gabor-hosu / feature-extraction updated 8 months ago

e5-mistral-7b-instruct-bnb-4bit

Name: e5-mistral-7b-instruct-bnb-4bit
Author: gabor-hosu

This model is a quantized version of the original model intfloat/e5-mistral-7b-instruct.

Params

7.3 B

Context

32,768

Downloads 30d

1.0 M

Likes

Commercial use: allowed mit Not gated SAFETENSORS 1 languages View on Hugging Face ↗

Download history

daily snapshots · 25 days

1.0 M442 K

Aug 27Sep 4Sep 12Sep 20

Estimated VRAM at 8K context unless noted. Pick your hardware to see the verdict per quantization.

File	Quant	Size	Est. VRAM	Verdict on RTX 4090 · 24 GB
model.safetensors	u8	3.9 GB	5.8 GB	✅ Runs comfortably
model.safetensors (bf16, full)	bf16 + 32K ctx	3.9 GB	9.1 GB	✅ Runs comfortably

Estimate: file size × 1.1 + KV cache at 8K + 0.5 GB overhead. Not a benchmark — how we calculate this.

Architecture: MistralModel
Parameters: 7.3 B
Tensor type: U8
Context length: 32,768
Vocabulary: 32,000
Layers / heads: 32 / 32
Licence: mit
First seen on the Hub: 2026-01-09
Base model: e5-mistral-7b-instruct
Training datasets: undisclosed
MTEB ATEC (reported): 42.657334751856
MTEB AFQMC (reported): 38.709750161904
MTEB AmazonPolarityClassification (reported): 95.904856347086
MTEB AmazonReviewsClassification (de) (reported): 52.156230111545
MTEB AmazonReviewsClassification (en) (reported): 55.312119958151
MTEB AmazonReviewsClassification (es) (reported): 49.195023008878
MTEB AmazonReviewsClassification (fr) (reported): 48.434470184108
MTEB AmazonReviewsClassification (ja) (reported): 48.686
MTEB AmazonCounterfactualClassification (de) (reported): 72.143882579058
MTEB AmazonCounterfactualClassification (en) (reported): 72.372077035326
MTEB AmazonCounterfactualClassification (ja) (reported): 63.877335968499
MTEB AmazonCounterfactualClassification (en-ext) (reported): 64.810929544507
Added to our catalog: 2026-08-27

Base model and the most-downloaded derivatives in the catalog.

Compare with any feature-extraction model

Popular comparisons