You can deploy this model and start earning money today!

RichardErkhov/
INSAIT-Institute_-_BgGPT-7B-Instruct-v0.2-gguf

Name: INSAIT-Institute_-_BgGPT-7B-Instruct-v0.2-gguf
Availability: InStock

Views

No views yet

Quantization made by Richard Erkhov.

Github

Discord

Request more models

BgGPT-7B-Instruct-v0.2 - GGUF

Model creator: https://huggingface.co/INSAIT-Institute/
Original model: https://huggingface.co/INSAIT-Institute/BgGPT-7B-Instruct-v0.2/

Name	Quant method	Size
BgGPT-7B-Instruct-v0.2.Q2_K.gguf	Q2_K	2.56GB
BgGPT-7B-Instruct-v0.2.IQ3_XS.gguf	IQ3_XS	2.84GB
BgGPT-7B-Instruct-v0.2.IQ3_S.gguf	IQ3_S	2.99GB
BgGPT-7B-Instruct-v0.2.Q3_K_S.gguf	Q3_K_S	2.98GB
BgGPT-7B-Instruct-v0.2.IQ3_M.gguf	IQ3_M	3.09GB
BgGPT-7B-Instruct-v0.2.Q3_K.gguf	Q3_K	3.31GB
BgGPT-7B-Instruct-v0.2.Q3_K_M.gguf	Q3_K_M	3.31GB
BgGPT-7B-Instruct-v0.2.Q3_K_L.gguf	Q3_K_L	3.59GB
BgGPT-7B-Instruct-v0.2.IQ4_XS.gguf	IQ4_XS	3.7GB
BgGPT-7B-Instruct-v0.2.Q4_0.gguf	Q4_0	3.86GB
BgGPT-7B-Instruct-v0.2.IQ4_NL.gguf	IQ4_NL	3.9GB
BgGPT-7B-Instruct-v0.2.Q4_K_S.gguf	Q4_K_S	3.89GB
BgGPT-7B-Instruct-v0.2.Q4_K.gguf	Q4_K	4.1GB
BgGPT-7B-Instruct-v0.2.Q4_K_M.gguf	Q4_K_M	4.1GB
BgGPT-7B-Instruct-v0.2.Q4_1.gguf	Q4_1	4.27GB
BgGPT-7B-Instruct-v0.2.Q5_0.gguf	Q5_0	4.69GB
BgGPT-7B-Instruct-v0.2.Q5_K_S.gguf	Q5_K_S	4.69GB
BgGPT-7B-Instruct-v0.2.Q5_K.gguf	Q5_K	4.81GB
BgGPT-7B-Instruct-v0.2.Q5_K_M.gguf	Q5_K_M	4.81GB
BgGPT-7B-Instruct-v0.2.Q5_1.gguf	Q5_1	5.1GB
BgGPT-7B-Instruct-v0.2.Q6_K.gguf	Q6_K	5.57GB
BgGPT-7B-Instruct-v0.2.Q8_0.gguf	Q8_0	7.22GB

Original model description:

base_model: mistralai/Mistral-7B-v0.1 tags:

mistral
instruct
bggpt
insait language:
bg
en library_name: transformers pipeline_tag: text-generation license: apache-2.0

INSAIT-Institute/BgGPT-7B-Instruct-v0.2

Meet BgGPT-7B, a Bulgarian language model trained from mistralai/Mistral-7B-v0.1. BgGPT is distributed under Apache 2.0 license.

This model was created by INSAIT Institute, part of Sofia University, in Sofia, Bulgaria.

This is an improved version of the model - v0.2.

Model description

The model is continously pretrained to gain its Bulgarian language and culture capabilities using multiple datasets, including Bulgarian web crawl data, a range of specialized Bulgarian datasets sourced by INSAIT Institute, and machine translations of popular English datasets. This Bulgarian data was augmented with English datasets to retain English and logical reasoning skills.

The model's tokenizer has been extended to allow for a more efficient encoding of Bulgarian words written in Cyrillic. This not only increases throughput of Cyrillic text but also performance.

Instruction format

In order to leverage instruction fine-tuning, your prompt should be surrounded by [INST] and [/INST] tokens. The very first instruction should begin with a begin of sequence token <s>. Following instructions should not. The assistant generation will be ended by the end-of-sequence token.

E.g.

text = "<s>[INST] Кога е основан Софийският университет? [/INST]"
"Софийският университет „Св. Климент Охридски“ е създаден на 1 октомври 1888 г.</s> "
"[INST] Кой го е основал? [/INST]"

This format is available as a chat template via the apply_chat_template() method:

Benchmarks

The model comes with a set of Benchmarks that are translations of the corresponding English-benchmarks. These are provided at https://github.com/insait-institute/lm-evaluation-harness-bg

As this is an improved version over version 0.1 of the same model and we include benchmark comparisons.

Summary

Finetuned from: mistralai/Mistral-7B-v0.1
Model type: Causal decoder-only transformer language model
Language: Bulgarian and English
License: Apache 2.0
Contact: bggpt@insait.ai

Use in 🤗Transformers

First install direct dependencies:

pip install transformers torch accelerate

If you want faster inference using flash-attention2, you need to install these dependencies:

bash

pip install packaging ninja
pip install flash-attn

Then load the model in transformers:

python

from transformers import AutoModelForCausalLM, AutoTokenizer
import torch
model = AutoModelForCausalLM.from_pretrained(
    model="INSAIT-Institute/BgGPT-7B-Instruct-v0.2",
    device_map="auto",
    torch_dtype=torch.bfloat16,
    use_flash_attn_2=True # optional
)

Use with GGML / llama.cpp

The model in GGUF format INSAIT-Institute/BgGPT-7B-Instruct-v0.2-GGUF

RichardErkhov/INSAIT-Institute_-_BgGPT-7B-Instruct-v0.2-gguf