🚀 开源AI新突破:Mixtral 8x22b模型来袭!
摘要
Mixtral 8x22b,一款开源的稀疏混合专家模型,以其高效性能和多语言支持,在AI领域引起广泛关注。

内容
🌟 探索AI的边界,开源模型Mixtral 8x22b以其卓越的性能和效率,成为AI社区的新宠!这款模型仅用39B活跃参数,从141B中脱颖而出,成本效率极高。
🌍 多语言支持:Mixtral 8x22b精通英、法、意、德、西语,数学和编程能力出众,原生支持函数调用,64K的上下文窗口让大文档信息精准回忆。
🔍 社区评价:在Hugging Face社区的推动下,Mixtral经过严格测试,成为可靠且高效的LLM。它在文本生成、问答、代码补全等任务上表现出色,用户反馈良好。
📈 发展趋势:新版本Mixtral-8x7B的发布,标志着模型向更大更强的方向发展。Hugging Face的开源精神,确保了这些进步是基于用户反馈和实际应用。
Run the model locally
Id
a3b6bef0f836
e8479ee1cb51
e8479ee1cb51
be69186cc79b
cf96f64cdf60
85c9d74f72e2
d552821493b8
50ec4fbf79f2
e8479ee1cb51
0ea2df035a44
de58127cf267
a9be4647cd05
c4d73325754c
a8e4c57d9fdb
040d78ec94bc
7dc2b27d2ce5
2a597519778d
74d06fe380a1
eb6b1cd64bb1
bada034b7898
9a187644e490
5637d20e8d2b
a45b937d5d88
4f5b5d113e8d
eb6b1cd64bb1
068cdfff8170
48fd198637f7
f43943467ac8
4f7e4cb4532b
77414be8fbac
92a2e32d7f87
9e736b483924
2fb8e840c060
619b5b9ddad2
96c204672c06
b5f8be13d719
44a28a0d7d9a
0f4e85e45e0c
8214ed645fc2
a3b6bef0f836
4ca263ed25e3
a69121c90013
39c18795563b
94bb1b270af9
1084fd5133d1
d011e1981bd1
68f12ddfe676
ceaac0dbeffa
4b4321b5a853
29c986d45fba
6c08977f2b56
279c2f61bb77
09d5395d1226
2ec7294409e7
b8bdfbd641af
29c986d45fba
20f1c9c7d6ee
2b19c246b20c
45ad2aba6509
1692f253599b
24ef911cb6dc
957a4d1fad36
0f550cd04214
0331d26f4ac1
50d4ad519912
a3b6bef0f836
29c986d45fba
e8479ee1cb51
e8479ee1cb51
Model
8x7b
8x22b
8x22b-instruct
8x22b-instruct-v0.1-fp16
8x22b-instruct-v0.1-q2_K
8x22b-instruct-v0.1-q3_K_L
8x22b-instruct-v0.1-q3_K_M
8x22b-instruct-v0.1-q3_K_S
8x22b-instruct-v0.1-q4_0
8x22b-instruct-v0.1-q4_1
8x22b-instruct-v0.1-q4_K_M
8x22b-instruct-v0.1-q4_K_S
8x22b-instruct-v0.1-q5_0
8x22b-instruct-v0.1-q5_1
8x22b-instruct-v0.1-q5_K_M
8x22b-instruct-v0.1-q5_K_S
8x22b-instruct-v0.1-q6_K
8x22b-instruct-v0.1-q8_0
8x22b-text
8x22b-text-v0.1-fp16
8x22b-text-v0.1-q2_K
8x22b-text-v0.1-q3_K_L
8x22b-text-v0.1-q3_K_M
8x22b-text-v0.1-q3_K_S
8x22b-text-v0.1-q4_0
8x22b-text-v0.1-q4_1
8x22b-text-v0.1-q4_K_M
8x22b-text-v0.1-q4_K_S
8x22b-text-v0.1-q5_0
8x22b-text-v0.1-q5_1
8x22b-text-v0.1-q5_K_M
8x22b-text-v0.1-q5_K_S
8x22b-text-v0.1-q6_K
8x22b-text-v0.1-q8_0
8x7b-instruct-v0.1-fp16
8x7b-instruct-v0.1-q2_K
8x7b-instruct-v0.1-q3_K_L
8x7b-instruct-v0.1-q3_K_M
8x7b-instruct-v0.1-q3_K_S
8x7b-instruct-v0.1-q4_0
8x7b-instruct-v0.1-q4_1
8x7b-instruct-v0.1-q4_K_M
8x7b-instruct-v0.1-q4_K_S
8x7b-instruct-v0.1-q5_0
8x7b-instruct-v0.1-q5_1
8x7b-instruct-v0.1-q5_K_M
8x7b-instruct-v0.1-q5_K_S
8x7b-instruct-v0.1-q6_K
8x7b-instruct-v0.1-q8_0
8x7b-text
8x7b-text-v0.1-fp16
8x7b-text-v0.1-q2_K
8x7b-text-v0.1-q3_K_L
8x7b-text-v0.1-q3_K_M
8x7b-text-v0.1-q3_K_S
8x7b-text-v0.1-q4_0
8x7b-text-v0.1-q4_1
8x7b-text-v0.1-q4_K_M
8x7b-text-v0.1-q4_K_S
8x7b-text-v0.1-q5_0
8x7b-text-v0.1-q5_1
8x7b-text-v0.1-q5_K_M
8x7b-text-v0.1-q5_K_S
8x7b-text-v0.1-q6_K
8x7b-text-v0.1-q8_0
instruct
text
v0.1
v0.1-instruct
Size
26GB
80GB
80GB
281GB
52GB
73GB
68GB
62GB
80GB
88GB
86GB
80GB
97GB
106GB
100GB
97GB
116GB
149GB
80GB
281GB
52GB
73GB
68GB
61GB
80GB
88GB
86GB
80GB
97GB
106GB
100GB
97GB
116GB
149GB
93GB
17GB
24GB
23GB
20GB
26GB
29GB
28GB
27GB
32GB
35GB
33GB
32GB
38GB
50GB
26GB
93GB
17GB
24GB
23GB
20GB
26GB
29GB
28GB
27GB
32GB
35GB
33GB
32GB
38GB
50GB
26GB
26GB
80GB
80GB