Models
Docs
Pricing
Sign in
Download
Models
Download
Docs
Pricing
Sign in
ninjasatwork
/
qwen3-omni-abliterated
:30b
116
Downloads
Updated
2 weeks ago
Cancel
vision
tools
thinking
30b
qwen3-omni-abliterated:30b
...
/
projector
8d39e7365621 · 1.3GB
Metadata
general.architecture
clip
clip
general.file_type
Q8_0
Q8_0
clip.audio.attention.head_count
20
20
clip.audio.attention.layer_norm_epsilon
1e-05
1e-05
clip.audio.block_count
32
32
clip.audio.embedding_length
1280
1280
clip.audio.feed_forward_length
5120
5120
clip.audio.num_mel_bins
128
128
clip.audio.projection_dim
2048
2048
clip.audio.projector_type
qwen3a
qwen3a
clip.has_audio_encoder
true
true
clip.has_vision_encoder
true
true
clip.use_gelu
true
true
clip.vision.attention.head_count
16
16
clip.vision.attention.layer_norm_epsilon
1e-06
1e-06
clip.vision.block_count
27
27
clip.vision.embedding_length
1152
1152
clip.vision.feed_forward_length
4304
4304
clip.vision.image_mean
[0.5, 0.5, 0.5]
[0.5, 0.5, 0.5]
clip.vision.image_size
768
768
clip.vision.image_std
[0.5, 0.5, 0.5]
[0.5, 0.5, 0.5]
clip.vision.is_deepstack_layers
[false, false, false, false, false, ...]
[false, false, false, false, false, ...]
clip.vision.patch_size
16
16
clip.vision.projection_dim
2048
2048
clip.vision.projector_type
qwen3vl_merger
qwen3vl_merger
clip.vision.spatial_merge_size
2
2
Tensor
Name
Type
Shape
a.blk.0
a.blk.0.attn_k.bias
F32
F32
[1280]
a.blk.0.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.0.attn_out.bias
F32
F32
[1280]
a.blk.0.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.0.attn_q.bias
F32
F32
[1280]
a.blk.0.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.0.attn_v.bias
F32
F32
[1280]
a.blk.0.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.0.ffn_down.bias
F32
F32
[1280]
a.blk.0.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.0.ffn_up.bias
F32
F32
[5120]
a.blk.0.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.0.ln1.bias
F32
F32
[1280]
a.blk.0.ln1.weight
F32
F32
[1280]
a.blk.0.ln2.bias
F32
F32
[1280]
a.blk.0.ln2.weight
F32
F32
[1280]
a.blk.1
a.blk.1.attn_k.bias
F32
F32
[1280]
a.blk.1.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.1.attn_out.bias
F32
F32
[1280]
a.blk.1.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.1.attn_q.bias
F32
F32
[1280]
a.blk.1.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.1.attn_v.bias
F32
F32
[1280]
a.blk.1.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.1.ffn_down.bias
F32
F32
[1280]
a.blk.1.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.1.ffn_up.bias
F32
F32
[5120]
a.blk.1.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.1.ln1.bias
F32
F32
[1280]
a.blk.1.ln1.weight
F32
F32
[1280]
a.blk.1.ln2.bias
F32
F32
[1280]
a.blk.1.ln2.weight
F32
F32
[1280]
a.blk.2
a.blk.2.attn_k.bias
F32
F32
[1280]
a.blk.2.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.2.attn_out.bias
F32
F32
[1280]
a.blk.2.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.2.attn_q.bias
F32
F32
[1280]
a.blk.2.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.2.attn_v.bias
F32
F32
[1280]
a.blk.2.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.2.ffn_down.bias
F32
F32
[1280]
a.blk.2.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.2.ffn_up.bias
F32
F32
[5120]
a.blk.2.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.2.ln1.bias
F32
F32
[1280]
a.blk.2.ln1.weight
F32
F32
[1280]
a.blk.2.ln2.bias
F32
F32
[1280]
a.blk.2.ln2.weight
F32
F32
[1280]
a.blk.3
a.blk.3.attn_k.bias
F32
F32
[1280]
a.blk.3.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.3.attn_out.bias
F32
F32
[1280]
a.blk.3.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.3.attn_q.bias
F32
F32
[1280]
a.blk.3.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.3.attn_v.bias
F32
F32
[1280]
a.blk.3.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.3.ffn_down.bias
F32
F32
[1280]
a.blk.3.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.3.ffn_up.bias
F32
F32
[5120]
a.blk.3.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.3.ln1.bias
F32
F32
[1280]
a.blk.3.ln1.weight
F32
F32
[1280]
a.blk.3.ln2.bias
F32
F32
[1280]
a.blk.3.ln2.weight
F32
F32
[1280]
a.blk.4
a.blk.4.attn_k.bias
F32
F32
[1280]
a.blk.4.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.4.attn_out.bias
F32
F32
[1280]
a.blk.4.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.4.attn_q.bias
F32
F32
[1280]
a.blk.4.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.4.attn_v.bias
F32
F32
[1280]
a.blk.4.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.4.ffn_down.bias
F32
F32
[1280]
a.blk.4.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.4.ffn_up.bias
F32
F32
[5120]
a.blk.4.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.4.ln1.bias
F32
F32
[1280]
a.blk.4.ln1.weight
F32
F32
[1280]
a.blk.4.ln2.bias
F32
F32
[1280]
a.blk.4.ln2.weight
F32
F32
[1280]
a.blk.5
a.blk.5.attn_k.bias
F32
F32
[1280]
a.blk.5.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.5.attn_out.bias
F32
F32
[1280]
a.blk.5.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.5.attn_q.bias
F32
F32
[1280]
a.blk.5.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.5.attn_v.bias
F32
F32
[1280]
a.blk.5.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.5.ffn_down.bias
F32
F32
[1280]
a.blk.5.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.5.ffn_up.bias
F32
F32
[5120]
a.blk.5.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.5.ln1.bias
F32
F32
[1280]
a.blk.5.ln1.weight
F32
F32
[1280]
a.blk.5.ln2.bias
F32
F32
[1280]
a.blk.5.ln2.weight
F32
F32
[1280]
a.blk.6
a.blk.6.attn_k.bias
F32
F32
[1280]
a.blk.6.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.6.attn_out.bias
F32
F32
[1280]
a.blk.6.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.6.attn_q.bias
F32
F32
[1280]
a.blk.6.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.6.attn_v.bias
F32
F32
[1280]
a.blk.6.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.6.ffn_down.bias
F32
F32
[1280]
a.blk.6.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.6.ffn_up.bias
F32
F32
[5120]
a.blk.6.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.6.ln1.bias
F32
F32
[1280]
a.blk.6.ln1.weight
F32
F32
[1280]
a.blk.6.ln2.bias
F32
F32
[1280]
a.blk.6.ln2.weight
F32
F32
[1280]
a.blk.7
a.blk.7.attn_k.bias
F32
F32
[1280]
a.blk.7.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.7.attn_out.bias
F32
F32
[1280]
a.blk.7.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.7.attn_q.bias
F32
F32
[1280]
a.blk.7.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.7.attn_v.bias
F32
F32
[1280]
a.blk.7.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.7.ffn_down.bias
F32
F32
[1280]
a.blk.7.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.7.ffn_up.bias
F32
F32
[5120]
a.blk.7.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.7.ln1.bias
F32
F32
[1280]
a.blk.7.ln1.weight
F32
F32
[1280]
a.blk.7.ln2.bias
F32
F32
[1280]
a.blk.7.ln2.weight
F32
F32
[1280]
a.blk.8
a.blk.8.attn_k.bias
F32
F32
[1280]
a.blk.8.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.8.attn_out.bias
F32
F32
[1280]
a.blk.8.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.8.attn_q.bias
F32
F32
[1280]
a.blk.8.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.8.attn_v.bias
F32
F32
[1280]
a.blk.8.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.8.ffn_down.bias
F32
F32
[1280]
a.blk.8.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.8.ffn_up.bias
F32
F32
[5120]
a.blk.8.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.8.ln1.bias
F32
F32
[1280]
a.blk.8.ln1.weight
F32
F32
[1280]
a.blk.8.ln2.bias
F32
F32
[1280]
a.blk.8.ln2.weight
F32
F32
[1280]
a.blk.9
a.blk.9.attn_k.bias
F32
F32
[1280]
a.blk.9.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.9.attn_out.bias
F32
F32
[1280]
a.blk.9.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.9.attn_q.bias
F32
F32
[1280]
a.blk.9.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.9.attn_v.bias
F32
F32
[1280]
a.blk.9.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.9.ffn_down.bias
F32
F32
[1280]
a.blk.9.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.9.ffn_up.bias
F32
F32
[5120]
a.blk.9.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.9.ln1.bias
F32
F32
[1280]
a.blk.9.ln1.weight
F32
F32
[1280]
a.blk.9.ln2.bias
F32
F32
[1280]
a.blk.9.ln2.weight
F32
F32
[1280]
a.blk.10
a.blk.10.attn_k.bias
F32
F32
[1280]
a.blk.10.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.10.attn_out.bias
F32
F32
[1280]
a.blk.10.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.10.attn_q.bias
F32
F32
[1280]
a.blk.10.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.10.attn_v.bias
F32
F32
[1280]
a.blk.10.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.10.ffn_down.bias
F32
F32
[1280]
a.blk.10.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.10.ffn_up.bias
F32
F32
[5120]
a.blk.10.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.10.ln1.bias
F32
F32
[1280]
a.blk.10.ln1.weight
F32
F32
[1280]
a.blk.10.ln2.bias
F32
F32
[1280]
a.blk.10.ln2.weight
F32
F32
[1280]
a.blk.11
a.blk.11.attn_k.bias
F32
F32
[1280]
a.blk.11.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.11.attn_out.bias
F32
F32
[1280]
a.blk.11.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.11.attn_q.bias
F32
F32
[1280]
a.blk.11.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.11.attn_v.bias
F32
F32
[1280]
a.blk.11.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.11.ffn_down.bias
F32
F32
[1280]
a.blk.11.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.11.ffn_up.bias
F32
F32
[5120]
a.blk.11.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.11.ln1.bias
F32
F32
[1280]
a.blk.11.ln1.weight
F32
F32
[1280]
a.blk.11.ln2.bias
F32
F32
[1280]
a.blk.11.ln2.weight
F32
F32
[1280]
a.blk.12
a.blk.12.attn_k.bias
F32
F32
[1280]
a.blk.12.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.12.attn_out.bias
F32
F32
[1280]
a.blk.12.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.12.attn_q.bias
F32
F32
[1280]
a.blk.12.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.12.attn_v.bias
F32
F32
[1280]
a.blk.12.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.12.ffn_down.bias
F32
F32
[1280]
a.blk.12.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.12.ffn_up.bias
F32
F32
[5120]
a.blk.12.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.12.ln1.bias
F32
F32
[1280]
a.blk.12.ln1.weight
F32
F32
[1280]
a.blk.12.ln2.bias
F32
F32
[1280]
a.blk.12.ln2.weight
F32
F32
[1280]
a.blk.13
a.blk.13.attn_k.bias
F32
F32
[1280]
a.blk.13.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.13.attn_out.bias
F32
F32
[1280]
a.blk.13.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.13.attn_q.bias
F32
F32
[1280]
a.blk.13.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.13.attn_v.bias
F32
F32
[1280]
a.blk.13.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.13.ffn_down.bias
F32
F32
[1280]
a.blk.13.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.13.ffn_up.bias
F32
F32
[5120]
a.blk.13.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.13.ln1.bias
F32
F32
[1280]
a.blk.13.ln1.weight
F32
F32
[1280]
a.blk.13.ln2.bias
F32
F32
[1280]
a.blk.13.ln2.weight
F32
F32
[1280]
a.blk.14
a.blk.14.attn_k.bias
F32
F32
[1280]
a.blk.14.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.14.attn_out.bias
F32
F32
[1280]
a.blk.14.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.14.attn_q.bias
F32
F32
[1280]
a.blk.14.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.14.attn_v.bias
F32
F32
[1280]
a.blk.14.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.14.ffn_down.bias
F32
F32
[1280]
a.blk.14.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.14.ffn_up.bias
F32
F32
[5120]
a.blk.14.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.14.ln1.bias
F32
F32
[1280]
a.blk.14.ln1.weight
F32
F32
[1280]
a.blk.14.ln2.bias
F32
F32
[1280]
a.blk.14.ln2.weight
F32
F32
[1280]
a.blk.15
a.blk.15.attn_k.bias
F32
F32
[1280]
a.blk.15.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.15.attn_out.bias
F32
F32
[1280]
a.blk.15.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.15.attn_q.bias
F32
F32
[1280]
a.blk.15.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.15.attn_v.bias
F32
F32
[1280]
a.blk.15.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.15.ffn_down.bias
F32
F32
[1280]
a.blk.15.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.15.ffn_up.bias
F32
F32
[5120]
a.blk.15.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.15.ln1.bias
F32
F32
[1280]
a.blk.15.ln1.weight
F32
F32
[1280]
a.blk.15.ln2.bias
F32
F32
[1280]
a.blk.15.ln2.weight
F32
F32
[1280]
a.blk.16
a.blk.16.attn_k.bias
F32
F32
[1280]
a.blk.16.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.16.attn_out.bias
F32
F32
[1280]
a.blk.16.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.16.attn_q.bias
F32
F32
[1280]
a.blk.16.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.16.attn_v.bias
F32
F32
[1280]
a.blk.16.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.16.ffn_down.bias
F32
F32
[1280]
a.blk.16.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.16.ffn_up.bias
F32
F32
[5120]
a.blk.16.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.16.ln1.bias
F32
F32
[1280]
a.blk.16.ln1.weight
F32
F32
[1280]
a.blk.16.ln2.bias
F32
F32
[1280]
a.blk.16.ln2.weight
F32
F32
[1280]
a.blk.17
a.blk.17.attn_k.bias
F32
F32
[1280]
a.blk.17.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.17.attn_out.bias
F32
F32
[1280]
a.blk.17.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.17.attn_q.bias
F32
F32
[1280]
a.blk.17.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.17.attn_v.bias
F32
F32
[1280]
a.blk.17.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.17.ffn_down.bias
F32
F32
[1280]
a.blk.17.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.17.ffn_up.bias
F32
F32
[5120]
a.blk.17.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.17.ln1.bias
F32
F32
[1280]
a.blk.17.ln1.weight
F32
F32
[1280]
a.blk.17.ln2.bias
F32
F32
[1280]
a.blk.17.ln2.weight
F32
F32
[1280]
a.blk.18
a.blk.18.attn_k.bias
F32
F32
[1280]
a.blk.18.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.18.attn_out.bias
F32
F32
[1280]
a.blk.18.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.18.attn_q.bias
F32
F32
[1280]
a.blk.18.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.18.attn_v.bias
F32
F32
[1280]
a.blk.18.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.18.ffn_down.bias
F32
F32
[1280]
a.blk.18.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.18.ffn_up.bias
F32
F32
[5120]
a.blk.18.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.18.ln1.bias
F32
F32
[1280]
a.blk.18.ln1.weight
F32
F32
[1280]
a.blk.18.ln2.bias
F32
F32
[1280]
a.blk.18.ln2.weight
F32
F32
[1280]
a.blk.19
a.blk.19.attn_k.bias
F32
F32
[1280]
a.blk.19.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.19.attn_out.bias
F32
F32
[1280]
a.blk.19.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.19.attn_q.bias
F32
F32
[1280]
a.blk.19.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.19.attn_v.bias
F32
F32
[1280]
a.blk.19.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.19.ffn_down.bias
F32
F32
[1280]
a.blk.19.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.19.ffn_up.bias
F32
F32
[5120]
a.blk.19.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.19.ln1.bias
F32
F32
[1280]
a.blk.19.ln1.weight
F32
F32
[1280]
a.blk.19.ln2.bias
F32
F32
[1280]
a.blk.19.ln2.weight
F32
F32
[1280]
a.blk.20
a.blk.20.attn_k.bias
F32
F32
[1280]
a.blk.20.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.20.attn_out.bias
F32
F32
[1280]
a.blk.20.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.20.attn_q.bias
F32
F32
[1280]
a.blk.20.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.20.attn_v.bias
F32
F32
[1280]
a.blk.20.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.20.ffn_down.bias
F32
F32
[1280]
a.blk.20.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.20.ffn_up.bias
F32
F32
[5120]
a.blk.20.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.20.ln1.bias
F32
F32
[1280]
a.blk.20.ln1.weight
F32
F32
[1280]
a.blk.20.ln2.bias
F32
F32
[1280]
a.blk.20.ln2.weight
F32
F32
[1280]
a.blk.21
a.blk.21.attn_k.bias
F32
F32
[1280]
a.blk.21.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.21.attn_out.bias
F32
F32
[1280]
a.blk.21.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.21.attn_q.bias
F32
F32
[1280]
a.blk.21.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.21.attn_v.bias
F32
F32
[1280]
a.blk.21.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.21.ffn_down.bias
F32
F32
[1280]
a.blk.21.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.21.ffn_up.bias
F32
F32
[5120]
a.blk.21.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.21.ln1.bias
F32
F32
[1280]
a.blk.21.ln1.weight
F32
F32
[1280]
a.blk.21.ln2.bias
F32
F32
[1280]
a.blk.21.ln2.weight
F32
F32
[1280]
a.blk.22
a.blk.22.attn_k.bias
F32
F32
[1280]
a.blk.22.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.22.attn_out.bias
F32
F32
[1280]
a.blk.22.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.22.attn_q.bias
F32
F32
[1280]
a.blk.22.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.22.attn_v.bias
F32
F32
[1280]
a.blk.22.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.22.ffn_down.bias
F32
F32
[1280]
a.blk.22.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.22.ffn_up.bias
F32
F32
[5120]
a.blk.22.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.22.ln1.bias
F32
F32
[1280]
a.blk.22.ln1.weight
F32
F32
[1280]
a.blk.22.ln2.bias
F32
F32
[1280]
a.blk.22.ln2.weight
F32
F32
[1280]
a.blk.23
a.blk.23.attn_k.bias
F32
F32
[1280]
a.blk.23.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.23.attn_out.bias
F32
F32
[1280]
a.blk.23.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.23.attn_q.bias
F32
F32
[1280]
a.blk.23.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.23.attn_v.bias
F32
F32
[1280]
a.blk.23.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.23.ffn_down.bias
F32
F32
[1280]
a.blk.23.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.23.ffn_up.bias
F32
F32
[5120]
a.blk.23.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.23.ln1.bias
F32
F32
[1280]
a.blk.23.ln1.weight
F32
F32
[1280]
a.blk.23.ln2.bias
F32
F32
[1280]
a.blk.23.ln2.weight
F32
F32
[1280]
a.blk.24
a.blk.24.attn_k.bias
F32
F32
[1280]
a.blk.24.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.24.attn_out.bias
F32
F32
[1280]
a.blk.24.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.24.attn_q.bias
F32
F32
[1280]
a.blk.24.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.24.attn_v.bias
F32
F32
[1280]
a.blk.24.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.24.ffn_down.bias
F32
F32
[1280]
a.blk.24.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.24.ffn_up.bias
F32
F32
[5120]
a.blk.24.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.24.ln1.bias
F32
F32
[1280]
a.blk.24.ln1.weight
F32
F32
[1280]
a.blk.24.ln2.bias
F32
F32
[1280]
a.blk.24.ln2.weight
F32
F32
[1280]
a.blk.25
a.blk.25.attn_k.bias
F32
F32
[1280]
a.blk.25.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.25.attn_out.bias
F32
F32
[1280]
a.blk.25.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.25.attn_q.bias
F32
F32
[1280]
a.blk.25.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.25.attn_v.bias
F32
F32
[1280]
a.blk.25.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.25.ffn_down.bias
F32
F32
[1280]
a.blk.25.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.25.ffn_up.bias
F32
F32
[5120]
a.blk.25.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.25.ln1.bias
F32
F32
[1280]
a.blk.25.ln1.weight
F32
F32
[1280]
a.blk.25.ln2.bias
F32
F32
[1280]
a.blk.25.ln2.weight
F32
F32
[1280]
a.blk.26
a.blk.26.attn_k.bias
F32
F32
[1280]
a.blk.26.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.26.attn_out.bias
F32
F32
[1280]
a.blk.26.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.26.attn_q.bias
F32
F32
[1280]
a.blk.26.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.26.attn_v.bias
F32
F32
[1280]
a.blk.26.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.26.ffn_down.bias
F32
F32
[1280]
a.blk.26.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.26.ffn_up.bias
F32
F32
[5120]
a.blk.26.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.26.ln1.bias
F32
F32
[1280]
a.blk.26.ln1.weight
F32
F32
[1280]
a.blk.26.ln2.bias
F32
F32
[1280]
a.blk.26.ln2.weight
F32
F32
[1280]
a.blk.27
a.blk.27.attn_k.bias
F32
F32
[1280]
a.blk.27.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.27.attn_out.bias
F32
F32
[1280]
a.blk.27.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.27.attn_q.bias
F32
F32
[1280]
a.blk.27.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.27.attn_v.bias
F32
F32
[1280]
a.blk.27.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.27.ffn_down.bias
F32
F32
[1280]
a.blk.27.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.27.ffn_up.bias
F32
F32
[5120]
a.blk.27.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.27.ln1.bias
F32
F32
[1280]
a.blk.27.ln1.weight
F32
F32
[1280]
a.blk.27.ln2.bias
F32
F32
[1280]
a.blk.27.ln2.weight
F32
F32
[1280]
a.blk.28
a.blk.28.attn_k.bias
F32
F32
[1280]
a.blk.28.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.28.attn_out.bias
F32
F32
[1280]
a.blk.28.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.28.attn_q.bias
F32
F32
[1280]
a.blk.28.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.28.attn_v.bias
F32
F32
[1280]
a.blk.28.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.28.ffn_down.bias
F32
F32
[1280]
a.blk.28.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.28.ffn_up.bias
F32
F32
[5120]
a.blk.28.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.28.ln1.bias
F32
F32
[1280]
a.blk.28.ln1.weight
F32
F32
[1280]
a.blk.28.ln2.bias
F32
F32
[1280]
a.blk.28.ln2.weight
F32
F32
[1280]
a.blk.29
a.blk.29.attn_k.bias
F32
F32
[1280]
a.blk.29.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.29.attn_out.bias
F32
F32
[1280]
a.blk.29.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.29.attn_q.bias
F32
F32
[1280]
a.blk.29.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.29.attn_v.bias
F32
F32
[1280]
a.blk.29.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.29.ffn_down.bias
F32
F32
[1280]
a.blk.29.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.29.ffn_up.bias
F32
F32
[5120]
a.blk.29.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.29.ln1.bias
F32
F32
[1280]
a.blk.29.ln1.weight
F32
F32
[1280]
a.blk.29.ln2.bias
F32
F32
[1280]
a.blk.29.ln2.weight
F32
F32
[1280]
a.blk.30
a.blk.30.attn_k.bias
F32
F32
[1280]
a.blk.30.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.30.attn_out.bias
F32
F32
[1280]
a.blk.30.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.30.attn_q.bias
F32
F32
[1280]
a.blk.30.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.30.attn_v.bias
F32
F32
[1280]
a.blk.30.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.30.ffn_down.bias
F32
F32
[1280]
a.blk.30.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.30.ffn_up.bias
F32
F32
[5120]
a.blk.30.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.30.ln1.bias
F32
F32
[1280]
a.blk.30.ln1.weight
F32
F32
[1280]
a.blk.30.ln2.bias
F32
F32
[1280]
a.blk.30.ln2.weight
F32
F32
[1280]
a.blk.31
a.blk.31.attn_k.bias
F32
F32
[1280]
a.blk.31.attn_k.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.31.attn_out.bias
F32
F32
[1280]
a.blk.31.attn_out.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.31.attn_q.bias
F32
F32
[1280]
a.blk.31.attn_q.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.31.attn_v.bias
F32
F32
[1280]
a.blk.31.attn_v.weight
Q8_0
Q8_0
[1280, 1280]
a.blk.31.ffn_down.bias
F32
F32
[1280]
a.blk.31.ffn_down.weight
Q8_0
Q8_0
[5120, 1280]
a.blk.31.ffn_up.bias
F32
F32
[5120]
a.blk.31.ffn_up.weight
Q8_0
Q8_0
[1280, 5120]
a.blk.31.ln1.bias
F32
F32
[1280]
a.blk.31.ln1.weight
F32
F32
[1280]
a.blk.31.ln2.bias
F32
F32
[1280]
a.blk.31.ln2.weight
F32
F32
[1280]
a.conv2d.1.bias
F32
F32
[1, 1, 480]
a.conv2d.1.weight
F16
F16
[3, 3, 1, 480]
a.conv2d.2.bias
F32
F32
[1, 1, 480]
a.conv2d.2.weight
F16
F16
[3, 3, 480, 480]
a.conv2d.3.bias
F32
F32
[1, 1, 480]
a.conv2d.3.weight
F16
F16
[3, 3, 480, 480]
a.conv_out.weight
F16
F16
[7680, 1280]
a.position_embd.weight
F32
F32
[1280, 1500]
a.post_ln.bias
F32
F32
[1280]
a.post_ln.weight
F32
F32
[1280]
mm.0.bias
F32
F32
[4608]
mm.0.weight
Q8_0
Q8_0
[4608, 4608]
mm.2.bias
F32
F32
[2048]
mm.2.weight
Q8_0
Q8_0
[4608, 2048]
mm.a.mlp.1.bias
F32
F32
[1280]
mm.a.mlp.1.weight
Q8_0
Q8_0
[1280, 1280]
mm.a.mlp.2.bias
F32
F32
[2048]
mm.a.mlp.2.weight
Q8_0
Q8_0
[1280, 2048]
v.blk.0
v.blk.0.attn_out.bias
F32
F32
[1152]
v.blk.0.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.0.attn_qkv.bias
F32
F32
[3456]
v.blk.0.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.0.ffn_down.bias
F32
F32
[1152]
v.blk.0.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.0.ffn_up.bias
F32
F32
[4304]
v.blk.0.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.0.ln1.bias
F32
F32
[1152]
v.blk.0.ln1.weight
F32
F32
[1152]
v.blk.0.ln2.bias
F32
F32
[1152]
v.blk.0.ln2.weight
F32
F32
[1152]
v.blk.1
v.blk.1.attn_out.bias
F32
F32
[1152]
v.blk.1.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.1.attn_qkv.bias
F32
F32
[3456]
v.blk.1.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.1.ffn_down.bias
F32
F32
[1152]
v.blk.1.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.1.ffn_up.bias
F32
F32
[4304]
v.blk.1.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.1.ln1.bias
F32
F32
[1152]
v.blk.1.ln1.weight
F32
F32
[1152]
v.blk.1.ln2.bias
F32
F32
[1152]
v.blk.1.ln2.weight
F32
F32
[1152]
v.blk.2
v.blk.2.attn_out.bias
F32
F32
[1152]
v.blk.2.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.2.attn_qkv.bias
F32
F32
[3456]
v.blk.2.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.2.ffn_down.bias
F32
F32
[1152]
v.blk.2.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.2.ffn_up.bias
F32
F32
[4304]
v.blk.2.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.2.ln1.bias
F32
F32
[1152]
v.blk.2.ln1.weight
F32
F32
[1152]
v.blk.2.ln2.bias
F32
F32
[1152]
v.blk.2.ln2.weight
F32
F32
[1152]
v.blk.3
v.blk.3.attn_out.bias
F32
F32
[1152]
v.blk.3.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.3.attn_qkv.bias
F32
F32
[3456]
v.blk.3.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.3.ffn_down.bias
F32
F32
[1152]
v.blk.3.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.3.ffn_up.bias
F32
F32
[4304]
v.blk.3.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.3.ln1.bias
F32
F32
[1152]
v.blk.3.ln1.weight
F32
F32
[1152]
v.blk.3.ln2.bias
F32
F32
[1152]
v.blk.3.ln2.weight
F32
F32
[1152]
v.blk.4
v.blk.4.attn_out.bias
F32
F32
[1152]
v.blk.4.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.4.attn_qkv.bias
F32
F32
[3456]
v.blk.4.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.4.ffn_down.bias
F32
F32
[1152]
v.blk.4.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.4.ffn_up.bias
F32
F32
[4304]
v.blk.4.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.4.ln1.bias
F32
F32
[1152]
v.blk.4.ln1.weight
F32
F32
[1152]
v.blk.4.ln2.bias
F32
F32
[1152]
v.blk.4.ln2.weight
F32
F32
[1152]
v.blk.5
v.blk.5.attn_out.bias
F32
F32
[1152]
v.blk.5.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.5.attn_qkv.bias
F32
F32
[3456]
v.blk.5.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.5.ffn_down.bias
F32
F32
[1152]
v.blk.5.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.5.ffn_up.bias
F32
F32
[4304]
v.blk.5.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.5.ln1.bias
F32
F32
[1152]
v.blk.5.ln1.weight
F32
F32
[1152]
v.blk.5.ln2.bias
F32
F32
[1152]
v.blk.5.ln2.weight
F32
F32
[1152]
v.blk.6
v.blk.6.attn_out.bias
F32
F32
[1152]
v.blk.6.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.6.attn_qkv.bias
F32
F32
[3456]
v.blk.6.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.6.ffn_down.bias
F32
F32
[1152]
v.blk.6.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.6.ffn_up.bias
F32
F32
[4304]
v.blk.6.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.6.ln1.bias
F32
F32
[1152]
v.blk.6.ln1.weight
F32
F32
[1152]
v.blk.6.ln2.bias
F32
F32
[1152]
v.blk.6.ln2.weight
F32
F32
[1152]
v.blk.7
v.blk.7.attn_out.bias
F32
F32
[1152]
v.blk.7.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.7.attn_qkv.bias
F32
F32
[3456]
v.blk.7.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.7.ffn_down.bias
F32
F32
[1152]
v.blk.7.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.7.ffn_up.bias
F32
F32
[4304]
v.blk.7.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.7.ln1.bias
F32
F32
[1152]
v.blk.7.ln1.weight
F32
F32
[1152]
v.blk.7.ln2.bias
F32
F32
[1152]
v.blk.7.ln2.weight
F32
F32
[1152]
v.blk.8
v.blk.8.attn_out.bias
F32
F32
[1152]
v.blk.8.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.8.attn_qkv.bias
F32
F32
[3456]
v.blk.8.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.8.ffn_down.bias
F32
F32
[1152]
v.blk.8.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.8.ffn_up.bias
F32
F32
[4304]
v.blk.8.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.8.ln1.bias
F32
F32
[1152]
v.blk.8.ln1.weight
F32
F32
[1152]
v.blk.8.ln2.bias
F32
F32
[1152]
v.blk.8.ln2.weight
F32
F32
[1152]
v.blk.9
v.blk.9.attn_out.bias
F32
F32
[1152]
v.blk.9.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.9.attn_qkv.bias
F32
F32
[3456]
v.blk.9.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.9.ffn_down.bias
F32
F32
[1152]
v.blk.9.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.9.ffn_up.bias
F32
F32
[4304]
v.blk.9.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.9.ln1.bias
F32
F32
[1152]
v.blk.9.ln1.weight
F32
F32
[1152]
v.blk.9.ln2.bias
F32
F32
[1152]
v.blk.9.ln2.weight
F32
F32
[1152]
v.blk.10
v.blk.10.attn_out.bias
F32
F32
[1152]
v.blk.10.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.10.attn_qkv.bias
F32
F32
[3456]
v.blk.10.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.10.ffn_down.bias
F32
F32
[1152]
v.blk.10.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.10.ffn_up.bias
F32
F32
[4304]
v.blk.10.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.10.ln1.bias
F32
F32
[1152]
v.blk.10.ln1.weight
F32
F32
[1152]
v.blk.10.ln2.bias
F32
F32
[1152]
v.blk.10.ln2.weight
F32
F32
[1152]
v.blk.11
v.blk.11.attn_out.bias
F32
F32
[1152]
v.blk.11.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.11.attn_qkv.bias
F32
F32
[3456]
v.blk.11.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.11.ffn_down.bias
F32
F32
[1152]
v.blk.11.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.11.ffn_up.bias
F32
F32
[4304]
v.blk.11.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.11.ln1.bias
F32
F32
[1152]
v.blk.11.ln1.weight
F32
F32
[1152]
v.blk.11.ln2.bias
F32
F32
[1152]
v.blk.11.ln2.weight
F32
F32
[1152]
v.blk.12
v.blk.12.attn_out.bias
F32
F32
[1152]
v.blk.12.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.12.attn_qkv.bias
F32
F32
[3456]
v.blk.12.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.12.ffn_down.bias
F32
F32
[1152]
v.blk.12.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.12.ffn_up.bias
F32
F32
[4304]
v.blk.12.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.12.ln1.bias
F32
F32
[1152]
v.blk.12.ln1.weight
F32
F32
[1152]
v.blk.12.ln2.bias
F32
F32
[1152]
v.blk.12.ln2.weight
F32
F32
[1152]
v.blk.13
v.blk.13.attn_out.bias
F32
F32
[1152]
v.blk.13.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.13.attn_qkv.bias
F32
F32
[3456]
v.blk.13.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.13.ffn_down.bias
F32
F32
[1152]
v.blk.13.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.13.ffn_up.bias
F32
F32
[4304]
v.blk.13.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.13.ln1.bias
F32
F32
[1152]
v.blk.13.ln1.weight
F32
F32
[1152]
v.blk.13.ln2.bias
F32
F32
[1152]
v.blk.13.ln2.weight
F32
F32
[1152]
v.blk.14
v.blk.14.attn_out.bias
F32
F32
[1152]
v.blk.14.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.14.attn_qkv.bias
F32
F32
[3456]
v.blk.14.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.14.ffn_down.bias
F32
F32
[1152]
v.blk.14.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.14.ffn_up.bias
F32
F32
[4304]
v.blk.14.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.14.ln1.bias
F32
F32
[1152]
v.blk.14.ln1.weight
F32
F32
[1152]
v.blk.14.ln2.bias
F32
F32
[1152]
v.blk.14.ln2.weight
F32
F32
[1152]
v.blk.15
v.blk.15.attn_out.bias
F32
F32
[1152]
v.blk.15.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.15.attn_qkv.bias
F32
F32
[3456]
v.blk.15.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.15.ffn_down.bias
F32
F32
[1152]
v.blk.15.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.15.ffn_up.bias
F32
F32
[4304]
v.blk.15.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.15.ln1.bias
F32
F32
[1152]
v.blk.15.ln1.weight
F32
F32
[1152]
v.blk.15.ln2.bias
F32
F32
[1152]
v.blk.15.ln2.weight
F32
F32
[1152]
v.blk.16
v.blk.16.attn_out.bias
F32
F32
[1152]
v.blk.16.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.16.attn_qkv.bias
F32
F32
[3456]
v.blk.16.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.16.ffn_down.bias
F32
F32
[1152]
v.blk.16.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.16.ffn_up.bias
F32
F32
[4304]
v.blk.16.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.16.ln1.bias
F32
F32
[1152]
v.blk.16.ln1.weight
F32
F32
[1152]
v.blk.16.ln2.bias
F32
F32
[1152]
v.blk.16.ln2.weight
F32
F32
[1152]
v.blk.17
v.blk.17.attn_out.bias
F32
F32
[1152]
v.blk.17.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.17.attn_qkv.bias
F32
F32
[3456]
v.blk.17.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.17.ffn_down.bias
F32
F32
[1152]
v.blk.17.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.17.ffn_up.bias
F32
F32
[4304]
v.blk.17.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.17.ln1.bias
F32
F32
[1152]
v.blk.17.ln1.weight
F32
F32
[1152]
v.blk.17.ln2.bias
F32
F32
[1152]
v.blk.17.ln2.weight
F32
F32
[1152]
v.blk.18
v.blk.18.attn_out.bias
F32
F32
[1152]
v.blk.18.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.18.attn_qkv.bias
F32
F32
[3456]
v.blk.18.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.18.ffn_down.bias
F32
F32
[1152]
v.blk.18.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.18.ffn_up.bias
F32
F32
[4304]
v.blk.18.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.18.ln1.bias
F32
F32
[1152]
v.blk.18.ln1.weight
F32
F32
[1152]
v.blk.18.ln2.bias
F32
F32
[1152]
v.blk.18.ln2.weight
F32
F32
[1152]
v.blk.19
v.blk.19.attn_out.bias
F32
F32
[1152]
v.blk.19.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.19.attn_qkv.bias
F32
F32
[3456]
v.blk.19.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.19.ffn_down.bias
F32
F32
[1152]
v.blk.19.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.19.ffn_up.bias
F32
F32
[4304]
v.blk.19.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.19.ln1.bias
F32
F32
[1152]
v.blk.19.ln1.weight
F32
F32
[1152]
v.blk.19.ln2.bias
F32
F32
[1152]
v.blk.19.ln2.weight
F32
F32
[1152]
v.blk.20
v.blk.20.attn_out.bias
F32
F32
[1152]
v.blk.20.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.20.attn_qkv.bias
F32
F32
[3456]
v.blk.20.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.20.ffn_down.bias
F32
F32
[1152]
v.blk.20.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.20.ffn_up.bias
F32
F32
[4304]
v.blk.20.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.20.ln1.bias
F32
F32
[1152]
v.blk.20.ln1.weight
F32
F32
[1152]
v.blk.20.ln2.bias
F32
F32
[1152]
v.blk.20.ln2.weight
F32
F32
[1152]
v.blk.21
v.blk.21.attn_out.bias
F32
F32
[1152]
v.blk.21.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.21.attn_qkv.bias
F32
F32
[3456]
v.blk.21.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.21.ffn_down.bias
F32
F32
[1152]
v.blk.21.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.21.ffn_up.bias
F32
F32
[4304]
v.blk.21.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.21.ln1.bias
F32
F32
[1152]
v.blk.21.ln1.weight
F32
F32
[1152]
v.blk.21.ln2.bias
F32
F32
[1152]
v.blk.21.ln2.weight
F32
F32
[1152]
v.blk.22
v.blk.22.attn_out.bias
F32
F32
[1152]
v.blk.22.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.22.attn_qkv.bias
F32
F32
[3456]
v.blk.22.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.22.ffn_down.bias
F32
F32
[1152]
v.blk.22.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.22.ffn_up.bias
F32
F32
[4304]
v.blk.22.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.22.ln1.bias
F32
F32
[1152]
v.blk.22.ln1.weight
F32
F32
[1152]
v.blk.22.ln2.bias
F32
F32
[1152]
v.blk.22.ln2.weight
F32
F32
[1152]
v.blk.23
v.blk.23.attn_out.bias
F32
F32
[1152]
v.blk.23.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.23.attn_qkv.bias
F32
F32
[3456]
v.blk.23.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.23.ffn_down.bias
F32
F32
[1152]
v.blk.23.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.23.ffn_up.bias
F32
F32
[4304]
v.blk.23.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.23.ln1.bias
F32
F32
[1152]
v.blk.23.ln1.weight
F32
F32
[1152]
v.blk.23.ln2.bias
F32
F32
[1152]
v.blk.23.ln2.weight
F32
F32
[1152]
v.blk.24
v.blk.24.attn_out.bias
F32
F32
[1152]
v.blk.24.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.24.attn_qkv.bias
F32
F32
[3456]
v.blk.24.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.24.ffn_down.bias
F32
F32
[1152]
v.blk.24.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.24.ffn_up.bias
F32
F32
[4304]
v.blk.24.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.24.ln1.bias
F32
F32
[1152]
v.blk.24.ln1.weight
F32
F32
[1152]
v.blk.24.ln2.bias
F32
F32
[1152]
v.blk.24.ln2.weight
F32
F32
[1152]
v.blk.25
v.blk.25.attn_out.bias
F32
F32
[1152]
v.blk.25.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.25.attn_qkv.bias
F32
F32
[3456]
v.blk.25.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.25.ffn_down.bias
F32
F32
[1152]
v.blk.25.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.25.ffn_up.bias
F32
F32
[4304]
v.blk.25.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.25.ln1.bias
F32
F32
[1152]
v.blk.25.ln1.weight
F32
F32
[1152]
v.blk.25.ln2.bias
F32
F32
[1152]
v.blk.25.ln2.weight
F32
F32
[1152]
v.blk.26
v.blk.26.attn_out.bias
F32
F32
[1152]
v.blk.26.attn_out.weight
Q8_0
Q8_0
[1152, 1152]
v.blk.26.attn_qkv.bias
F32
F32
[3456]
v.blk.26.attn_qkv.weight
Q8_0
Q8_0
[1152, 3456]
v.blk.26.ffn_down.bias
F32
F32
[1152]
v.blk.26.ffn_down.weight
F16
F16
[4304, 1152]
v.blk.26.ffn_up.bias
F32
F32
[4304]
v.blk.26.ffn_up.weight
Q8_0
Q8_0
[1152, 4304]
v.blk.26.ln1.bias
F32
F32
[1152]
v.blk.26.ln1.weight
F32
F32
[1152]
v.blk.26.ln2.bias
F32
F32
[1152]
v.blk.26.ln2.weight
F32
F32
[1152]
v.patch_embd.bias
F32
F32
[1152]
v.patch_embd.weight
F32
F32
[16, 16, 3, 1152]
v.patch_embd.weight.1
F32
F32
[16, 16, 3, 1152]
v.position_embd.weight
F32
F32
[1152, 2304]
v.post_ln.bias
F32
F32
[1152]
v.post_ln.weight
F32
F32
[1152]