Feedback and support: TensorBlock's Twitter/X, Telegram Group and Discord server

acrastt/RedPajama-INCITE-Chat-Instruct-3B-V1 - GGUF

This repo contains GGUF format model files for acrastt/RedPajama-INCITE-Chat-Instruct-3B-V1.

The files were quantized using machines provided by TensorBlock, and they are compatible with llama.cpp as of commit b4011.

Run them on the TensorBlock client using your local machine ↗

Prompt template

Model file specification

Filename	Quant type	File Size	Description
RedPajama-INCITE-Chat-Instruct-3B-V1-Q2_K.gguf	Q2_K	1.012 GB	smallest, significant quality loss - not recommended for most purposes
RedPajama-INCITE-Chat-Instruct-3B-V1-Q3_K_S.gguf	Q3_K_S	1.163 GB	very small, high quality loss
RedPajama-INCITE-Chat-Instruct-3B-V1-Q3_K_M.gguf	Q3_K_M	1.377 GB	very small, high quality loss
RedPajama-INCITE-Chat-Instruct-3B-V1-Q3_K_L.gguf	Q3_K_L	1.493 GB	small, substantial quality loss
RedPajama-INCITE-Chat-Instruct-3B-V1-Q4_0.gguf	Q4_0	1.490 GB	legacy; small, very high quality loss - prefer using Q3_K_M
RedPajama-INCITE-Chat-Instruct-3B-V1-Q4_K_S.gguf	Q4_K_S	1.502 GB	small, greater quality loss
RedPajama-INCITE-Chat-Instruct-3B-V1-Q4_K_M.gguf	Q4_K_M	1.664 GB	medium, balanced quality - recommended
RedPajama-INCITE-Chat-Instruct-3B-V1-Q5_0.gguf	Q5_0	1.798 GB	legacy; medium, balanced quality - prefer using Q4_K_M
RedPajama-INCITE-Chat-Instruct-3B-V1-Q5_K_S.gguf	Q5_K_S	1.798 GB	large, low quality loss - recommended
RedPajama-INCITE-Chat-Instruct-3B-V1-Q5_K_M.gguf	Q5_K_M	1.928 GB	large, very low quality loss - recommended
RedPajama-INCITE-Chat-Instruct-3B-V1-Q6_K.gguf	Q6_K	2.126 GB	very large, extremely low quality loss
RedPajama-INCITE-Chat-Instruct-3B-V1-Q8_0.gguf	Q8_0	2.751 GB	very large, extremely low quality loss - not recommended

Downloading instruction

Command line

Firstly, install Huggingface Client

pip install -U "huggingface_hub[cli]"

Then, downoad the individual model file the a local directory

huggingface-cli download tensorblock/RedPajama-INCITE-Chat-Instruct-3B-V1-GGUF --include "RedPajama-INCITE-Chat-Instruct-3B-V1-Q2_K.gguf" --local-dir MY_LOCAL_DIR

If you wanna download multiple model files with a pattern (e.g., *Q4_K*gguf), you can try:

huggingface-cli download tensorblock/RedPajama-INCITE-Chat-Instruct-3B-V1-GGUF --local-dir MY_LOCAL_DIR --local-dir-use-symlinks False --include='*Q4_K*gguf'

Downloads last month: 135

GGUF

Model size

2.78B params

Architecture

gptneox

2-bit

3-bit

4-bit

5-bit

6-bit

8-bit

Inference Examples

Text Generation

This model does not have enough activity to be deployed to Inference API (serverless) yet. Increase its social visibility and check back later, or deploy to Inference Endpoints (dedicated) instead.

Model tree for tensorblock/RedPajama-INCITE-Chat-Instruct-3B-V1-GGUF

Base model

acrastt/RedPajama-INCITE-Chat-Instruct-3B-V1

Quantized

(3)

this model

Datasets used to train tensorblock/RedPajama-INCITE-Chat-Instruct-3B-V1-GGUF

Evaluation results

normalized accuracy on AI2 Reasoning Challenge (25-Shot)
test set Open LLM Leaderboard

42.580
normalized accuracy on HellaSwag (10-Shot)
validation set Open LLM Leaderboard

67.480
accuracy on MMLU (5-Shot)
test set Open LLM Leaderboard

25.990
mc2 on TruthfulQA (0-shot)
validation set Open LLM Leaderboard

33.620
accuracy on Winogrande (5-shot)
validation set Open LLM Leaderboard

64.800
accuracy on GSM8k (5-shot)
test set Open LLM Leaderboard

0.910

View on Papers With Code