RichardErkhov/AtAndDev_-_ShortKing-1.4b-v0.1-gguf

Quantization made by Richard Erkhov.

Github

Discord

Request more models

ShortKing-1.4b-v0.1 - GGUF

• Model creator: https://huggingface.co/AtAndDev/
• Original model: https://huggingface.co/AtAndDev/ShortKing-1.4b-v0.1/

Name	Quant method	Size
ShortKing-1.4b-v0.1.Q2K.gguf	Q2K	0.53GB
ShortKing-1.4b-v0.1.IQ3XS.gguf	IQ3XS	0.59GB
ShortKing-1.4b-v0.1.IQ3S.gguf	IQ3S	0.61GB
ShortKing-1.4b-v0.1.Q3KS.gguf	Q3K_S	0.61GB
ShortKing-1.4b-v0.1.IQ3M.gguf	IQ3M	0.66GB
ShortKing-1.4b-v0.1.Q3K.gguf	Q3K	0.71GB
ShortKing-1.4b-v0.1.Q3KM.gguf	Q3K_M	0.71GB
ShortKing-1.4b-v0.1.Q3KL.gguf	Q3K_L	0.77GB
ShortKing-1.4b-v0.1.IQ4XS.gguf	IQ4XS	0.74GB
ShortKing-1.4b-v0.1.Q40.gguf	Q40	0.77GB
ShortKing-1.4b-v0.1.IQ4NL.gguf	IQ4NL	0.78GB
ShortKing-1.4b-v0.1.Q4KS.gguf	Q4K_S	0.78GB
ShortKing-1.4b-v0.1.Q4K.gguf	Q4K	0.85GB
ShortKing-1.4b-v0.1.Q4KM.gguf	Q4K_M	0.85GB
ShortKing-1.4b-v0.1.Q41.gguf	Q41	0.85GB
ShortKing-1.4b-v0.1.Q50.gguf	Q50	0.92GB
ShortKing-1.4b-v0.1.Q5KS.gguf	Q5K_S	0.92GB
ShortKing-1.4b-v0.1.Q5K.gguf	Q5K	0.98GB
ShortKing-1.4b-v0.1.Q5KM.gguf	Q5K_M	0.98GB
ShortKing-1.4b-v0.1.Q51.gguf	Q51	1.0GB
ShortKing-1.4b-v0.1.Q6K.gguf	Q6K	1.08GB
ShortKing-1.4b-v0.1.Q80.gguf	Q80	1.4GB

Original model description:

---

license: cc-by-nc-4.0
datasets:

• vicgalle/alpaca-gpt4

language:

• en

---

Model Overview

Model license: cc-by-nc-4.0
This model is trained based on EleutherAI/pythia-1.4b-deduped model that is LoRA finetuned on vicgalle/alpaca-gpt4 dataset.

Prompt Template: Alpaca

<system_prompt>

Instruction:
<user_message>

Response:
<assistant_response>

Intended Use

THIS IS A TEST MODEL, IT IS NOT INTENDED FOR REAL APPLICATIONS BY ANY MEANS. HOWEVER, A NEW MODEL IS COMING IN THE SAME TOPIC.
This model series will be used for small but intense applications.

Training Details

This model took 2:31:23 to train in QLoRA on a single T4 GPU.
- epochs: 1 - train batch size: 12 - eval batch size: 12 - gradient accumulation steps: 1 - maximum gradient normal: 0.3 - learning rate: 2e-4 - weight decay: 0.001 - optimizer: pagedadamw32bit - learning rate schedule: cosine - warmup ratio (linear): 0.03