Models
Docs
Pricing
Sign in
Download
Models
Download
Docs
Pricing
Sign in
qwen3.5
:27b-q4_K_M
36.2K
Downloads
Updated
just now
Qwen 3.5 is a family of open-source multimodal models that delivers exceptional utility and performance.
Qwen 3.5 is a family of open-source multimodal models that delivers exceptional utility and performance.
Cancel
vision
tools
thinking
cloud
35b
122b
qwen3.5:27b-q4_K_M
...
/
model
7935de6e08f9 · 17GB
Metadata
general.architecture
qwen35
qwen35
general.file_type
Q4_K_M
Q4_K_M
qwen35.attention.head_count
24
24
qwen35.attention.head_count_kv
[0, 0, 0, 4, 0, ...]
[0, 0, 0, 4, 0, ...]
qwen35.attention.key_length
256
256
qwen35.attention.layer_norm_rms_epsilon
1e-06
1e-06
qwen35.attention.value_length
256
256
qwen35.block_count
64
64
qwen35.context_length
262144
262144
qwen35.embedding_length
5120
5120
qwen35.feed_forward_length
17408
17408
qwen35.full_attention_interval
4
4
qwen35.image_token_id
248056
248056
qwen35.mrope_sections
[11, 11, 10]
[11, 11, 10]
qwen35.rope.dimension_count
64
64
qwen35.rope.dimension_sections
[11, 11, 10]
[11, 11, 10]
qwen35.rope.freq_base
1e+07
1e+07
qwen35.rope.mrope_interleaved
true
true
qwen35.rope.mrope_section
[11, 11, 10]
[11, 11, 10]
qwen35.ssm.conv_kernel
4
4
qwen35.ssm.group_count
16
16
qwen35.ssm.inner_size
6144
6144
qwen35.ssm.state_size
128
128
qwen35.ssm.time_step_rank
48
48
qwen35.ssm.v_head_reordered
true
true
qwen35.vision.attention.head_count
16
16
qwen35.vision.block_count
27
27
qwen35.vision.deepstack_visual_indexes
[]
[]
qwen35.vision.embedding_length
1152
1152
qwen35.vision.image_mean
[0.5, 0.5, 0.5]
[0.5, 0.5, 0.5]
qwen35.vision.image_std
[0.5, 0.5, 0.5]
[0.5, 0.5, 0.5]
qwen35.vision.longest_edge
16777216
16777216
qwen35.vision.num_channels
3
3
qwen35.vision.patch_size
16
16
qwen35.vision.shortest_edge
65536
65536
qwen35.vision.spatial_merge_size
2
2
qwen35.vision.temporal_patch_size
2
2
qwen35.vision_end_token_id
248054
248054
qwen35.vision_start_token_id
248053
248053
tokenizer.ggml.add_eos_token
false
false
tokenizer.ggml.add_padding_token
false
false
tokenizer.ggml.eos_token_id
248046
248046
tokenizer.ggml.eos_token_ids
[248046, 248044]
[248046, 248044]
tokenizer.ggml.merges
[Ġ Ġ, ĠĠ ĠĠ, i n, Ġ t, ĠĠĠĠ ĠĠĠĠ, ...]
[Ġ Ġ, ĠĠ ĠĠ, i n, Ġ t, ĠĠĠĠ ĠĠĠĠ, ...]
tokenizer.ggml.model
gpt2
gpt2
tokenizer.ggml.padding_token_id
248044
248044
tokenizer.ggml.pre
qwen35
qwen35
tokenizer.ggml.scores
[0, 1, 2, 3, 4, ...]
[0, 1, 2, 3, 4, ...]
tokenizer.ggml.token_type
[1, 1, 1, 1, 1, ...]
[1, 1, 1, 1, 1, ...]
tokenizer.ggml.tokens
[!, ", #, $, %, ...]
[!, ", #, $, %, ...]
Tensor
Name
Type
Shape
token_embd.weight
Q4_K
Q4_K
[5120, 248320]
blk.0
blk.0.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.0.attn_norm.weight
F32
F32
[5120]
blk.0.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.0.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.0.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.0.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.0.post_attention_norm.weight
F32
F32
[5120]
blk.0.ssm_a
F32
F32
[48]
blk.0.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.0.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.0.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.0.ssm_dt
F32
F32
[48]
blk.0.ssm_norm.weight
F32
F32
[128]
blk.0.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.1
blk.1.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.1.attn_norm.weight
F32
F32
[5120]
blk.1.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.1.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.1.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.1.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.1.post_attention_norm.weight
F32
F32
[5120]
blk.1.ssm_a
F32
F32
[48]
blk.1.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.1.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.1.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.1.ssm_dt
F32
F32
[48]
blk.1.ssm_norm.weight
F32
F32
[128]
blk.1.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.2
blk.2.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.2.attn_norm.weight
F32
F32
[5120]
blk.2.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.2.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.2.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.2.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.2.post_attention_norm.weight
F32
F32
[5120]
blk.2.ssm_a
F32
F32
[48]
blk.2.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.2.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.2.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.2.ssm_dt
F32
F32
[48]
blk.2.ssm_norm.weight
F32
F32
[128]
blk.2.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.3
blk.3.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.3.attn_k_norm.weight
F32
F32
[256]
blk.3.attn_norm.weight
F32
F32
[5120]
blk.3.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.3.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.3.attn_q_norm.weight
F32
F32
[256]
blk.3.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.3.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.3.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.3.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.3.post_attention_norm.weight
F32
F32
[5120]
blk.4
blk.4.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.4.attn_norm.weight
F32
F32
[5120]
blk.4.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.4.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.4.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.4.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.4.post_attention_norm.weight
F32
F32
[5120]
blk.4.ssm_a
F32
F32
[48]
blk.4.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.4.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.4.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.4.ssm_dt
F32
F32
[48]
blk.4.ssm_norm.weight
F32
F32
[128]
blk.4.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.5
blk.5.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.5.attn_norm.weight
F32
F32
[5120]
blk.5.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.5.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.5.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.5.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.5.post_attention_norm.weight
F32
F32
[5120]
blk.5.ssm_a
F32
F32
[48]
blk.5.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.5.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.5.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.5.ssm_dt
F32
F32
[48]
blk.5.ssm_norm.weight
F32
F32
[128]
blk.5.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.6
blk.6.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.6.attn_norm.weight
F32
F32
[5120]
blk.6.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.6.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.6.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.6.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.6.post_attention_norm.weight
F32
F32
[5120]
blk.6.ssm_a
F32
F32
[48]
blk.6.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.6.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.6.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.6.ssm_dt
F32
F32
[48]
blk.6.ssm_norm.weight
F32
F32
[128]
blk.6.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.7
blk.7.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.7.attn_k_norm.weight
F32
F32
[256]
blk.7.attn_norm.weight
F32
F32
[5120]
blk.7.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.7.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.7.attn_q_norm.weight
F32
F32
[256]
blk.7.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.7.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.7.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.7.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.7.post_attention_norm.weight
F32
F32
[5120]
blk.8
blk.8.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.8.attn_norm.weight
F32
F32
[5120]
blk.8.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.8.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.8.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.8.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.8.post_attention_norm.weight
F32
F32
[5120]
blk.8.ssm_a
F32
F32
[48]
blk.8.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.8.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.8.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.8.ssm_dt
F32
F32
[48]
blk.8.ssm_norm.weight
F32
F32
[128]
blk.8.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.9
blk.9.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.9.attn_norm.weight
F32
F32
[5120]
blk.9.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.9.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.9.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.9.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.9.post_attention_norm.weight
F32
F32
[5120]
blk.9.ssm_a
F32
F32
[48]
blk.9.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.9.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.9.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.9.ssm_dt
F32
F32
[48]
blk.9.ssm_norm.weight
F32
F32
[128]
blk.9.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.10
blk.10.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.10.attn_norm.weight
F32
F32
[5120]
blk.10.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.10.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.10.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.10.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.10.post_attention_norm.weight
F32
F32
[5120]
blk.10.ssm_a
F32
F32
[48]
blk.10.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.10.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.10.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.10.ssm_dt
F32
F32
[48]
blk.10.ssm_norm.weight
F32
F32
[128]
blk.10.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.11
blk.11.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.11.attn_k_norm.weight
F32
F32
[256]
blk.11.attn_norm.weight
F32
F32
[5120]
blk.11.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.11.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.11.attn_q_norm.weight
F32
F32
[256]
blk.11.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.11.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.11.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.11.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.11.post_attention_norm.weight
F32
F32
[5120]
blk.12
blk.12.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.12.attn_norm.weight
F32
F32
[5120]
blk.12.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.12.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.12.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.12.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.12.post_attention_norm.weight
F32
F32
[5120]
blk.12.ssm_a
F32
F32
[48]
blk.12.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.12.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.12.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.12.ssm_dt
F32
F32
[48]
blk.12.ssm_norm.weight
F32
F32
[128]
blk.12.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.13
blk.13.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.13.attn_norm.weight
F32
F32
[5120]
blk.13.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.13.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.13.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.13.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.13.post_attention_norm.weight
F32
F32
[5120]
blk.13.ssm_a
F32
F32
[48]
blk.13.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.13.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.13.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.13.ssm_dt
F32
F32
[48]
blk.13.ssm_norm.weight
F32
F32
[128]
blk.13.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.14
blk.14.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.14.attn_norm.weight
F32
F32
[5120]
blk.14.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.14.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.14.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.14.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.14.post_attention_norm.weight
F32
F32
[5120]
blk.14.ssm_a
F32
F32
[48]
blk.14.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.14.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.14.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.14.ssm_dt
F32
F32
[48]
blk.14.ssm_norm.weight
F32
F32
[128]
blk.14.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.15
blk.15.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.15.attn_k_norm.weight
F32
F32
[256]
blk.15.attn_norm.weight
F32
F32
[5120]
blk.15.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.15.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.15.attn_q_norm.weight
F32
F32
[256]
blk.15.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.15.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.15.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.15.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.15.post_attention_norm.weight
F32
F32
[5120]
blk.16
blk.16.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.16.attn_norm.weight
F32
F32
[5120]
blk.16.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.16.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.16.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.16.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.16.post_attention_norm.weight
F32
F32
[5120]
blk.16.ssm_a
F32
F32
[48]
blk.16.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.16.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.16.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.16.ssm_dt
F32
F32
[48]
blk.16.ssm_norm.weight
F32
F32
[128]
blk.16.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.17
blk.17.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.17.attn_norm.weight
F32
F32
[5120]
blk.17.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.17.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.17.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.17.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.17.post_attention_norm.weight
F32
F32
[5120]
blk.17.ssm_a
F32
F32
[48]
blk.17.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.17.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.17.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.17.ssm_dt
F32
F32
[48]
blk.17.ssm_norm.weight
F32
F32
[128]
blk.17.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.18
blk.18.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.18.attn_norm.weight
F32
F32
[5120]
blk.18.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.18.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.18.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.18.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.18.post_attention_norm.weight
F32
F32
[5120]
blk.18.ssm_a
F32
F32
[48]
blk.18.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.18.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.18.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.18.ssm_dt
F32
F32
[48]
blk.18.ssm_norm.weight
F32
F32
[128]
blk.18.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.19
blk.19.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.19.attn_k_norm.weight
F32
F32
[256]
blk.19.attn_norm.weight
F32
F32
[5120]
blk.19.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.19.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.19.attn_q_norm.weight
F32
F32
[256]
blk.19.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.19.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.19.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.19.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.19.post_attention_norm.weight
F32
F32
[5120]
blk.20
blk.20.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.20.attn_norm.weight
F32
F32
[5120]
blk.20.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.20.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.20.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.20.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.20.post_attention_norm.weight
F32
F32
[5120]
blk.20.ssm_a
F32
F32
[48]
blk.20.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.20.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.20.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.20.ssm_dt
F32
F32
[48]
blk.20.ssm_norm.weight
F32
F32
[128]
blk.20.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.21
blk.21.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.21.attn_norm.weight
F32
F32
[5120]
blk.21.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.21.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.21.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.21.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.21.post_attention_norm.weight
F32
F32
[5120]
blk.21.ssm_a
F32
F32
[48]
blk.21.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.21.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.21.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.21.ssm_dt
F32
F32
[48]
blk.21.ssm_norm.weight
F32
F32
[128]
blk.21.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.22
blk.22.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.22.attn_norm.weight
F32
F32
[5120]
blk.22.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.22.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.22.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.22.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.22.post_attention_norm.weight
F32
F32
[5120]
blk.22.ssm_a
F32
F32
[48]
blk.22.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.22.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.22.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.22.ssm_dt
F32
F32
[48]
blk.22.ssm_norm.weight
F32
F32
[128]
blk.22.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.23
blk.23.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.23.attn_k_norm.weight
F32
F32
[256]
blk.23.attn_norm.weight
F32
F32
[5120]
blk.23.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.23.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.23.attn_q_norm.weight
F32
F32
[256]
blk.23.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.23.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.23.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.23.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.23.post_attention_norm.weight
F32
F32
[5120]
blk.24
blk.24.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.24.attn_norm.weight
F32
F32
[5120]
blk.24.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.24.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.24.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.24.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.24.post_attention_norm.weight
F32
F32
[5120]
blk.24.ssm_a
F32
F32
[48]
blk.24.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.24.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.24.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.24.ssm_dt
F32
F32
[48]
blk.24.ssm_norm.weight
F32
F32
[128]
blk.24.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.25
blk.25.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.25.attn_norm.weight
F32
F32
[5120]
blk.25.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.25.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.25.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.25.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.25.post_attention_norm.weight
F32
F32
[5120]
blk.25.ssm_a
F32
F32
[48]
blk.25.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.25.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.25.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.25.ssm_dt
F32
F32
[48]
blk.25.ssm_norm.weight
F32
F32
[128]
blk.25.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.26
blk.26.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.26.attn_norm.weight
F32
F32
[5120]
blk.26.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.26.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.26.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.26.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.26.post_attention_norm.weight
F32
F32
[5120]
blk.26.ssm_a
F32
F32
[48]
blk.26.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.26.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.26.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.26.ssm_dt
F32
F32
[48]
blk.26.ssm_norm.weight
F32
F32
[128]
blk.26.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.27
blk.27.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.27.attn_k_norm.weight
F32
F32
[256]
blk.27.attn_norm.weight
F32
F32
[5120]
blk.27.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.27.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.27.attn_q_norm.weight
F32
F32
[256]
blk.27.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.27.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.27.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.27.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.27.post_attention_norm.weight
F32
F32
[5120]
blk.28
blk.28.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.28.attn_norm.weight
F32
F32
[5120]
blk.28.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.28.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.28.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.28.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.28.post_attention_norm.weight
F32
F32
[5120]
blk.28.ssm_a
F32
F32
[48]
blk.28.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.28.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.28.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.28.ssm_dt
F32
F32
[48]
blk.28.ssm_norm.weight
F32
F32
[128]
blk.28.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.29
blk.29.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.29.attn_norm.weight
F32
F32
[5120]
blk.29.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.29.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.29.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.29.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.29.post_attention_norm.weight
F32
F32
[5120]
blk.29.ssm_a
F32
F32
[48]
blk.29.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.29.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.29.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.29.ssm_dt
F32
F32
[48]
blk.29.ssm_norm.weight
F32
F32
[128]
blk.29.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.30
blk.30.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.30.attn_norm.weight
F32
F32
[5120]
blk.30.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.30.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.30.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.30.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.30.post_attention_norm.weight
F32
F32
[5120]
blk.30.ssm_a
F32
F32
[48]
blk.30.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.30.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.30.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.30.ssm_dt
F32
F32
[48]
blk.30.ssm_norm.weight
F32
F32
[128]
blk.30.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.31
blk.31.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.31.attn_k_norm.weight
F32
F32
[256]
blk.31.attn_norm.weight
F32
F32
[5120]
blk.31.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.31.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.31.attn_q_norm.weight
F32
F32
[256]
blk.31.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.31.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.31.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.31.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.31.post_attention_norm.weight
F32
F32
[5120]
blk.32
blk.32.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.32.attn_norm.weight
F32
F32
[5120]
blk.32.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.32.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.32.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.32.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.32.post_attention_norm.weight
F32
F32
[5120]
blk.32.ssm_a
F32
F32
[48]
blk.32.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.32.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.32.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.32.ssm_dt
F32
F32
[48]
blk.32.ssm_norm.weight
F32
F32
[128]
blk.32.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.33
blk.33.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.33.attn_norm.weight
F32
F32
[5120]
blk.33.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.33.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.33.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.33.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.33.post_attention_norm.weight
F32
F32
[5120]
blk.33.ssm_a
F32
F32
[48]
blk.33.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.33.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.33.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.33.ssm_dt
F32
F32
[48]
blk.33.ssm_norm.weight
F32
F32
[128]
blk.33.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.34
blk.34.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.34.attn_norm.weight
F32
F32
[5120]
blk.34.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.34.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.34.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.34.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.34.post_attention_norm.weight
F32
F32
[5120]
blk.34.ssm_a
F32
F32
[48]
blk.34.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.34.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.34.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.34.ssm_dt
F32
F32
[48]
blk.34.ssm_norm.weight
F32
F32
[128]
blk.34.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.35
blk.35.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.35.attn_k_norm.weight
F32
F32
[256]
blk.35.attn_norm.weight
F32
F32
[5120]
blk.35.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.35.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.35.attn_q_norm.weight
F32
F32
[256]
blk.35.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.35.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.35.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.35.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.35.post_attention_norm.weight
F32
F32
[5120]
blk.36
blk.36.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.36.attn_norm.weight
F32
F32
[5120]
blk.36.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.36.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.36.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.36.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.36.post_attention_norm.weight
F32
F32
[5120]
blk.36.ssm_a
F32
F32
[48]
blk.36.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.36.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.36.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.36.ssm_dt
F32
F32
[48]
blk.36.ssm_norm.weight
F32
F32
[128]
blk.36.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.37
blk.37.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.37.attn_norm.weight
F32
F32
[5120]
blk.37.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.37.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.37.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.37.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.37.post_attention_norm.weight
F32
F32
[5120]
blk.37.ssm_a
F32
F32
[48]
blk.37.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.37.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.37.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.37.ssm_dt
F32
F32
[48]
blk.37.ssm_norm.weight
F32
F32
[128]
blk.37.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.38
blk.38.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.38.attn_norm.weight
F32
F32
[5120]
blk.38.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.38.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.38.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.38.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.38.post_attention_norm.weight
F32
F32
[5120]
blk.38.ssm_a
F32
F32
[48]
blk.38.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.38.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.38.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.38.ssm_dt
F32
F32
[48]
blk.38.ssm_norm.weight
F32
F32
[128]
blk.38.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.39
blk.39.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.39.attn_k_norm.weight
F32
F32
[256]
blk.39.attn_norm.weight
F32
F32
[5120]
blk.39.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.39.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.39.attn_q_norm.weight
F32
F32
[256]
blk.39.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.39.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.39.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.39.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.39.post_attention_norm.weight
F32
F32
[5120]
blk.40
blk.40.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.40.attn_norm.weight
F32
F32
[5120]
blk.40.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.40.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.40.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.40.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.40.post_attention_norm.weight
F32
F32
[5120]
blk.40.ssm_a
F32
F32
[48]
blk.40.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.40.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.40.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.40.ssm_dt
F32
F32
[48]
blk.40.ssm_norm.weight
F32
F32
[128]
blk.40.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.41
blk.41.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.41.attn_norm.weight
F32
F32
[5120]
blk.41.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.41.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.41.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.41.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.41.post_attention_norm.weight
F32
F32
[5120]
blk.41.ssm_a
F32
F32
[48]
blk.41.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.41.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.41.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.41.ssm_dt
F32
F32
[48]
blk.41.ssm_norm.weight
F32
F32
[128]
blk.41.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.42
blk.42.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.42.attn_norm.weight
F32
F32
[5120]
blk.42.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.42.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.42.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.42.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.42.post_attention_norm.weight
F32
F32
[5120]
blk.42.ssm_a
F32
F32
[48]
blk.42.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.42.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.42.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.42.ssm_dt
F32
F32
[48]
blk.42.ssm_norm.weight
F32
F32
[128]
blk.42.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.43
blk.43.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.43.attn_k_norm.weight
F32
F32
[256]
blk.43.attn_norm.weight
F32
F32
[5120]
blk.43.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.43.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.43.attn_q_norm.weight
F32
F32
[256]
blk.43.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.43.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.43.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.43.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.43.post_attention_norm.weight
F32
F32
[5120]
blk.44
blk.44.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.44.attn_norm.weight
F32
F32
[5120]
blk.44.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.44.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.44.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.44.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.44.post_attention_norm.weight
F32
F32
[5120]
blk.44.ssm_a
F32
F32
[48]
blk.44.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.44.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.44.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.44.ssm_dt
F32
F32
[48]
blk.44.ssm_norm.weight
F32
F32
[128]
blk.44.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.45
blk.45.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.45.attn_norm.weight
F32
F32
[5120]
blk.45.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.45.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.45.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.45.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.45.post_attention_norm.weight
F32
F32
[5120]
blk.45.ssm_a
F32
F32
[48]
blk.45.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.45.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.45.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.45.ssm_dt
F32
F32
[48]
blk.45.ssm_norm.weight
F32
F32
[128]
blk.45.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.46
blk.46.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.46.attn_norm.weight
F32
F32
[5120]
blk.46.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.46.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.46.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.46.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.46.post_attention_norm.weight
F32
F32
[5120]
blk.46.ssm_a
F32
F32
[48]
blk.46.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.46.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.46.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.46.ssm_dt
F32
F32
[48]
blk.46.ssm_norm.weight
F32
F32
[128]
blk.46.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.47
blk.47.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.47.attn_k_norm.weight
F32
F32
[256]
blk.47.attn_norm.weight
F32
F32
[5120]
blk.47.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.47.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.47.attn_q_norm.weight
F32
F32
[256]
blk.47.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.47.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.47.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.47.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.47.post_attention_norm.weight
F32
F32
[5120]
blk.48
blk.48.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.48.attn_norm.weight
F32
F32
[5120]
blk.48.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.48.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.48.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.48.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.48.post_attention_norm.weight
F32
F32
[5120]
blk.48.ssm_a
F32
F32
[48]
blk.48.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.48.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.48.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.48.ssm_dt
F32
F32
[48]
blk.48.ssm_norm.weight
F32
F32
[128]
blk.48.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.49
blk.49.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.49.attn_norm.weight
F32
F32
[5120]
blk.49.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.49.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.49.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.49.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.49.post_attention_norm.weight
F32
F32
[5120]
blk.49.ssm_a
F32
F32
[48]
blk.49.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.49.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.49.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.49.ssm_dt
F32
F32
[48]
blk.49.ssm_norm.weight
F32
F32
[128]
blk.49.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.50
blk.50.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.50.attn_norm.weight
F32
F32
[5120]
blk.50.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.50.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.50.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.50.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.50.post_attention_norm.weight
F32
F32
[5120]
blk.50.ssm_a
F32
F32
[48]
blk.50.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.50.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.50.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.50.ssm_dt
F32
F32
[48]
blk.50.ssm_norm.weight
F32
F32
[128]
blk.50.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.51
blk.51.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.51.attn_k_norm.weight
F32
F32
[256]
blk.51.attn_norm.weight
F32
F32
[5120]
blk.51.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.51.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.51.attn_q_norm.weight
F32
F32
[256]
blk.51.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.51.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.51.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.51.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.51.post_attention_norm.weight
F32
F32
[5120]
blk.52
blk.52.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.52.attn_norm.weight
F32
F32
[5120]
blk.52.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.52.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.52.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.52.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.52.post_attention_norm.weight
F32
F32
[5120]
blk.52.ssm_a
F32
F32
[48]
blk.52.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.52.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.52.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.52.ssm_dt
F32
F32
[48]
blk.52.ssm_norm.weight
F32
F32
[128]
blk.52.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.53
blk.53.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.53.attn_norm.weight
F32
F32
[5120]
blk.53.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.53.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.53.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.53.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.53.post_attention_norm.weight
F32
F32
[5120]
blk.53.ssm_a
F32
F32
[48]
blk.53.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.53.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.53.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.53.ssm_dt
F32
F32
[48]
blk.53.ssm_norm.weight
F32
F32
[128]
blk.53.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.54
blk.54.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.54.attn_norm.weight
F32
F32
[5120]
blk.54.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.54.ffn_down.weight
Q4_K
Q4_K
[17408, 5120]
blk.54.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.54.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.54.post_attention_norm.weight
F32
F32
[5120]
blk.54.ssm_a
F32
F32
[48]
blk.54.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.54.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.54.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.54.ssm_dt
F32
F32
[48]
blk.54.ssm_norm.weight
F32
F32
[128]
blk.54.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.55
blk.55.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.55.attn_k_norm.weight
F32
F32
[256]
blk.55.attn_norm.weight
F32
F32
[5120]
blk.55.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.55.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.55.attn_q_norm.weight
F32
F32
[256]
blk.55.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.55.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.55.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.55.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.55.post_attention_norm.weight
F32
F32
[5120]
blk.56
blk.56.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.56.attn_norm.weight
F32
F32
[5120]
blk.56.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.56.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.56.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.56.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.56.post_attention_norm.weight
F32
F32
[5120]
blk.56.ssm_a
F32
F32
[48]
blk.56.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.56.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.56.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.56.ssm_dt
F32
F32
[48]
blk.56.ssm_norm.weight
F32
F32
[128]
blk.56.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.57
blk.57.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.57.attn_norm.weight
F32
F32
[5120]
blk.57.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.57.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.57.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.57.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.57.post_attention_norm.weight
F32
F32
[5120]
blk.57.ssm_a
F32
F32
[48]
blk.57.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.57.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.57.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.57.ssm_dt
F32
F32
[48]
blk.57.ssm_norm.weight
F32
F32
[128]
blk.57.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.58
blk.58.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.58.attn_norm.weight
F32
F32
[5120]
blk.58.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.58.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.58.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.58.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.58.post_attention_norm.weight
F32
F32
[5120]
blk.58.ssm_a
F32
F32
[48]
blk.58.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.58.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.58.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.58.ssm_dt
F32
F32
[48]
blk.58.ssm_norm.weight
F32
F32
[128]
blk.58.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.59
blk.59.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.59.attn_k_norm.weight
F32
F32
[256]
blk.59.attn_norm.weight
F32
F32
[5120]
blk.59.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.59.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.59.attn_q_norm.weight
F32
F32
[256]
blk.59.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.59.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.59.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.59.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.59.post_attention_norm.weight
F32
F32
[5120]
blk.60
blk.60.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.60.attn_norm.weight
F32
F32
[5120]
blk.60.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.60.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.60.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.60.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.60.post_attention_norm.weight
F32
F32
[5120]
blk.60.ssm_a
F32
F32
[48]
blk.60.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.60.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.60.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.60.ssm_dt
F32
F32
[48]
blk.60.ssm_norm.weight
F32
F32
[128]
blk.60.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.61
blk.61.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.61.attn_norm.weight
F32
F32
[5120]
blk.61.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.61.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.61.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.61.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.61.post_attention_norm.weight
F32
F32
[5120]
blk.61.ssm_a
F32
F32
[48]
blk.61.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.61.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.61.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.61.ssm_dt
F32
F32
[48]
blk.61.ssm_norm.weight
F32
F32
[128]
blk.61.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.62
blk.62.attn_gate.weight
Q4_K
Q4_K
[5120, 6144]
blk.62.attn_norm.weight
F32
F32
[5120]
blk.62.attn_qkv.weight
Q4_K
Q4_K
[5120, 10240]
blk.62.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.62.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.62.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.62.post_attention_norm.weight
F32
F32
[5120]
blk.62.ssm_a
F32
F32
[48]
blk.62.ssm_alpha.weight
Q4_K
Q4_K
[5120, 48]
blk.62.ssm_beta.weight
Q4_K
Q4_K
[5120, 48]
blk.62.ssm_conv1d.weight
F32
F32
[4, 10240]
blk.62.ssm_dt
F32
F32
[48]
blk.62.ssm_norm.weight
F32
F32
[128]
blk.62.ssm_out.weight
Q4_K
Q4_K
[6144, 5120]
blk.63
blk.63.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
blk.63.attn_k_norm.weight
F32
F32
[256]
blk.63.attn_norm.weight
F32
F32
[5120]
blk.63.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
blk.63.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
blk.63.attn_q_norm.weight
F32
F32
[256]
blk.63.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
blk.63.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
blk.63.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
blk.63.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
blk.63.post_attention_norm.weight
F32
F32
[5120]
mtp.fc.weight
Q4_K
Q4_K
[10240, 5120]
mtp.layers.0.attn_k.weight
Q4_K
Q4_K
[5120, 1024]
mtp.layers.0.attn_k_norm.weight
F32
F32
[256]
mtp.layers.0.attn_norm.weight
F32
F32
[5120]
mtp.layers.0.attn_output.weight
Q4_K
Q4_K
[6144, 5120]
mtp.layers.0.attn_q.weight
Q4_K
Q4_K
[5120, 12288]
mtp.layers.0.attn_q_norm.weight
F32
F32
[256]
mtp.layers.0.attn_v.weight
Q6_K
Q6_K
[5120, 1024]
mtp.layers.0.ffn_down.weight
Q6_K
Q6_K
[17408, 5120]
mtp.layers.0.ffn_gate.weight
Q4_K
Q4_K
[5120, 17408]
mtp.layers.0.ffn_up.weight
Q4_K
Q4_K
[5120, 17408]
mtp.layers.0.post_attention_norm.weight
F32
F32
[5120]
mtp.norm.weight
F32
F32
[5120]
mtp.pre_fc_norm_embedding.weight
F32
F32
[5120]
mtp.pre_fc_norm_hidden.weight
F32
F32
[5120]
output.weight
Q6_K
Q6_K
[5120, 248320]
v.blk.0
v.blk.0.attn_k.bias
F32
F32
[1152]
v.blk.0.attn_k.weight
F16
F16
[1152, 1152]
v.blk.0.attn_out.bias
F32
F32
[1152]
v.blk.0.attn_out.weight
F16
F16
[1152, 1152]
v.blk.0.attn_q.bias
F32
F32
[1152]
v.blk.0.attn_q.weight
F16
F16
[1152, 1152]
v.blk.0.attn_v.bias
F32
F32
[1152]
v.blk.0.attn_v.weight
F16
F16
[1152, 1152]
v.blk.0.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.0.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.0.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.0.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.0.norm1.bias
F32
F32
[1152]
v.blk.0.norm1.weight
F32
F32
[1152]
v.blk.0.norm2.bias
F32
F32
[1152]
v.blk.0.norm2.weight
F32
F32
[1152]
v.blk.1
v.blk.1.attn_k.bias
F32
F32
[1152]
v.blk.1.attn_k.weight
F16
F16
[1152, 1152]
v.blk.1.attn_out.bias
F32
F32
[1152]
v.blk.1.attn_out.weight
F16
F16
[1152, 1152]
v.blk.1.attn_q.bias
F32
F32
[1152]
v.blk.1.attn_q.weight
F16
F16
[1152, 1152]
v.blk.1.attn_v.bias
F32
F32
[1152]
v.blk.1.attn_v.weight
F16
F16
[1152, 1152]
v.blk.1.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.1.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.1.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.1.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.1.norm1.bias
F32
F32
[1152]
v.blk.1.norm1.weight
F32
F32
[1152]
v.blk.1.norm2.bias
F32
F32
[1152]
v.blk.1.norm2.weight
F32
F32
[1152]
v.blk.2
v.blk.2.attn_k.bias
F32
F32
[1152]
v.blk.2.attn_k.weight
F16
F16
[1152, 1152]
v.blk.2.attn_out.bias
F32
F32
[1152]
v.blk.2.attn_out.weight
F16
F16
[1152, 1152]
v.blk.2.attn_q.bias
F32
F32
[1152]
v.blk.2.attn_q.weight
F16
F16
[1152, 1152]
v.blk.2.attn_v.bias
F32
F32
[1152]
v.blk.2.attn_v.weight
F16
F16
[1152, 1152]
v.blk.2.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.2.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.2.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.2.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.2.norm1.bias
F32
F32
[1152]
v.blk.2.norm1.weight
F32
F32
[1152]
v.blk.2.norm2.bias
F32
F32
[1152]
v.blk.2.norm2.weight
F32
F32
[1152]
v.blk.3
v.blk.3.attn_k.bias
F32
F32
[1152]
v.blk.3.attn_k.weight
F16
F16
[1152, 1152]
v.blk.3.attn_out.bias
F32
F32
[1152]
v.blk.3.attn_out.weight
F16
F16
[1152, 1152]
v.blk.3.attn_q.bias
F32
F32
[1152]
v.blk.3.attn_q.weight
F16
F16
[1152, 1152]
v.blk.3.attn_v.bias
F32
F32
[1152]
v.blk.3.attn_v.weight
F16
F16
[1152, 1152]
v.blk.3.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.3.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.3.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.3.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.3.norm1.bias
F32
F32
[1152]
v.blk.3.norm1.weight
F32
F32
[1152]
v.blk.3.norm2.bias
F32
F32
[1152]
v.blk.3.norm2.weight
F32
F32
[1152]
v.blk.4
v.blk.4.attn_k.bias
F32
F32
[1152]
v.blk.4.attn_k.weight
F16
F16
[1152, 1152]
v.blk.4.attn_out.bias
F32
F32
[1152]
v.blk.4.attn_out.weight
F16
F16
[1152, 1152]
v.blk.4.attn_q.bias
F32
F32
[1152]
v.blk.4.attn_q.weight
F16
F16
[1152, 1152]
v.blk.4.attn_v.bias
F32
F32
[1152]
v.blk.4.attn_v.weight
F16
F16
[1152, 1152]
v.blk.4.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.4.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.4.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.4.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.4.norm1.bias
F32
F32
[1152]
v.blk.4.norm1.weight
F32
F32
[1152]
v.blk.4.norm2.bias
F32
F32
[1152]
v.blk.4.norm2.weight
F32
F32
[1152]
v.blk.5
v.blk.5.attn_k.bias
F32
F32
[1152]
v.blk.5.attn_k.weight
F16
F16
[1152, 1152]
v.blk.5.attn_out.bias
F32
F32
[1152]
v.blk.5.attn_out.weight
F16
F16
[1152, 1152]
v.blk.5.attn_q.bias
F32
F32
[1152]
v.blk.5.attn_q.weight
F16
F16
[1152, 1152]
v.blk.5.attn_v.bias
F32
F32
[1152]
v.blk.5.attn_v.weight
F16
F16
[1152, 1152]
v.blk.5.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.5.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.5.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.5.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.5.norm1.bias
F32
F32
[1152]
v.blk.5.norm1.weight
F32
F32
[1152]
v.blk.5.norm2.bias
F32
F32
[1152]
v.blk.5.norm2.weight
F32
F32
[1152]
v.blk.6
v.blk.6.attn_k.bias
F32
F32
[1152]
v.blk.6.attn_k.weight
F16
F16
[1152, 1152]
v.blk.6.attn_out.bias
F32
F32
[1152]
v.blk.6.attn_out.weight
F16
F16
[1152, 1152]
v.blk.6.attn_q.bias
F32
F32
[1152]
v.blk.6.attn_q.weight
F16
F16
[1152, 1152]
v.blk.6.attn_v.bias
F32
F32
[1152]
v.blk.6.attn_v.weight
F16
F16
[1152, 1152]
v.blk.6.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.6.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.6.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.6.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.6.norm1.bias
F32
F32
[1152]
v.blk.6.norm1.weight
F32
F32
[1152]
v.blk.6.norm2.bias
F32
F32
[1152]
v.blk.6.norm2.weight
F32
F32
[1152]
v.blk.7
v.blk.7.attn_k.bias
F32
F32
[1152]
v.blk.7.attn_k.weight
F16
F16
[1152, 1152]
v.blk.7.attn_out.bias
F32
F32
[1152]
v.blk.7.attn_out.weight
F16
F16
[1152, 1152]
v.blk.7.attn_q.bias
F32
F32
[1152]
v.blk.7.attn_q.weight
F16
F16
[1152, 1152]
v.blk.7.attn_v.bias
F32
F32
[1152]
v.blk.7.attn_v.weight
F16
F16
[1152, 1152]
v.blk.7.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.7.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.7.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.7.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.7.norm1.bias
F32
F32
[1152]
v.blk.7.norm1.weight
F32
F32
[1152]
v.blk.7.norm2.bias
F32
F32
[1152]
v.blk.7.norm2.weight
F32
F32
[1152]
v.blk.8
v.blk.8.attn_k.bias
F32
F32
[1152]
v.blk.8.attn_k.weight
F16
F16
[1152, 1152]
v.blk.8.attn_out.bias
F32
F32
[1152]
v.blk.8.attn_out.weight
F16
F16
[1152, 1152]
v.blk.8.attn_q.bias
F32
F32
[1152]
v.blk.8.attn_q.weight
F16
F16
[1152, 1152]
v.blk.8.attn_v.bias
F32
F32
[1152]
v.blk.8.attn_v.weight
F16
F16
[1152, 1152]
v.blk.8.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.8.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.8.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.8.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.8.norm1.bias
F32
F32
[1152]
v.blk.8.norm1.weight
F32
F32
[1152]
v.blk.8.norm2.bias
F32
F32
[1152]
v.blk.8.norm2.weight
F32
F32
[1152]
v.blk.9
v.blk.9.attn_k.bias
F32
F32
[1152]
v.blk.9.attn_k.weight
F16
F16
[1152, 1152]
v.blk.9.attn_out.bias
F32
F32
[1152]
v.blk.9.attn_out.weight
F16
F16
[1152, 1152]
v.blk.9.attn_q.bias
F32
F32
[1152]
v.blk.9.attn_q.weight
F16
F16
[1152, 1152]
v.blk.9.attn_v.bias
F32
F32
[1152]
v.blk.9.attn_v.weight
F16
F16
[1152, 1152]
v.blk.9.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.9.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.9.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.9.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.9.norm1.bias
F32
F32
[1152]
v.blk.9.norm1.weight
F32
F32
[1152]
v.blk.9.norm2.bias
F32
F32
[1152]
v.blk.9.norm2.weight
F32
F32
[1152]
v.blk.10
v.blk.10.attn_k.bias
F32
F32
[1152]
v.blk.10.attn_k.weight
F16
F16
[1152, 1152]
v.blk.10.attn_out.bias
F32
F32
[1152]
v.blk.10.attn_out.weight
F16
F16
[1152, 1152]
v.blk.10.attn_q.bias
F32
F32
[1152]
v.blk.10.attn_q.weight
F16
F16
[1152, 1152]
v.blk.10.attn_v.bias
F32
F32
[1152]
v.blk.10.attn_v.weight
F16
F16
[1152, 1152]
v.blk.10.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.10.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.10.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.10.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.10.norm1.bias
F32
F32
[1152]
v.blk.10.norm1.weight
F32
F32
[1152]
v.blk.10.norm2.bias
F32
F32
[1152]
v.blk.10.norm2.weight
F32
F32
[1152]
v.blk.11
v.blk.11.attn_k.bias
F32
F32
[1152]
v.blk.11.attn_k.weight
F16
F16
[1152, 1152]
v.blk.11.attn_out.bias
F32
F32
[1152]
v.blk.11.attn_out.weight
F16
F16
[1152, 1152]
v.blk.11.attn_q.bias
F32
F32
[1152]
v.blk.11.attn_q.weight
F16
F16
[1152, 1152]
v.blk.11.attn_v.bias
F32
F32
[1152]
v.blk.11.attn_v.weight
F16
F16
[1152, 1152]
v.blk.11.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.11.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.11.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.11.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.11.norm1.bias
F32
F32
[1152]
v.blk.11.norm1.weight
F32
F32
[1152]
v.blk.11.norm2.bias
F32
F32
[1152]
v.blk.11.norm2.weight
F32
F32
[1152]
v.blk.12
v.blk.12.attn_k.bias
F32
F32
[1152]
v.blk.12.attn_k.weight
F16
F16
[1152, 1152]
v.blk.12.attn_out.bias
F32
F32
[1152]
v.blk.12.attn_out.weight
F16
F16
[1152, 1152]
v.blk.12.attn_q.bias
F32
F32
[1152]
v.blk.12.attn_q.weight
F16
F16
[1152, 1152]
v.blk.12.attn_v.bias
F32
F32
[1152]
v.blk.12.attn_v.weight
F16
F16
[1152, 1152]
v.blk.12.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.12.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.12.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.12.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.12.norm1.bias
F32
F32
[1152]
v.blk.12.norm1.weight
F32
F32
[1152]
v.blk.12.norm2.bias
F32
F32
[1152]
v.blk.12.norm2.weight
F32
F32
[1152]
v.blk.13
v.blk.13.attn_k.bias
F32
F32
[1152]
v.blk.13.attn_k.weight
F16
F16
[1152, 1152]
v.blk.13.attn_out.bias
F32
F32
[1152]
v.blk.13.attn_out.weight
F16
F16
[1152, 1152]
v.blk.13.attn_q.bias
F32
F32
[1152]
v.blk.13.attn_q.weight
F16
F16
[1152, 1152]
v.blk.13.attn_v.bias
F32
F32
[1152]
v.blk.13.attn_v.weight
F16
F16
[1152, 1152]
v.blk.13.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.13.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.13.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.13.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.13.norm1.bias
F32
F32
[1152]
v.blk.13.norm1.weight
F32
F32
[1152]
v.blk.13.norm2.bias
F32
F32
[1152]
v.blk.13.norm2.weight
F32
F32
[1152]
v.blk.14
v.blk.14.attn_k.bias
F32
F32
[1152]
v.blk.14.attn_k.weight
F16
F16
[1152, 1152]
v.blk.14.attn_out.bias
F32
F32
[1152]
v.blk.14.attn_out.weight
F16
F16
[1152, 1152]
v.blk.14.attn_q.bias
F32
F32
[1152]
v.blk.14.attn_q.weight
F16
F16
[1152, 1152]
v.blk.14.attn_v.bias
F32
F32
[1152]
v.blk.14.attn_v.weight
F16
F16
[1152, 1152]
v.blk.14.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.14.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.14.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.14.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.14.norm1.bias
F32
F32
[1152]
v.blk.14.norm1.weight
F32
F32
[1152]
v.blk.14.norm2.bias
F32
F32
[1152]
v.blk.14.norm2.weight
F32
F32
[1152]
v.blk.15
v.blk.15.attn_k.bias
F32
F32
[1152]
v.blk.15.attn_k.weight
F16
F16
[1152, 1152]
v.blk.15.attn_out.bias
F32
F32
[1152]
v.blk.15.attn_out.weight
F16
F16
[1152, 1152]
v.blk.15.attn_q.bias
F32
F32
[1152]
v.blk.15.attn_q.weight
F16
F16
[1152, 1152]
v.blk.15.attn_v.bias
F32
F32
[1152]
v.blk.15.attn_v.weight
F16
F16
[1152, 1152]
v.blk.15.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.15.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.15.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.15.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.15.norm1.bias
F32
F32
[1152]
v.blk.15.norm1.weight
F32
F32
[1152]
v.blk.15.norm2.bias
F32
F32
[1152]
v.blk.15.norm2.weight
F32
F32
[1152]
v.blk.16
v.blk.16.attn_k.bias
F32
F32
[1152]
v.blk.16.attn_k.weight
F16
F16
[1152, 1152]
v.blk.16.attn_out.bias
F32
F32
[1152]
v.blk.16.attn_out.weight
F16
F16
[1152, 1152]
v.blk.16.attn_q.bias
F32
F32
[1152]
v.blk.16.attn_q.weight
F16
F16
[1152, 1152]
v.blk.16.attn_v.bias
F32
F32
[1152]
v.blk.16.attn_v.weight
F16
F16
[1152, 1152]
v.blk.16.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.16.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.16.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.16.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.16.norm1.bias
F32
F32
[1152]
v.blk.16.norm1.weight
F32
F32
[1152]
v.blk.16.norm2.bias
F32
F32
[1152]
v.blk.16.norm2.weight
F32
F32
[1152]
v.blk.17
v.blk.17.attn_k.bias
F32
F32
[1152]
v.blk.17.attn_k.weight
F16
F16
[1152, 1152]
v.blk.17.attn_out.bias
F32
F32
[1152]
v.blk.17.attn_out.weight
F16
F16
[1152, 1152]
v.blk.17.attn_q.bias
F32
F32
[1152]
v.blk.17.attn_q.weight
F16
F16
[1152, 1152]
v.blk.17.attn_v.bias
F32
F32
[1152]
v.blk.17.attn_v.weight
F16
F16
[1152, 1152]
v.blk.17.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.17.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.17.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.17.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.17.norm1.bias
F32
F32
[1152]
v.blk.17.norm1.weight
F32
F32
[1152]
v.blk.17.norm2.bias
F32
F32
[1152]
v.blk.17.norm2.weight
F32
F32
[1152]
v.blk.18
v.blk.18.attn_k.bias
F32
F32
[1152]
v.blk.18.attn_k.weight
F16
F16
[1152, 1152]
v.blk.18.attn_out.bias
F32
F32
[1152]
v.blk.18.attn_out.weight
F16
F16
[1152, 1152]
v.blk.18.attn_q.bias
F32
F32
[1152]
v.blk.18.attn_q.weight
F16
F16
[1152, 1152]
v.blk.18.attn_v.bias
F32
F32
[1152]
v.blk.18.attn_v.weight
F16
F16
[1152, 1152]
v.blk.18.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.18.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.18.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.18.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.18.norm1.bias
F32
F32
[1152]
v.blk.18.norm1.weight
F32
F32
[1152]
v.blk.18.norm2.bias
F32
F32
[1152]
v.blk.18.norm2.weight
F32
F32
[1152]
v.blk.19
v.blk.19.attn_k.bias
F32
F32
[1152]
v.blk.19.attn_k.weight
F16
F16
[1152, 1152]
v.blk.19.attn_out.bias
F32
F32
[1152]
v.blk.19.attn_out.weight
F16
F16
[1152, 1152]
v.blk.19.attn_q.bias
F32
F32
[1152]
v.blk.19.attn_q.weight
F16
F16
[1152, 1152]
v.blk.19.attn_v.bias
F32
F32
[1152]
v.blk.19.attn_v.weight
F16
F16
[1152, 1152]
v.blk.19.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.19.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.19.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.19.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.19.norm1.bias
F32
F32
[1152]
v.blk.19.norm1.weight
F32
F32
[1152]
v.blk.19.norm2.bias
F32
F32
[1152]
v.blk.19.norm2.weight
F32
F32
[1152]
v.blk.20
v.blk.20.attn_k.bias
F32
F32
[1152]
v.blk.20.attn_k.weight
F16
F16
[1152, 1152]
v.blk.20.attn_out.bias
F32
F32
[1152]
v.blk.20.attn_out.weight
F16
F16
[1152, 1152]
v.blk.20.attn_q.bias
F32
F32
[1152]
v.blk.20.attn_q.weight
F16
F16
[1152, 1152]
v.blk.20.attn_v.bias
F32
F32
[1152]
v.blk.20.attn_v.weight
F16
F16
[1152, 1152]
v.blk.20.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.20.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.20.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.20.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.20.norm1.bias
F32
F32
[1152]
v.blk.20.norm1.weight
F32
F32
[1152]
v.blk.20.norm2.bias
F32
F32
[1152]
v.blk.20.norm2.weight
F32
F32
[1152]
v.blk.21
v.blk.21.attn_k.bias
F32
F32
[1152]
v.blk.21.attn_k.weight
F16
F16
[1152, 1152]
v.blk.21.attn_out.bias
F32
F32
[1152]
v.blk.21.attn_out.weight
F16
F16
[1152, 1152]
v.blk.21.attn_q.bias
F32
F32
[1152]
v.blk.21.attn_q.weight
F16
F16
[1152, 1152]
v.blk.21.attn_v.bias
F32
F32
[1152]
v.blk.21.attn_v.weight
F16
F16
[1152, 1152]
v.blk.21.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.21.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.21.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.21.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.21.norm1.bias
F32
F32
[1152]
v.blk.21.norm1.weight
F32
F32
[1152]
v.blk.21.norm2.bias
F32
F32
[1152]
v.blk.21.norm2.weight
F32
F32
[1152]
v.blk.22
v.blk.22.attn_k.bias
F32
F32
[1152]
v.blk.22.attn_k.weight
F16
F16
[1152, 1152]
v.blk.22.attn_out.bias
F32
F32
[1152]
v.blk.22.attn_out.weight
F16
F16
[1152, 1152]
v.blk.22.attn_q.bias
F32
F32
[1152]
v.blk.22.attn_q.weight
F16
F16
[1152, 1152]
v.blk.22.attn_v.bias
F32
F32
[1152]
v.blk.22.attn_v.weight
F16
F16
[1152, 1152]
v.blk.22.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.22.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.22.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.22.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.22.norm1.bias
F32
F32
[1152]
v.blk.22.norm1.weight
F32
F32
[1152]
v.blk.22.norm2.bias
F32
F32
[1152]
v.blk.22.norm2.weight
F32
F32
[1152]
v.blk.23
v.blk.23.attn_k.bias
F32
F32
[1152]
v.blk.23.attn_k.weight
F16
F16
[1152, 1152]
v.blk.23.attn_out.bias
F32
F32
[1152]
v.blk.23.attn_out.weight
F16
F16
[1152, 1152]
v.blk.23.attn_q.bias
F32
F32
[1152]
v.blk.23.attn_q.weight
F16
F16
[1152, 1152]
v.blk.23.attn_v.bias
F32
F32
[1152]
v.blk.23.attn_v.weight
F16
F16
[1152, 1152]
v.blk.23.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.23.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.23.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.23.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.23.norm1.bias
F32
F32
[1152]
v.blk.23.norm1.weight
F32
F32
[1152]
v.blk.23.norm2.bias
F32
F32
[1152]
v.blk.23.norm2.weight
F32
F32
[1152]
v.blk.24
v.blk.24.attn_k.bias
F32
F32
[1152]
v.blk.24.attn_k.weight
F16
F16
[1152, 1152]
v.blk.24.attn_out.bias
F32
F32
[1152]
v.blk.24.attn_out.weight
F16
F16
[1152, 1152]
v.blk.24.attn_q.bias
F32
F32
[1152]
v.blk.24.attn_q.weight
F16
F16
[1152, 1152]
v.blk.24.attn_v.bias
F32
F32
[1152]
v.blk.24.attn_v.weight
F16
F16
[1152, 1152]
v.blk.24.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.24.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.24.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.24.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.24.norm1.bias
F32
F32
[1152]
v.blk.24.norm1.weight
F32
F32
[1152]
v.blk.24.norm2.bias
F32
F32
[1152]
v.blk.24.norm2.weight
F32
F32
[1152]
v.blk.25
v.blk.25.attn_k.bias
F32
F32
[1152]
v.blk.25.attn_k.weight
F16
F16
[1152, 1152]
v.blk.25.attn_out.bias
F32
F32
[1152]
v.blk.25.attn_out.weight
F16
F16
[1152, 1152]
v.blk.25.attn_q.bias
F32
F32
[1152]
v.blk.25.attn_q.weight
F16
F16
[1152, 1152]
v.blk.25.attn_v.bias
F32
F32
[1152]
v.blk.25.attn_v.weight
F16
F16
[1152, 1152]
v.blk.25.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.25.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.25.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.25.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.25.norm1.bias
F32
F32
[1152]
v.blk.25.norm1.weight
F32
F32
[1152]
v.blk.25.norm2.bias
F32
F32
[1152]
v.blk.25.norm2.weight
F32
F32
[1152]
v.blk.26
v.blk.26.attn_k.bias
F32
F32
[1152]
v.blk.26.attn_k.weight
F16
F16
[1152, 1152]
v.blk.26.attn_out.bias
F32
F32
[1152]
v.blk.26.attn_out.weight
F16
F16
[1152, 1152]
v.blk.26.attn_q.bias
F32
F32
[1152]
v.blk.26.attn_q.weight
F16
F16
[1152, 1152]
v.blk.26.attn_v.bias
F32
F32
[1152]
v.blk.26.attn_v.weight
F16
F16
[1152, 1152]
v.blk.26.mlp.linear_fc1.bias
F32
F32
[4304]
v.blk.26.mlp.linear_fc1.weight
F16
F16
[1152, 4304]
v.blk.26.mlp.linear_fc2.bias
F32
F32
[1152]
v.blk.26.mlp.linear_fc2.weight
F16
F16
[4304, 1152]
v.blk.26.norm1.bias
F32
F32
[1152]
v.blk.26.norm1.weight
F32
F32
[1152]
v.blk.26.norm2.bias
F32
F32
[1152]
v.blk.26.norm2.weight
F32
F32
[1152]
v.merger.linear_fc1.bias
F32
F32
[4608]
v.merger.linear_fc1.weight
F16
F16
[4608, 4608]
v.merger.linear_fc2.bias
F32
F32
[5120]
v.merger.linear_fc2.weight
F16
F16
[4608, 5120]
v.merger.norm.bias
F32
F32
[1152]
v.merger.norm.weight
F32
F32
[1152]
v.patch_embed.bias
F32
F32
[1152]
v.patch_embed.weight
F16
F16
[16, 16, 2, 3456]
v.pos_embed.weight
F16
F16
[1152, 2304]
output_norm.weight
F32
F32
[5120]