mixtral

A set of Mixture of Experts (MoE) model with open weights by Mistral AI in 8x7b and 8x22b parameter sizes.

Araç Kullanımı 8x7b 8x22b

Hızlı Kurulum (Ollama kuruluysa)

ollama run mixtral

Ollama kurulu değil mi? ollama.com/download — Windows, macOS ve Linux için ücretsiz. İlk çalıştırmada model indirilir, sonrası tamamen çevrimdışıdır.

Varyantlar

Boyut büyüdükçe kalite artar, donanım ihtiyacı yükselir. Başlangıç için küçük varyantı deneyin.

Etiket	Boyut	Bağlam	Girdi	Komut
latest	26GB	32K	Text	`ollama run mixtral:latest`
instruct	26GB	32K	Text	`ollama run mixtral:instruct`
text	26GB	32K	Text	`ollama run mixtral:text`
v0.1	80GB	64K	Text	`ollama run mixtral:v0.1`
8x7b	26GB	32K	Text	`ollama run mixtral:8x7b`
8x22b	80GB	64K	Text	`ollama run mixtral:8x22b`
8x7b-instruct-v0.1-q2_K	17GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q2_K`
8x7b-instruct-v0.1-q3_K_S	20GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q3_K_S`
8x7b-instruct-v0.1-q3_K_M	23GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q3_K_M`
8x7b-instruct-v0.1-q3_K_L	24GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q3_K_L`
8x7b-instruct-v0.1-q4_0	26GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q4_0`
8x7b-instruct-v0.1-q4_1	29GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q4_1`
8x7b-instruct-v0.1-q4_K_S	27GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q4_K_S`
8x7b-instruct-v0.1-q4_K_M	28GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q4_K_M`
8x7b-instruct-v0.1-q5_0	32GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q5_0`
8x7b-instruct-v0.1-q5_1	35GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q5_1`
8x7b-instruct-v0.1-q5_K_S	32GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q5_K_S`
8x7b-instruct-v0.1-q5_K_M	33GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q5_K_M`
8x7b-instruct-v0.1-q6_K	38GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q6_K`
8x7b-instruct-v0.1-q8_0	50GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-q8_0`
8x7b-instruct-v0.1-fp16	93GB	32K	Text	`ollama run mixtral:8x7b-instruct-v0.1-fp16`
8x7b-text	26GB	32K	Text	`ollama run mixtral:8x7b-text`
8x7b-text-v0.1-q2_K	17GB	32K	Text	`ollama run mixtral:8x7b-text-v0.1-q2_K`
8x7b-text-v0.1-q3_K_S	20GB	32K	Text	`ollama run mixtral:8x7b-text-v0.1-q3_K_S`
8x7b-text-v0.1-q3_K_M	23GB	32K	Text	`ollama run mixtral:8x7b-text-v0.1-q3_K_M`

45 varyant daha var — tümü için ollama.com.

Model Detayları ve Benchmarklar (kaynak: ollama.com)

The Mixtral large Language Models (LLM) are a set of pretrained generative Sparse Mixture of Experts.

Sizes

mixtral:8x22b
mixtral:8x7b

Mixtral 8x22b

ollama run mixtral:8x22b

Mixtral 8x22B sets a new standard for performance and efficiency within the AI community. It is a sparse Mixture-of-Experts (SMoE) model that uses only 39B active parameters out of 141B, offering unparalleled cost efficiency for its size.

Mixtral 8x22B comes with the following strengths:

It is fluent in English, French, Italian, German, and Spanish
It has strong maths and coding capabilities
It is natively capable of function calling
64K tokens context window allows precise information recall from large documents

References

Announcement

HuggingFace