GitHub
183.3K
Download
Models
Docs
Pricing
Sign in
Download
Models
Docs
Pricing
Sign in
oza6ut0ne
/
d1-omni-600M
:latest
2
Downloads
Updated
yesterday
Cancel
vision
d1-omni-600M:latest
...
/
projector
df887978550c ยท 263MB
Metadata
general.architecture
clip
clip
general.file_type
Q8_0
Q8_0
clip.audio.attention.head_count
8
8
clip.audio.attention.layer_norm_epsilon
1e-05
1e-05
clip.audio.block_count
17
17
clip.audio.embedding_length
512
512
clip.audio.feed_forward_length
2048
2048
clip.audio.num_mel_bins
128
128
clip.audio.projection_dim
1024
1024
clip.audio.projector_type
d1omni_a
d1omni_a
clip.has_audio_encoder
true
true
clip.has_vision_encoder
true
true
clip.use_gelu
true
true
clip.vision.attention.head_count
12
12
clip.vision.attention.layer_norm_epsilon
1e-06
1e-06
clip.vision.block_count
12
12
clip.vision.embedding_length
768
768
clip.vision.feed_forward_length
3072
3072
clip.vision.image_mean
[0.5, 0.5, 0.5]
[0.5, 0.5, 0.5]
clip.vision.image_size
256
256
clip.vision.image_std
[0.5, 0.5, 0.5]
[0.5, 0.5, 0.5]
clip.vision.patch_size
16
16
clip.vision.projection_dim
1024
1024
clip.vision.projector.scale_factor
2
2
clip.vision.projector_type
d1omni_v
d1omni_v
Tensor
Name
Type
Shape
a.blk.0
a.blk.0.attn_k.bias
F32
F32
[512]
a.blk.0.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.0.attn_out.bias
F32
F32
[512]
a.blk.0.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.0.attn_q.bias
F32
F32
[512]
a.blk.0.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.0.attn_v.bias
F32
F32
[512]
a.blk.0.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.0.conv_dw.bias
F32
F32
[512]
a.blk.0.conv_dw.weight
F32
F32
[9, 512]
a.blk.0.conv_norm.bias
F32
F32
[512]
a.blk.0.conv_norm.weight
F32
F32
[512]
a.blk.0.conv_pw1.bias
F32
F32
[1024]
a.blk.0.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.0.conv_pw2.bias
F32
F32
[512]
a.blk.0.conv_pw2.weight
F32
F32
[512, 512]
a.blk.0.ffn_down.bias
F32
F32
[512]
a.blk.0.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.0.ffn_down_1.bias
F32
F32
[512]
a.blk.0.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.0.ffn_norm.bias
F32
F32
[512]
a.blk.0.ffn_norm.weight
F32
F32
[512]
a.blk.0.ffn_norm_1.bias
F32
F32
[512]
a.blk.0.ffn_norm_1.weight
F32
F32
[512]
a.blk.0.ffn_up.bias
F32
F32
[2048]
a.blk.0.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.0.ffn_up_1.bias
F32
F32
[2048]
a.blk.0.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.0.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.0.ln1.bias
F32
F32
[512]
a.blk.0.ln1.weight
F32
F32
[512]
a.blk.0.ln2.bias
F32
F32
[512]
a.blk.0.ln2.weight
F32
F32
[512]
a.blk.0.norm_conv.bias
F32
F32
[512]
a.blk.0.norm_conv.weight
F32
F32
[512]
a.blk.0.pos_bias_u
F32
F32
[64, 8]
a.blk.0.pos_bias_v
F32
F32
[64, 8]
a.blk.1
a.blk.1.attn_k.bias
F32
F32
[512]
a.blk.1.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.1.attn_out.bias
F32
F32
[512]
a.blk.1.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.1.attn_q.bias
F32
F32
[512]
a.blk.1.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.1.attn_v.bias
F32
F32
[512]
a.blk.1.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.1.conv_dw.bias
F32
F32
[512]
a.blk.1.conv_dw.weight
F32
F32
[9, 512]
a.blk.1.conv_norm.bias
F32
F32
[512]
a.blk.1.conv_norm.weight
F32
F32
[512]
a.blk.1.conv_pw1.bias
F32
F32
[1024]
a.blk.1.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.1.conv_pw2.bias
F32
F32
[512]
a.blk.1.conv_pw2.weight
F32
F32
[512, 512]
a.blk.1.ffn_down.bias
F32
F32
[512]
a.blk.1.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.1.ffn_down_1.bias
F32
F32
[512]
a.blk.1.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.1.ffn_norm.bias
F32
F32
[512]
a.blk.1.ffn_norm.weight
F32
F32
[512]
a.blk.1.ffn_norm_1.bias
F32
F32
[512]
a.blk.1.ffn_norm_1.weight
F32
F32
[512]
a.blk.1.ffn_up.bias
F32
F32
[2048]
a.blk.1.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.1.ffn_up_1.bias
F32
F32
[2048]
a.blk.1.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.1.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.1.ln1.bias
F32
F32
[512]
a.blk.1.ln1.weight
F32
F32
[512]
a.blk.1.ln2.bias
F32
F32
[512]
a.blk.1.ln2.weight
F32
F32
[512]
a.blk.1.norm_conv.bias
F32
F32
[512]
a.blk.1.norm_conv.weight
F32
F32
[512]
a.blk.1.pos_bias_u
F32
F32
[64, 8]
a.blk.1.pos_bias_v
F32
F32
[64, 8]
a.blk.2
a.blk.2.attn_k.bias
F32
F32
[512]
a.blk.2.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.2.attn_out.bias
F32
F32
[512]
a.blk.2.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.2.attn_q.bias
F32
F32
[512]
a.blk.2.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.2.attn_v.bias
F32
F32
[512]
a.blk.2.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.2.conv_dw.bias
F32
F32
[512]
a.blk.2.conv_dw.weight
F32
F32
[9, 512]
a.blk.2.conv_norm.bias
F32
F32
[512]
a.blk.2.conv_norm.weight
F32
F32
[512]
a.blk.2.conv_pw1.bias
F32
F32
[1024]
a.blk.2.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.2.conv_pw2.bias
F32
F32
[512]
a.blk.2.conv_pw2.weight
F32
F32
[512, 512]
a.blk.2.ffn_down.bias
F32
F32
[512]
a.blk.2.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.2.ffn_down_1.bias
F32
F32
[512]
a.blk.2.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.2.ffn_norm.bias
F32
F32
[512]
a.blk.2.ffn_norm.weight
F32
F32
[512]
a.blk.2.ffn_norm_1.bias
F32
F32
[512]
a.blk.2.ffn_norm_1.weight
F32
F32
[512]
a.blk.2.ffn_up.bias
F32
F32
[2048]
a.blk.2.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.2.ffn_up_1.bias
F32
F32
[2048]
a.blk.2.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.2.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.2.ln1.bias
F32
F32
[512]
a.blk.2.ln1.weight
F32
F32
[512]
a.blk.2.ln2.bias
F32
F32
[512]
a.blk.2.ln2.weight
F32
F32
[512]
a.blk.2.norm_conv.bias
F32
F32
[512]
a.blk.2.norm_conv.weight
F32
F32
[512]
a.blk.2.pos_bias_u
F32
F32
[64, 8]
a.blk.2.pos_bias_v
F32
F32
[64, 8]
a.blk.3
a.blk.3.attn_k.bias
F32
F32
[512]
a.blk.3.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.3.attn_out.bias
F32
F32
[512]
a.blk.3.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.3.attn_q.bias
F32
F32
[512]
a.blk.3.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.3.attn_v.bias
F32
F32
[512]
a.blk.3.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.3.conv_dw.bias
F32
F32
[512]
a.blk.3.conv_dw.weight
F32
F32
[9, 512]
a.blk.3.conv_norm.bias
F32
F32
[512]
a.blk.3.conv_norm.weight
F32
F32
[512]
a.blk.3.conv_pw1.bias
F32
F32
[1024]
a.blk.3.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.3.conv_pw2.bias
F32
F32
[512]
a.blk.3.conv_pw2.weight
F32
F32
[512, 512]
a.blk.3.ffn_down.bias
F32
F32
[512]
a.blk.3.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.3.ffn_down_1.bias
F32
F32
[512]
a.blk.3.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.3.ffn_norm.bias
F32
F32
[512]
a.blk.3.ffn_norm.weight
F32
F32
[512]
a.blk.3.ffn_norm_1.bias
F32
F32
[512]
a.blk.3.ffn_norm_1.weight
F32
F32
[512]
a.blk.3.ffn_up.bias
F32
F32
[2048]
a.blk.3.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.3.ffn_up_1.bias
F32
F32
[2048]
a.blk.3.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.3.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.3.ln1.bias
F32
F32
[512]
a.blk.3.ln1.weight
F32
F32
[512]
a.blk.3.ln2.bias
F32
F32
[512]
a.blk.3.ln2.weight
F32
F32
[512]
a.blk.3.norm_conv.bias
F32
F32
[512]
a.blk.3.norm_conv.weight
F32
F32
[512]
a.blk.3.pos_bias_u
F32
F32
[64, 8]
a.blk.3.pos_bias_v
F32
F32
[64, 8]
a.blk.4
a.blk.4.attn_k.bias
F32
F32
[512]
a.blk.4.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.4.attn_out.bias
F32
F32
[512]
a.blk.4.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.4.attn_q.bias
F32
F32
[512]
a.blk.4.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.4.attn_v.bias
F32
F32
[512]
a.blk.4.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.4.conv_dw.bias
F32
F32
[512]
a.blk.4.conv_dw.weight
F32
F32
[9, 512]
a.blk.4.conv_norm.bias
F32
F32
[512]
a.blk.4.conv_norm.weight
F32
F32
[512]
a.blk.4.conv_pw1.bias
F32
F32
[1024]
a.blk.4.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.4.conv_pw2.bias
F32
F32
[512]
a.blk.4.conv_pw2.weight
F32
F32
[512, 512]
a.blk.4.ffn_down.bias
F32
F32
[512]
a.blk.4.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.4.ffn_down_1.bias
F32
F32
[512]
a.blk.4.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.4.ffn_norm.bias
F32
F32
[512]
a.blk.4.ffn_norm.weight
F32
F32
[512]
a.blk.4.ffn_norm_1.bias
F32
F32
[512]
a.blk.4.ffn_norm_1.weight
F32
F32
[512]
a.blk.4.ffn_up.bias
F32
F32
[2048]
a.blk.4.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.4.ffn_up_1.bias
F32
F32
[2048]
a.blk.4.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.4.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.4.ln1.bias
F32
F32
[512]
a.blk.4.ln1.weight
F32
F32
[512]
a.blk.4.ln2.bias
F32
F32
[512]
a.blk.4.ln2.weight
F32
F32
[512]
a.blk.4.norm_conv.bias
F32
F32
[512]
a.blk.4.norm_conv.weight
F32
F32
[512]
a.blk.4.pos_bias_u
F32
F32
[64, 8]
a.blk.4.pos_bias_v
F32
F32
[64, 8]
a.blk.5
a.blk.5.attn_k.bias
F32
F32
[512]
a.blk.5.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.5.attn_out.bias
F32
F32
[512]
a.blk.5.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.5.attn_q.bias
F32
F32
[512]
a.blk.5.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.5.attn_v.bias
F32
F32
[512]
a.blk.5.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.5.conv_dw.bias
F32
F32
[512]
a.blk.5.conv_dw.weight
F32
F32
[9, 512]
a.blk.5.conv_norm.bias
F32
F32
[512]
a.blk.5.conv_norm.weight
F32
F32
[512]
a.blk.5.conv_pw1.bias
F32
F32
[1024]
a.blk.5.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.5.conv_pw2.bias
F32
F32
[512]
a.blk.5.conv_pw2.weight
F32
F32
[512, 512]
a.blk.5.ffn_down.bias
F32
F32
[512]
a.blk.5.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.5.ffn_down_1.bias
F32
F32
[512]
a.blk.5.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.5.ffn_norm.bias
F32
F32
[512]
a.blk.5.ffn_norm.weight
F32
F32
[512]
a.blk.5.ffn_norm_1.bias
F32
F32
[512]
a.blk.5.ffn_norm_1.weight
F32
F32
[512]
a.blk.5.ffn_up.bias
F32
F32
[2048]
a.blk.5.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.5.ffn_up_1.bias
F32
F32
[2048]
a.blk.5.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.5.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.5.ln1.bias
F32
F32
[512]
a.blk.5.ln1.weight
F32
F32
[512]
a.blk.5.ln2.bias
F32
F32
[512]
a.blk.5.ln2.weight
F32
F32
[512]
a.blk.5.norm_conv.bias
F32
F32
[512]
a.blk.5.norm_conv.weight
F32
F32
[512]
a.blk.5.pos_bias_u
F32
F32
[64, 8]
a.blk.5.pos_bias_v
F32
F32
[64, 8]
a.blk.6
a.blk.6.attn_k.bias
F32
F32
[512]
a.blk.6.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.6.attn_out.bias
F32
F32
[512]
a.blk.6.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.6.attn_q.bias
F32
F32
[512]
a.blk.6.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.6.attn_v.bias
F32
F32
[512]
a.blk.6.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.6.conv_dw.bias
F32
F32
[512]
a.blk.6.conv_dw.weight
F32
F32
[9, 512]
a.blk.6.conv_norm.bias
F32
F32
[512]
a.blk.6.conv_norm.weight
F32
F32
[512]
a.blk.6.conv_pw1.bias
F32
F32
[1024]
a.blk.6.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.6.conv_pw2.bias
F32
F32
[512]
a.blk.6.conv_pw2.weight
F32
F32
[512, 512]
a.blk.6.ffn_down.bias
F32
F32
[512]
a.blk.6.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.6.ffn_down_1.bias
F32
F32
[512]
a.blk.6.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.6.ffn_norm.bias
F32
F32
[512]
a.blk.6.ffn_norm.weight
F32
F32
[512]
a.blk.6.ffn_norm_1.bias
F32
F32
[512]
a.blk.6.ffn_norm_1.weight
F32
F32
[512]
a.blk.6.ffn_up.bias
F32
F32
[2048]
a.blk.6.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.6.ffn_up_1.bias
F32
F32
[2048]
a.blk.6.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.6.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.6.ln1.bias
F32
F32
[512]
a.blk.6.ln1.weight
F32
F32
[512]
a.blk.6.ln2.bias
F32
F32
[512]
a.blk.6.ln2.weight
F32
F32
[512]
a.blk.6.norm_conv.bias
F32
F32
[512]
a.blk.6.norm_conv.weight
F32
F32
[512]
a.blk.6.pos_bias_u
F32
F32
[64, 8]
a.blk.6.pos_bias_v
F32
F32
[64, 8]
a.blk.7
a.blk.7.attn_k.bias
F32
F32
[512]
a.blk.7.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.7.attn_out.bias
F32
F32
[512]
a.blk.7.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.7.attn_q.bias
F32
F32
[512]
a.blk.7.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.7.attn_v.bias
F32
F32
[512]
a.blk.7.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.7.conv_dw.bias
F32
F32
[512]
a.blk.7.conv_dw.weight
F32
F32
[9, 512]
a.blk.7.conv_norm.bias
F32
F32
[512]
a.blk.7.conv_norm.weight
F32
F32
[512]
a.blk.7.conv_pw1.bias
F32
F32
[1024]
a.blk.7.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.7.conv_pw2.bias
F32
F32
[512]
a.blk.7.conv_pw2.weight
F32
F32
[512, 512]
a.blk.7.ffn_down.bias
F32
F32
[512]
a.blk.7.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.7.ffn_down_1.bias
F32
F32
[512]
a.blk.7.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.7.ffn_norm.bias
F32
F32
[512]
a.blk.7.ffn_norm.weight
F32
F32
[512]
a.blk.7.ffn_norm_1.bias
F32
F32
[512]
a.blk.7.ffn_norm_1.weight
F32
F32
[512]
a.blk.7.ffn_up.bias
F32
F32
[2048]
a.blk.7.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.7.ffn_up_1.bias
F32
F32
[2048]
a.blk.7.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.7.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.7.ln1.bias
F32
F32
[512]
a.blk.7.ln1.weight
F32
F32
[512]
a.blk.7.ln2.bias
F32
F32
[512]
a.blk.7.ln2.weight
F32
F32
[512]
a.blk.7.norm_conv.bias
F32
F32
[512]
a.blk.7.norm_conv.weight
F32
F32
[512]
a.blk.7.pos_bias_u
F32
F32
[64, 8]
a.blk.7.pos_bias_v
F32
F32
[64, 8]
a.blk.8
a.blk.8.attn_k.bias
F32
F32
[512]
a.blk.8.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.8.attn_out.bias
F32
F32
[512]
a.blk.8.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.8.attn_q.bias
F32
F32
[512]
a.blk.8.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.8.attn_v.bias
F32
F32
[512]
a.blk.8.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.8.conv_dw.bias
F32
F32
[512]
a.blk.8.conv_dw.weight
F32
F32
[9, 512]
a.blk.8.conv_norm.bias
F32
F32
[512]
a.blk.8.conv_norm.weight
F32
F32
[512]
a.blk.8.conv_pw1.bias
F32
F32
[1024]
a.blk.8.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.8.conv_pw2.bias
F32
F32
[512]
a.blk.8.conv_pw2.weight
F32
F32
[512, 512]
a.blk.8.ffn_down.bias
F32
F32
[512]
a.blk.8.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.8.ffn_down_1.bias
F32
F32
[512]
a.blk.8.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.8.ffn_norm.bias
F32
F32
[512]
a.blk.8.ffn_norm.weight
F32
F32
[512]
a.blk.8.ffn_norm_1.bias
F32
F32
[512]
a.blk.8.ffn_norm_1.weight
F32
F32
[512]
a.blk.8.ffn_up.bias
F32
F32
[2048]
a.blk.8.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.8.ffn_up_1.bias
F32
F32
[2048]
a.blk.8.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.8.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.8.ln1.bias
F32
F32
[512]
a.blk.8.ln1.weight
F32
F32
[512]
a.blk.8.ln2.bias
F32
F32
[512]
a.blk.8.ln2.weight
F32
F32
[512]
a.blk.8.norm_conv.bias
F32
F32
[512]
a.blk.8.norm_conv.weight
F32
F32
[512]
a.blk.8.pos_bias_u
F32
F32
[64, 8]
a.blk.8.pos_bias_v
F32
F32
[64, 8]
a.blk.9
a.blk.9.attn_k.bias
F32
F32
[512]
a.blk.9.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.9.attn_out.bias
F32
F32
[512]
a.blk.9.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.9.attn_q.bias
F32
F32
[512]
a.blk.9.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.9.attn_v.bias
F32
F32
[512]
a.blk.9.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.9.conv_dw.bias
F32
F32
[512]
a.blk.9.conv_dw.weight
F32
F32
[9, 512]
a.blk.9.conv_norm.bias
F32
F32
[512]
a.blk.9.conv_norm.weight
F32
F32
[512]
a.blk.9.conv_pw1.bias
F32
F32
[1024]
a.blk.9.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.9.conv_pw2.bias
F32
F32
[512]
a.blk.9.conv_pw2.weight
F32
F32
[512, 512]
a.blk.9.ffn_down.bias
F32
F32
[512]
a.blk.9.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.9.ffn_down_1.bias
F32
F32
[512]
a.blk.9.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.9.ffn_norm.bias
F32
F32
[512]
a.blk.9.ffn_norm.weight
F32
F32
[512]
a.blk.9.ffn_norm_1.bias
F32
F32
[512]
a.blk.9.ffn_norm_1.weight
F32
F32
[512]
a.blk.9.ffn_up.bias
F32
F32
[2048]
a.blk.9.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.9.ffn_up_1.bias
F32
F32
[2048]
a.blk.9.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.9.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.9.ln1.bias
F32
F32
[512]
a.blk.9.ln1.weight
F32
F32
[512]
a.blk.9.ln2.bias
F32
F32
[512]
a.blk.9.ln2.weight
F32
F32
[512]
a.blk.9.norm_conv.bias
F32
F32
[512]
a.blk.9.norm_conv.weight
F32
F32
[512]
a.blk.9.pos_bias_u
F32
F32
[64, 8]
a.blk.9.pos_bias_v
F32
F32
[64, 8]
a.blk.10
a.blk.10.attn_k.bias
F32
F32
[512]
a.blk.10.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.10.attn_out.bias
F32
F32
[512]
a.blk.10.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.10.attn_q.bias
F32
F32
[512]
a.blk.10.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.10.attn_v.bias
F32
F32
[512]
a.blk.10.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.10.conv_dw.bias
F32
F32
[512]
a.blk.10.conv_dw.weight
F32
F32
[9, 512]
a.blk.10.conv_norm.bias
F32
F32
[512]
a.blk.10.conv_norm.weight
F32
F32
[512]
a.blk.10.conv_pw1.bias
F32
F32
[1024]
a.blk.10.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.10.conv_pw2.bias
F32
F32
[512]
a.blk.10.conv_pw2.weight
F32
F32
[512, 512]
a.blk.10.ffn_down.bias
F32
F32
[512]
a.blk.10.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.10.ffn_down_1.bias
F32
F32
[512]
a.blk.10.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.10.ffn_norm.bias
F32
F32
[512]
a.blk.10.ffn_norm.weight
F32
F32
[512]
a.blk.10.ffn_norm_1.bias
F32
F32
[512]
a.blk.10.ffn_norm_1.weight
F32
F32
[512]
a.blk.10.ffn_up.bias
F32
F32
[2048]
a.blk.10.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.10.ffn_up_1.bias
F32
F32
[2048]
a.blk.10.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.10.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.10.ln1.bias
F32
F32
[512]
a.blk.10.ln1.weight
F32
F32
[512]
a.blk.10.ln2.bias
F32
F32
[512]
a.blk.10.ln2.weight
F32
F32
[512]
a.blk.10.norm_conv.bias
F32
F32
[512]
a.blk.10.norm_conv.weight
F32
F32
[512]
a.blk.10.pos_bias_u
F32
F32
[64, 8]
a.blk.10.pos_bias_v
F32
F32
[64, 8]
a.blk.11
a.blk.11.attn_k.bias
F32
F32
[512]
a.blk.11.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.11.attn_out.bias
F32
F32
[512]
a.blk.11.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.11.attn_q.bias
F32
F32
[512]
a.blk.11.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.11.attn_v.bias
F32
F32
[512]
a.blk.11.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.11.conv_dw.bias
F32
F32
[512]
a.blk.11.conv_dw.weight
F32
F32
[9, 512]
a.blk.11.conv_norm.bias
F32
F32
[512]
a.blk.11.conv_norm.weight
F32
F32
[512]
a.blk.11.conv_pw1.bias
F32
F32
[1024]
a.blk.11.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.11.conv_pw2.bias
F32
F32
[512]
a.blk.11.conv_pw2.weight
F32
F32
[512, 512]
a.blk.11.ffn_down.bias
F32
F32
[512]
a.blk.11.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.11.ffn_down_1.bias
F32
F32
[512]
a.blk.11.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.11.ffn_norm.bias
F32
F32
[512]
a.blk.11.ffn_norm.weight
F32
F32
[512]
a.blk.11.ffn_norm_1.bias
F32
F32
[512]
a.blk.11.ffn_norm_1.weight
F32
F32
[512]
a.blk.11.ffn_up.bias
F32
F32
[2048]
a.blk.11.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.11.ffn_up_1.bias
F32
F32
[2048]
a.blk.11.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.11.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.11.ln1.bias
F32
F32
[512]
a.blk.11.ln1.weight
F32
F32
[512]
a.blk.11.ln2.bias
F32
F32
[512]
a.blk.11.ln2.weight
F32
F32
[512]
a.blk.11.norm_conv.bias
F32
F32
[512]
a.blk.11.norm_conv.weight
F32
F32
[512]
a.blk.11.pos_bias_u
F32
F32
[64, 8]
a.blk.11.pos_bias_v
F32
F32
[64, 8]
a.blk.12
a.blk.12.attn_k.bias
F32
F32
[512]
a.blk.12.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.12.attn_out.bias
F32
F32
[512]
a.blk.12.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.12.attn_q.bias
F32
F32
[512]
a.blk.12.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.12.attn_v.bias
F32
F32
[512]
a.blk.12.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.12.conv_dw.bias
F32
F32
[512]
a.blk.12.conv_dw.weight
F32
F32
[9, 512]
a.blk.12.conv_norm.bias
F32
F32
[512]
a.blk.12.conv_norm.weight
F32
F32
[512]
a.blk.12.conv_pw1.bias
F32
F32
[1024]
a.blk.12.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.12.conv_pw2.bias
F32
F32
[512]
a.blk.12.conv_pw2.weight
F32
F32
[512, 512]
a.blk.12.ffn_down.bias
F32
F32
[512]
a.blk.12.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.12.ffn_down_1.bias
F32
F32
[512]
a.blk.12.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.12.ffn_norm.bias
F32
F32
[512]
a.blk.12.ffn_norm.weight
F32
F32
[512]
a.blk.12.ffn_norm_1.bias
F32
F32
[512]
a.blk.12.ffn_norm_1.weight
F32
F32
[512]
a.blk.12.ffn_up.bias
F32
F32
[2048]
a.blk.12.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.12.ffn_up_1.bias
F32
F32
[2048]
a.blk.12.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.12.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.12.ln1.bias
F32
F32
[512]
a.blk.12.ln1.weight
F32
F32
[512]
a.blk.12.ln2.bias
F32
F32
[512]
a.blk.12.ln2.weight
F32
F32
[512]
a.blk.12.norm_conv.bias
F32
F32
[512]
a.blk.12.norm_conv.weight
F32
F32
[512]
a.blk.12.pos_bias_u
F32
F32
[64, 8]
a.blk.12.pos_bias_v
F32
F32
[64, 8]
a.blk.13
a.blk.13.attn_k.bias
F32
F32
[512]
a.blk.13.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.13.attn_out.bias
F32
F32
[512]
a.blk.13.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.13.attn_q.bias
F32
F32
[512]
a.blk.13.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.13.attn_v.bias
F32
F32
[512]
a.blk.13.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.13.conv_dw.bias
F32
F32
[512]
a.blk.13.conv_dw.weight
F32
F32
[9, 512]
a.blk.13.conv_norm.bias
F32
F32
[512]
a.blk.13.conv_norm.weight
F32
F32
[512]
a.blk.13.conv_pw1.bias
F32
F32
[1024]
a.blk.13.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.13.conv_pw2.bias
F32
F32
[512]
a.blk.13.conv_pw2.weight
F32
F32
[512, 512]
a.blk.13.ffn_down.bias
F32
F32
[512]
a.blk.13.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.13.ffn_down_1.bias
F32
F32
[512]
a.blk.13.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.13.ffn_norm.bias
F32
F32
[512]
a.blk.13.ffn_norm.weight
F32
F32
[512]
a.blk.13.ffn_norm_1.bias
F32
F32
[512]
a.blk.13.ffn_norm_1.weight
F32
F32
[512]
a.blk.13.ffn_up.bias
F32
F32
[2048]
a.blk.13.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.13.ffn_up_1.bias
F32
F32
[2048]
a.blk.13.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.13.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.13.ln1.bias
F32
F32
[512]
a.blk.13.ln1.weight
F32
F32
[512]
a.blk.13.ln2.bias
F32
F32
[512]
a.blk.13.ln2.weight
F32
F32
[512]
a.blk.13.norm_conv.bias
F32
F32
[512]
a.blk.13.norm_conv.weight
F32
F32
[512]
a.blk.13.pos_bias_u
F32
F32
[64, 8]
a.blk.13.pos_bias_v
F32
F32
[64, 8]
a.blk.14
a.blk.14.attn_k.bias
F32
F32
[512]
a.blk.14.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.14.attn_out.bias
F32
F32
[512]
a.blk.14.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.14.attn_q.bias
F32
F32
[512]
a.blk.14.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.14.attn_v.bias
F32
F32
[512]
a.blk.14.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.14.conv_dw.bias
F32
F32
[512]
a.blk.14.conv_dw.weight
F32
F32
[9, 512]
a.blk.14.conv_norm.bias
F32
F32
[512]
a.blk.14.conv_norm.weight
F32
F32
[512]
a.blk.14.conv_pw1.bias
F32
F32
[1024]
a.blk.14.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.14.conv_pw2.bias
F32
F32
[512]
a.blk.14.conv_pw2.weight
F32
F32
[512, 512]
a.blk.14.ffn_down.bias
F32
F32
[512]
a.blk.14.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.14.ffn_down_1.bias
F32
F32
[512]
a.blk.14.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.14.ffn_norm.bias
F32
F32
[512]
a.blk.14.ffn_norm.weight
F32
F32
[512]
a.blk.14.ffn_norm_1.bias
F32
F32
[512]
a.blk.14.ffn_norm_1.weight
F32
F32
[512]
a.blk.14.ffn_up.bias
F32
F32
[2048]
a.blk.14.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.14.ffn_up_1.bias
F32
F32
[2048]
a.blk.14.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.14.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.14.ln1.bias
F32
F32
[512]
a.blk.14.ln1.weight
F32
F32
[512]
a.blk.14.ln2.bias
F32
F32
[512]
a.blk.14.ln2.weight
F32
F32
[512]
a.blk.14.norm_conv.bias
F32
F32
[512]
a.blk.14.norm_conv.weight
F32
F32
[512]
a.blk.14.pos_bias_u
F32
F32
[64, 8]
a.blk.14.pos_bias_v
F32
F32
[64, 8]
a.blk.15
a.blk.15.attn_k.bias
F32
F32
[512]
a.blk.15.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.15.attn_out.bias
F32
F32
[512]
a.blk.15.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.15.attn_q.bias
F32
F32
[512]
a.blk.15.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.15.attn_v.bias
F32
F32
[512]
a.blk.15.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.15.conv_dw.bias
F32
F32
[512]
a.blk.15.conv_dw.weight
F32
F32
[9, 512]
a.blk.15.conv_norm.bias
F32
F32
[512]
a.blk.15.conv_norm.weight
F32
F32
[512]
a.blk.15.conv_pw1.bias
F32
F32
[1024]
a.blk.15.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.15.conv_pw2.bias
F32
F32
[512]
a.blk.15.conv_pw2.weight
F32
F32
[512, 512]
a.blk.15.ffn_down.bias
F32
F32
[512]
a.blk.15.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.15.ffn_down_1.bias
F32
F32
[512]
a.blk.15.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.15.ffn_norm.bias
F32
F32
[512]
a.blk.15.ffn_norm.weight
F32
F32
[512]
a.blk.15.ffn_norm_1.bias
F32
F32
[512]
a.blk.15.ffn_norm_1.weight
F32
F32
[512]
a.blk.15.ffn_up.bias
F32
F32
[2048]
a.blk.15.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.15.ffn_up_1.bias
F32
F32
[2048]
a.blk.15.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.15.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.15.ln1.bias
F32
F32
[512]
a.blk.15.ln1.weight
F32
F32
[512]
a.blk.15.ln2.bias
F32
F32
[512]
a.blk.15.ln2.weight
F32
F32
[512]
a.blk.15.norm_conv.bias
F32
F32
[512]
a.blk.15.norm_conv.weight
F32
F32
[512]
a.blk.15.pos_bias_u
F32
F32
[64, 8]
a.blk.15.pos_bias_v
F32
F32
[64, 8]
a.blk.16
a.blk.16.attn_k.bias
F32
F32
[512]
a.blk.16.attn_k.weight
Q8_0
Q8_0
[512, 512]
a.blk.16.attn_out.bias
F32
F32
[512]
a.blk.16.attn_out.weight
Q8_0
Q8_0
[512, 512]
a.blk.16.attn_q.bias
F32
F32
[512]
a.blk.16.attn_q.weight
Q8_0
Q8_0
[512, 512]
a.blk.16.attn_v.bias
F32
F32
[512]
a.blk.16.attn_v.weight
Q8_0
Q8_0
[512, 512]
a.blk.16.conv_dw.bias
F32
F32
[512]
a.blk.16.conv_dw.weight
F32
F32
[9, 512]
a.blk.16.conv_norm.bias
F32
F32
[512]
a.blk.16.conv_norm.weight
F32
F32
[512]
a.blk.16.conv_pw1.bias
F32
F32
[1024]
a.blk.16.conv_pw1.weight
F32
F32
[512, 1024]
a.blk.16.conv_pw2.bias
F32
F32
[512]
a.blk.16.conv_pw2.weight
F32
F32
[512, 512]
a.blk.16.ffn_down.bias
F32
F32
[512]
a.blk.16.ffn_down.weight
Q8_0
Q8_0
[2048, 512]
a.blk.16.ffn_down_1.bias
F32
F32
[512]
a.blk.16.ffn_down_1.weight
Q8_0
Q8_0
[2048, 512]
a.blk.16.ffn_norm.bias
F32
F32
[512]
a.blk.16.ffn_norm.weight
F32
F32
[512]
a.blk.16.ffn_norm_1.bias
F32
F32
[512]
a.blk.16.ffn_norm_1.weight
F32
F32
[512]
a.blk.16.ffn_up.bias
F32
F32
[2048]
a.blk.16.ffn_up.weight
Q8_0
Q8_0
[512, 2048]
a.blk.16.ffn_up_1.bias
F32
F32
[2048]
a.blk.16.ffn_up_1.weight
Q8_0
Q8_0
[512, 2048]
a.blk.16.linear_pos.weight
Q8_0
Q8_0
[512, 512]
a.blk.16.ln1.bias
F32
F32
[512]
a.blk.16.ln1.weight
F32
F32
[512]
a.blk.16.ln2.bias
F32
F32
[512]
a.blk.16.ln2.weight
F32
F32
[512]
a.blk.16.norm_conv.bias
F32
F32
[512]
a.blk.16.norm_conv.weight
F32
F32
[512]
a.blk.16.pos_bias_u
F32
F32
[64, 8]
a.blk.16.pos_bias_v
F32
F32
[64, 8]
a.conv1d.0.bias
F32
F32
[1, 1, 256]
a.conv1d.0.weight
F32
F32
[3, 3, 1, 256]
a.conv1d.2.bias
F32
F32
[1, 1, 256]
a.conv1d.2.weight
F32
F32
[3, 3, 1, 256]
a.conv1d.3.bias
F32
F32
[1, 1, 256]
a.conv1d.3.weight
F32
F32
[1, 1, 256, 256]
a.conv1d.5.bias
F32
F32
[1, 1, 256]
a.conv1d.5.weight
F32
F32
[3, 3, 1, 256]
a.conv1d.6.bias
F32
F32
[1, 1, 256]
a.conv1d.6.weight
F32
F32
[1, 1, 256, 256]
a.pre_encode.out.bias
F32
F32
[512]
a.pre_encode.out.weight
Q8_0
Q8_0
[4096, 512]
mm.1.bias
F32
F32
[2048]
mm.1.weight
Q8_0
Q8_0
[3072, 2048]
mm.2.bias
F32
F32
[1024]
mm.2.weight
Q8_0
Q8_0
[2048, 1024]
mm.a.mlp.0.bias
F32
F32
[512]
mm.a.mlp.0.weight
F32
F32
[512]
mm.a.mlp.1.bias
F32
F32
[1024]
mm.a.mlp.1.weight
Q8_0
Q8_0
[512, 1024]
mm.a.mlp.3.bias
F32
F32
[1024]
mm.a.mlp.3.weight
Q8_0
Q8_0
[1024, 1024]
mm.a.mlp.4.bias
F32
F32
[1024]
mm.a.mlp.4.weight
F32
F32
[1024]
mm.a.mlp.5.bias
F32
F32
[512]
mm.a.mlp.5.weight
Q8_0
Q8_0
[1024, 512]
mm.a.mlp.6.bias
F32
F32
[1024]
mm.a.mlp.6.weight
Q8_0
Q8_0
[512, 1024]
v.blk.0
v.blk.0.attn_k.bias
F32
F32
[768]
v.blk.0.attn_k.weight
Q8_0
Q8_0
[768, 768]
v.blk.0.attn_out.bias
F32
F32
[768]
v.blk.0.attn_out.weight
Q8_0
Q8_0
[768, 768]
v.blk.0.attn_q.bias
F32
F32
[768]
v.blk.0.attn_q.weight
Q8_0
Q8_0
[768, 768]
v.blk.0.attn_v.bias
F32
F32
[768]
v.blk.0.attn_v.weight
Q8_0
Q8_0
[768, 768]
v.blk.0.ffn_down.bias
F32
F32
[768]
v.blk.0.ffn_down.weight
Q8_0
Q8_0
[3072, 768]
v.blk.0.ffn_up.bias
F32
F32
[3072]
v.blk.0.ffn_up.weight
Q8_0
Q8_0
[768, 3072]
v.blk.0.ln1.bias
F32
F32
[768]
v.blk.0.ln1.weight
F32
F32
[768]
v.blk.0.ln2.bias
F32
F32
[768]
v.blk.0.ln2.weight
F32
F32
[768]
v.blk.1
v.blk.1.attn_k.bias
F32
F32
[768]
v.blk.1.attn_k.weight
Q8_0
Q8_0
[768, 768]
v.blk.1.attn_out.bias
F32
F32
[768]
v.blk.1.attn_out.weight
Q8_0
Q8_0
[768, 768]
v.blk.1.attn_q.bias
F32
F32
[768]
v.blk.1.attn_q.weight
Q8_0
Q8_0
[768, 768]
v.blk.1.attn_v.bias
F32
F32
[768]
v.blk.1.attn_v.weight
Q8_0
Q8_0
[768, 768]
v.blk.1.ffn_down.bias
F32
F32
[768]
v.blk.1.ffn_down.weight
Q8_0
Q8_0
[3072, 768]
v.blk.1.ffn_up.bias
F32
F32
[3072]
v.blk.1.ffn_up.weight
Q8_0
Q8_0
[768, 3072]
v.blk.1.ln1.bias
F32
F32
[768]
v.blk.1.ln1.weight
F32
F32
[768]
v.blk.1.ln2.bias
F32
F32
[768]
v.blk.1.ln2.weight
F32
F32
[768]
v.blk.2
v.blk.2.attn_k.bias
F32
F32
[768]
v.blk.2.attn_k.weight
Q8_0
Q8_0
[768, 768]
v.blk.2.attn_out.bias
F32
F32
[768]
v.blk.2.attn_out.weight
Q8_0
Q8_0
[768, 768]
v.blk.2.attn_q.bias
F32
F32
[768]
v.blk.2.attn_q.weight
Q8_0
Q8_0
[768, 768]
v.blk.2.attn_v.bias
F32
F32
[768]
v.blk.2.attn_v.weight
Q8_0
Q8_0
[768, 768]
v.blk.2.ffn_down.bias
F32
F32
[768]
v.blk.2.ffn_down.weight
Q8_0
Q8_0
[3072, 768]
v.blk.2.ffn_up.bias
F32
F32
[3072]
v.blk.2.ffn_up.weight
Q8_0
Q8_0
[768, 3072]
v.blk.2.ln1.bias
F32
F32
[768]
v.blk.2.ln1.weight
F32
F32
[768]
v.blk.2.ln2.bias
F32
F32
[768]
v.blk.2.ln2.weight
F32
F32
[768]
v.blk.3
v.blk.3.attn_k.bias
F32
F32
[768]
v.blk.3.attn_k.weight
Q8_0
Q8_0
[768, 768]
v.blk.3.attn_out.bias
F32
F32
[768]
v.blk.3.attn_out.weight
Q8_0
Q8_0
[768, 768]
v.blk.3.attn_q.bias
F32
F32
[768]
v.blk.3.attn_q.weight
Q8_0
Q8_0
[768, 768]
v.blk.3.attn_v.bias
F32
F32
[768]
v.blk.3.attn_v.weight
Q8_0
Q8_0
[768, 768]
v.blk.3.ffn_down.bias
F32
F32
[768]
v.blk.3.ffn_down.weight
Q8_0
Q8_0
[3072, 768]
v.blk.3.ffn_up.bias
F32
F32
[3072]
v.blk.3.ffn_up.weight
Q8_0
Q8_0
[768, 3072]
v.blk.3.ln1.bias
F32
F32
[768]
v.blk.3.ln1.weight
F32
F32
[768]
v.blk.3.ln2.bias
F32
F32
[768]
v.blk.3.ln2.weight
F32
F32
[768]
v.blk.4
v.blk.4.attn_k.bias
F32
F32
[768]
v.blk.4.attn_k.weight
Q8_0
Q8_0
[768, 768]
v.blk.4.attn_out.bias
F32
F32
[768]
v.blk.4.attn_out.weight
Q8_0
Q8_0
[768, 768]
v.blk.4.attn_q.bias
F32
F32
[768]
v.blk.4.attn_q.weight
Q8_0
Q8_0
[768, 768]
v.blk.4.attn_v.bias
F32
F32
[768]
v.blk.4.attn_v.weight
Q8_0
Q8_0
[768, 768]
v.blk.4.ffn_down.bias
F32
F32
[768]
v.blk.4.ffn_down.weight
Q8_0
Q8_0
[3072, 768]
v.blk.4.ffn_up.bias
F32
F32
[3072]
v.blk.4.ffn_up.weight
Q8_0
Q8_0
[768, 3072]
v.blk.4.ln1.bias
F32
F32
[768]
v.blk.4.ln1.weight
F32
F32
[768]
v.blk.4.ln2.bias
F32
F32
[768]
v.blk.4.ln2.weight
F32
F32
[768]
v.blk.5
v.blk.5.attn_k.bias
F32
F32
[768]
v.blk.5.attn_k.weight
Q8_0
Q8_0
[768, 768]
v.blk.5.attn_out.bias
F32
F32
[768]
v.blk.5.attn_out.weight
Q8_0
Q8_0
[768, 768]
v.blk.5.attn_q.bias
F32
F32
[768]
v.blk.5.attn_q.weight
Q8_0
Q8_0
[768, 768]
v.blk.5.attn_v.bias
F32
F32
[768]
v.blk.5.attn_v.weight
Q8_0
Q8_0
[768, 768]
v.blk.5.ffn_down.bias
F32
F32
[768]
v.blk.5.ffn_down.weight
Q8_0
Q8_0
[3072, 768]
v.blk.5.ffn_up.bias
F32
F32
[3072]
v.blk.5.ffn_up.weight
Q8_0
Q8_0
[768, 3072]
v.blk.5.ln1.bias
F32
F32
[768]
v.blk.5.ln1.weight
F32
F32
[768]
v.blk.5.ln2.bias
F32
F32
[768]
v.blk.5.ln2.weight
F32
F32
[768]
v.blk.6
v.blk.6.attn_k.bias
F32
F32
[768]
v.blk.6.attn_k.weight
Q8_0
Q8_0
[768, 768]
v.blk.6.attn_out.bias
F32
F32
[768]
v.blk.6.attn_out.weight
Q8_0
Q8_0
[768, 768]
v.blk.6.attn_q.bias
F32
F32
[768]
v.blk.6.attn_q.weight
Q8_0
Q8_0
[768, 768]
v.blk.6.attn_v.bias
F32
F32
[768]
v.blk.6.attn_v.weight
Q8_0
Q8_0
[768, 768]
v.blk.6.ffn_down.bias
F32
F32
[768]
v.blk.6.ffn_down.weight
Q8_0
Q8_0
[3072, 768]
v.blk.6.ffn_up.bias
F32
F32
[3072]
v.blk.6.ffn_up.weight
Q8_0
Q8_0
[768, 3072]
v.blk.6.ln1.bias
F32
F32
[768]
v.blk.6.ln1.weight
F32
F32
[768]
v.blk.6.ln2.bias
F32
F32
[768]
v.blk.6.ln2.weight
F32
F32
[768]
v.blk.7
v.blk.7.attn_k.bias
F32
F32
[768]
v.blk.7.attn_k.weight
Q8_0
Q8_0
[768, 768]
v.blk.7.attn_out.bias
F32
F32
[768]
v.blk.7.attn_out.weight
Q8_0
Q8_0
[768, 768]
v.blk.7.attn_q.bias
F32
F32
[768]
v.blk.7.attn_q.weight
Q8_0
Q8_0
[768, 768]
v.blk.7.attn_v.bias
F32
F32
[768]
v.blk.7.attn_v.weight
Q8_0
Q8_0
[768, 768]
v.blk.7.ffn_down.bias
F32
F32
[768]
v.blk.7.ffn_down.weight
Q8_0
Q8_0
[3072, 768]
v.blk.7.ffn_up.bias
F32
F32
[3072]
v.blk.7.ffn_up.weight
Q8_0
Q8_0
[768, 3072]
v.blk.7.ln1.bias
F32
F32
[768]
v.blk.7.ln1.weight
F32
F32
[768]
v.blk.7.ln2.bias
F32
F32
[768]
v.blk.7.ln2.weight
F32
F32
[768]
v.blk.8
v.blk.8.attn_k.bias
F32
F32
[768]
v.blk.8.attn_k.weight
Q8_0
Q8_0
[768, 768]
v.blk.8.attn_out.bias
F32
F32
[768]
v.blk.8.attn_out.weight
Q8_0
Q8_0
[768, 768]
v.blk.8.attn_q.bias
F32
F32
[768]
v.blk.8.attn_q.weight
Q8_0
Q8_0
[768, 768]
v.blk.8.attn_v.bias
F32
F32
[768]
v.blk.8.attn_v.weight
Q8_0
Q8_0
[768, 768]
v.blk.8.ffn_down.bias
F32
F32
[768]
v.blk.8.ffn_down.weight
Q8_0
Q8_0
[3072, 768]
v.blk.8.ffn_up.bias
F32
F32
[3072]
v.blk.8.ffn_up.weight
Q8_0
Q8_0
[768, 3072]
v.blk.8.ln1.bias
F32
F32
[768]
v.blk.8.ln1.weight
F32
F32
[768]
v.blk.8.ln2.bias
F32
F32
[768]
v.blk.8.ln2.weight
F32
F32
[768]
v.blk.9
v.blk.9.attn_k.bias
F32
F32
[768]
v.blk.9.attn_k.weight
Q8_0
Q8_0
[768, 768]
v.blk.9.attn_out.bias
F32
F32
[768]
v.blk.9.attn_out.weight
Q8_0
Q8_0
[768, 768]
v.blk.9.attn_q.bias
F32
F32
[768]
v.blk.9.attn_q.weight
Q8_0
Q8_0
[768, 768]
v.blk.9.attn_v.bias
F32
F32
[768]
v.blk.9.attn_v.weight
Q8_0
Q8_0
[768, 768]
v.blk.9.ffn_down.bias
F32
F32
[768]
v.blk.9.ffn_down.weight
Q8_0
Q8_0
[3072, 768]
v.blk.9.ffn_up.bias
F32
F32
[3072]
v.blk.9.ffn_up.weight
Q8_0
Q8_0
[768, 3072]
v.blk.9.ln1.bias
F32
F32
[768]
v.blk.9.ln1.weight
F32
F32
[768]
v.blk.9.ln2.bias
F32
F32
[768]
v.blk.9.ln2.weight
F32
F32
[768]
v.blk.10
v.blk.10.attn_k.bias
F32
F32
[768]
v.blk.10.attn_k.weight
Q8_0
Q8_0
[768, 768]
v.blk.10.attn_out.bias
F32
F32
[768]
v.blk.10.attn_out.weight
Q8_0
Q8_0
[768, 768]
v.blk.10.attn_q.bias
F32
F32
[768]
v.blk.10.attn_q.weight
Q8_0
Q8_0
[768, 768]
v.blk.10.attn_v.bias
F32
F32
[768]
v.blk.10.attn_v.weight
Q8_0
Q8_0
[768, 768]
v.blk.10.ffn_down.bias
F32
F32
[768]
v.blk.10.ffn_down.weight
Q8_0
Q8_0
[3072, 768]
v.blk.10.ffn_up.bias
F32
F32
[3072]
v.blk.10.ffn_up.weight
Q8_0
Q8_0
[768, 3072]
v.blk.10.ln1.bias
F32
F32
[768]
v.blk.10.ln1.weight
F32
F32
[768]
v.blk.10.ln2.bias
F32
F32
[768]
v.blk.10.ln2.weight
F32
F32
[768]
v.blk.11
v.blk.11.attn_k.bias
F32
F32
[768]
v.blk.11.attn_k.weight
Q8_0
Q8_0
[768, 768]
v.blk.11.attn_out.bias
F32
F32
[768]
v.blk.11.attn_out.weight
Q8_0
Q8_0
[768, 768]
v.blk.11.attn_q.bias
F32
F32
[768]
v.blk.11.attn_q.weight
Q8_0
Q8_0
[768, 768]
v.blk.11.attn_v.bias
F32
F32
[768]
v.blk.11.attn_v.weight
Q8_0
Q8_0
[768, 768]
v.blk.11.ffn_down.bias
F32
F32
[768]
v.blk.11.ffn_down.weight
Q8_0
Q8_0
[3072, 768]
v.blk.11.ffn_up.bias
F32
F32
[3072]
v.blk.11.ffn_up.weight
Q8_0
Q8_0
[768, 3072]
v.blk.11.ln1.bias
F32
F32
[768]
v.blk.11.ln1.weight
F32
F32
[768]
v.blk.11.ln2.bias
F32
F32
[768]
v.blk.11.ln2.weight
F32
F32
[768]
v.patch_embd.bias
F32
F32
[768]
v.patch_embd.weight
F32
F32
[16, 16, 3, 768]
v.position_embd.weight
F32
F32
[768, 256]
v.post_ln.bias
F32
F32
[768]
v.post_ln.weight
F32
F32
[768]