vince_ai/mmmistral8/adapter

vince_ai/

mmmistral8:latest

1 Download Updated 11 months ago

tools

mmmistral8:latest ... /

adapter

cbce60940941 · 27MB

Metadata

general.architecture

llama

llama
general.file_type

F16

F16
llama.attention.head_count

32

32
llama.attention.head_count_kv

8

8
adapter.lora.alpha

64

64
adapter.type

lora

lora

Tensor

Name

Type

Shape

blk.16

blk.16.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.16.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.16.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.16.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.17

blk.17.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.17.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.17.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.17.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.18

blk.18.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.18.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.18.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.18.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.19

blk.19.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.19.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.19.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.19.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.20

blk.20.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.20.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.20.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.20.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.21

blk.21.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.21.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.21.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.21.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.22

blk.22.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.22.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.22.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.22.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.23

blk.23.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.23.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.23.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.23.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.24

blk.24.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.24.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.24.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.24.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.25

blk.25.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.25.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.25.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.25.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.26

blk.26.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.26.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.26.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.26.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.27

blk.27.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.27.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.27.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.27.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.28

blk.28.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.28.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.28.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.28.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.29

blk.29.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.29.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.29.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.29.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.30

blk.30.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.30.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.30.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.30.attn_v.weight.lora_b

F16

F16

[64, 1024]

blk.31

blk.31.attn_q.weight.lora_a

F16

F16

[4096, 64]
blk.31.attn_q.weight.lora_b

F16

F16

[64, 4096]
blk.31.attn_v.weight.lora_a

F16

F16

[4096, 64]
blk.31.attn_v.weight.lora_b

F16

F16

[64, 1024]