You can deploy this model and start earning money today!

RichardErkhov/
kittn_-_mistral-7B-v0.1-hf-gguf

Name: kittn_-_mistral-7B-v0.1-hf-gguf
Availability: InStock

endpoints_compatible

gguf

template

Views

No views yet

Quantization made by Richard Erkhov.

Github

Discord

Request more models

mistral-7B-v0.1-hf - GGUF

Model creator: https://huggingface.co/kittn/
Original model: https://huggingface.co/kittn/mistral-7B-v0.1-hf/

Name	Quant method	Size
mistral-7B-v0.1-hf.Q2_K.gguf	Q2_K	2.53GB
mistral-7B-v0.1-hf.IQ3_XS.gguf	IQ3_XS	2.81GB
mistral-7B-v0.1-hf.IQ3_S.gguf	IQ3_S	2.96GB
mistral-7B-v0.1-hf.Q3_K_S.gguf	Q3_K_S	2.95GB
mistral-7B-v0.1-hf.IQ3_M.gguf	IQ3_M	3.06GB
mistral-7B-v0.1-hf.Q3_K.gguf	Q3_K	3.28GB
mistral-7B-v0.1-hf.Q3_K_M.gguf	Q3_K_M	3.28GB
mistral-7B-v0.1-hf.Q3_K_L.gguf	Q3_K_L	3.56GB
mistral-7B-v0.1-hf.IQ4_XS.gguf	IQ4_XS	3.67GB
mistral-7B-v0.1-hf.Q4_0.gguf	Q4_0	3.83GB
mistral-7B-v0.1-hf.IQ4_NL.gguf	IQ4_NL	3.87GB
mistral-7B-v0.1-hf.Q4_K_S.gguf	Q4_K_S	3.86GB
mistral-7B-v0.1-hf.Q4_K.gguf	Q4_K	4.07GB
mistral-7B-v0.1-hf.Q4_K_M.gguf	Q4_K_M	4.07GB
mistral-7B-v0.1-hf.Q4_1.gguf	Q4_1	4.24GB
mistral-7B-v0.1-hf.Q5_0.gguf	Q5_0	4.65GB
mistral-7B-v0.1-hf.Q5_K_S.gguf	Q5_K_S	4.65GB
mistral-7B-v0.1-hf.Q5_K.gguf	Q5_K	4.78GB
mistral-7B-v0.1-hf.Q5_K_M.gguf	Q5_K_M	4.78GB
mistral-7B-v0.1-hf.Q5_1.gguf	Q5_1	5.07GB
mistral-7B-v0.1-hf.Q6_K.gguf	Q6_K	5.53GB
mistral-7B-v0.1-hf.Q8_0.gguf	Q8_0	7.17GB

Original model description:

UPDATE: Official version is out, use it instead: https://huggingface.co/mistralai/Mistral-7B-v0.1

mistral-7B-v0.1-hf

Huggingface compatible version of Mistral's 7B model: https://twitter.com/MistralAI/status/1706877320844509405

Usage

Load in bfloat16 (16GB VRAM or higher)

python

import torch
from transformers import LlamaForCausalLM, LlamaTokenizer, pipeline, TextStreamer

tokenizer = LlamaTokenizer.from_pretrained("kittn/mistral-7B-v0.1-hf")
model = LlamaForCausalLM.from_pretrained(
    "kittn/mistral-7B-v0.1-hf",
    torch_dtype=torch.bfloat16,
    device_map={"": 0}
)

pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)

pipe("Hi, my name", streamer=TextStreamer(tokenizer), max_new_tokens=128)

Load in bitsandbytes nf4 (6GB VRAM or higher, maybe less with double_quant)

python

import torch
from transformers import LlamaForCausalLM, LlamaTokenizer, pipeline, TextStreamer, BitsAndBytesConfig

tokenizer = LlamaTokenizer.from_pretrained("kittn/mistral-7B-v0.1-hf")
model = LlamaForCausalLM.from_pretrained(
    "kittn/mistral-7B-v0.1-hf",
    device_map={"": 0},
    quantization_config=BitsAndBytesConfig(
        load_in_4bit=True,
        bnb_4bit_compute_dtype=torch.float16,
        bnb_4bit_quant_type="nf4",
        bnb_4bit_use_double_quant=False, # set to True to save more VRAM at the cost of some speed/accuracy
    ),
)

pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)

pipe("Hi, my name", streamer=TextStreamer(tokenizer), max_new_tokens=128)

Load in bitsandbytes int8 (8GB VRAM or higher). Quite slow; not recommended.

python

import torch
from transformers import LlamaForCausalLM, LlamaTokenizer, pipeline, TextStreamer, BitsAndBytesConfig

tokenizer = LlamaTokenizer.from_pretrained("kittn/mistral-7B-v0.1-hf")
model = LlamaForCausalLM.from_pretrained(
    "kittn/mistral-7B-v0.1-hf",
    device_map={"": 0},
    quantization_config=BitsAndBytesConfig(
        load_in_8bit=True,
    ),
)

pipe = pipeline("text-generation", model=model, tokenizer=tokenizer)

pipe("Hi, my name", streamer=TextStreamer(tokenizer), max_new_tokens=128)

Notes

The original huggingface conversion script converts the model from bf16 to fp16 before saving it. This script doesn't
The tokenizer is created with legacy=False, more about this here
Saved in safetensors format

Conversion script [link]

Unlike meta-llama/Llama-2-7b, this model uses GQA. This breaks some assumptions in the original conversion script, requiring a few changes.

Conversion script: link

Original conversion script: link

RichardErkhov/kittn_-_mistral-7B-v0.1-hf-gguf

UPDATE: Official version is out, use it instead: https://huggingface.co/mistralai/Mistral-7B-v0.1

mistral-7B-v0.1-hf

Usage

Load in bfloat16 (16GB VRAM or higher)

Load in bitsandbytes nf4 (6GB VRAM or higher, maybe less with double_quant)

Load in bitsandbytes int8 (8GB VRAM or higher). Quite slow; not recommended.

Notes

Conversion script [link]

RichardErkhov/
kittn_-_mistral-7B-v0.1-hf-gguf