_tokens: 250
    repetition_penalty: 1.1
pipeline_tag: text-generation

Hercules-5.0-Qwen2-1.5B-GGUF

This is quantized version of M4-ai/Hercules-5.0-Qwen2-1.5B created using llama.cpp

Model Description

We fine-tuned qwen2-1.5B on a high quality mix for general-purpose assistants. A DPO version of this will be released soon. We use the ChatML prompt format.

Model Details

This model has capabilities in math, coding, writing, and more. We fine-tuned it using a high quality mix for general-purpose assistants.

  • Developed by: M4-ai
  • Language(s) (NLP): English and maybe Chinese
  • License: apache-2.0
  • Finetuned from model: qwen2-1.5B

Uses

General purpose assistant, question answering, chain-of-thought, etc..

This language model made an impressive achievement, and correctly implemented a Multi Head Attention for use in a transformer neural network.

Recommendations

Users (both direct and downstream) should be made aware of the risks, biases and limitations of the model. More information needed for further recommendations.

Training Details

Training Data

  • Locutusque/hercules-v5.0

Evaluations

coming soon

Training Hyperparameters

  • Training regime: bf16 non-mixed precision

Technical Specifications

Hardware

We used 8 Kaggle TPUs, and we trained at a global batch size of 256 and sequence length of 1536.